Microscopía virtual en el diagnóstico de
rutina y la docencia en un hospital
universitario
TESIS DOCTORAL
Adela Saco Álvarez
Microscopía virtual en el diagnóstico rutinario y la docencia
[2]
Tesis Doctoral. Adela Saco Álvarez
[3]
Facultad de Medicina
Departament de Fonaments Clínics
Microscopía virtual en el diagnóstico de
rutina y la docencia en un hospital
universitario
Tesis Doctoral presentada por: Adela Saco Álvarez
Directores: Jaume Ordi Maja y José Ramírez Ruz
Barcelona, 2017
Tesis Doctoral. Adela Saco Álvarez
[5]
Jaume Ordi Maja y José Ramírez Ruz, profesores titulares de Anatomía Patológica de la
Universidad de Barcelona, certifican que la tesis doctoral titulada “Microscopía virtual
en el diagnóstico de rutina y la docencia en un hospital universitario” y presentada
por Adela Saco Álvarez, ha sido realizada bajo su dirección y cumple todos los
requisitos que dicta la normativa vigente para la presentación de tesis doctorales como
compendio de publicaciones en la Facultad de Medicina de la Universidad de
Barcelona.
Dr. Jaume Ordi Maja Dr. José Ramírez Ruz
Microscopía virtual en el diagnóstico rutinario y la docencia
[6]
Tesis Doctoral. Adela Saco Álvarez
[7]
Agradecimientos En primer lugar, me gustaría dar las gracias a Jaume y Rami por su inestimable ayuda y
porque sin ellos este trabajo no sería posible.
También me gustaría agradecer a mi familia, tanto gallega como leridana, y en especial
a Jose por apoyarme siempre y a Amelia por sus constantes sonrisas.
Microscopía virtual en el diagnóstico rutinario y la docencia
[8]
Tesis Doctoral. Adela Saco Álvarez
[9]
PUBLICACIONES INTERNACIONALES QUE COMPONEN ESTA TESIS DOCTORAL
Estudio 1
“Validation of Whole-slide imaging for histopathological diagnosis: current state”
Adela Saco, José Ramirez, Natalia Rakislova, Aurea Mira, Jaume Ordi
Pathobiology 2016; 83:89 – 98
Factor de impacto (2016): 1.703
Ranking (2016): 60/193, segundo cuartil
Estudio 2
“Validation of Whole-Slide Imaging in the Primary Diagnosis of Gynecological
Pathology in a University Hospital”
Jaume Ordi, Paola Castillo, Adela Saco, Marta del Pino, Oriol Ordi, Leonardo Rodríguez-
Carunchio, Jose Ramírez
Journal of Clinical Pathology 2015 Jan; 68(1): 33 - 9
Factor de impacto (2016): 2.687
Ranking (2016): 33/193, primer cuartil
Microscopía virtual en el diagnóstico rutinario y la docencia
[10]
Estudio 3
“Validation of Whole-Slide Imaging in the Primary Diagnosis of Liver Biopsies in a
University Hospital”
Adela Saco, Alba Diaz, Monica Hernandez, Daniel Martinez, Carla Montironi, Paola
Castillo, Natalia Rakislova, Marta del Pino, Antonio Martinez, Jaume Ordi
Dig Liver Dis. 2017 Jul 19. pii: S1590-8658(17)30977-5.
doi: 10.1016/j.dld.2017.07.002. [Epub ahead of print]
Factor de impacto (2016): 2.875
Ranking (2016): 35/134, Segundo cuartil
Estudio 4
“Current Status of Whole-Slide Imaging in Education”
Adela Saco, Josep Antoni Bombi, Adriana Garcia, José Ramirez, Jaume Ordi
Pathobiology 2016; 83:79 – 88
Factor de impacto (2016): 1.703
Ranking (2016): 60/193, segundo cuartil
Tesis Doctoral. Adela Saco Álvarez
[11]
Estudio 5
“Virtual Microscopy in the Undergraduate Teaching of Pathology”
Oriol Ordi, Josep Antoni Bombí, Antonio Martínez, Josep Ramírez, Llúcia Alòs,
Adela Saco, Teresa Ribalta, Pedro L. Fernández, Elias Campo, Jaume Ordi
Journal of Pathology Informatics 2015 Jan 29; 6:1
Factor de impacto (2016): 0
Microscopía virtual en el diagnóstico rutinario y la docencia
[12]
ÍNDICE
Lista de abreviaciones utilizadas…………………………………………………………………………. 15
INTRODUCCIÓN…………………………………………………………….…..………………………………………….. 17
1. Historia de la microscopía virtual…………………....................................... 19
2. Ventajas de la microscopía virtual
2.1 Ventajas de la microscopía virtual en el diagnóstico
rutinario……………………………………………………………………………
21
2.2 Ventajas de la microscopía virtual en la docencia……………. 23
3. Inconvenientes de la microscopía virtual
3.1 Inconvenientes de la microscopía virtual en el diagnóstico
de rutina..…………………………………………………………………………
25
3.2 Inconvenientes de la microscopía virtual en la docencia….. 26
4. Microscopía virtual en el diagnóstico de rutina
4.1 Integración del sistema de microscopía virtual en un
servicio de Anatomía Patológica……………………………………….
27
4.2 Validación de la microscopía virtual en el diagnóstico
rutinario……………………………………………………………………………
29
4.3 Estado de la validación en el momento actual………….……... 31
4.4 Aspectos técnicos en la digitalización de rutina………………… 32
Microscopía virtual en el diagnóstico rutinario y la docencia
[14]
4.5 Almacenamiento de archivos……………………………………………. 35
5. Microscopía virtual en la docencia……………………………………………………. 35
6. Microscopía virtual en comités multidisciplinares……………………………… 36
7. Microscopía virtual en la teleconsulta………………..……………………………… 37
HIPOTESIS DE LOS TRABAJOS……………………………………………………………..…………………………. 39
OBJETIVOS…………………………………………………………………….………………………………………………. 43
TRABAJOS REALIZADOS, MÉTODOS Y RESULTADOS………................................................. 47
ESTUDIO 1: Validation of Whole-slide Imaging for Histopathological Diagnosis:
Current State……………………………………………………………………………………………………
51
ESTUDIO 2: Validation of Whole-Slide Imaging in the Primary Diagnosis of
Gynecological Pathology in a University Hospital…………………………………………….
63
ESTUDIO 3: Validation of Whole-Slide Imaging in the Primary Diagnosis of
Liver Biopsies in a University Hospital………………………………………………………………
73
ESTUDIO 4: Current Status of Whole-Slide Imaging in Education…………………….. 83
ESTUDIO 5: Virtual Microscopy in the Undergraduate Teaching of Pathology…. 95
DISCUSIÓN…………………………………………………………………………………………………………………..... 103
CONCLUSIONES………………………………………………………………….………………………………………… 119
BIBLIOGRAFIA…………………………………………………………………………………………………................ 123
Tesis Doctoral. Adela Saco Álvarez
[15]
LISTA DE ABREVIACIONES UTILIZADAS
FISH: Hibridación in situ con fluorescencia (del ingles: fluorescence in situ hybridization)
H&E: Tinción de hematoxilina-eosina
H-SIL: lesión escamosa intraepitelial de alto grado (del inglés: high grade squamous
intraepithelial lesion)
LIS: Sistemas de información de laboratorio (del inglés: laboratory information
systems)
L-SIL: lesión escamosa intraepitelial de bajo grado (del inglés: low grade squamous
intraepithelial lesion)
MC: Microscopía convencional
MV: Microscopía virtual
QR: Respuesta rápida (del inglés: quick response)
Microscopía virtual en el diagnóstico rutinario y la docencia
[16]
Tesis Doctoral. Adela Saco Álvarez
[17]
I. Introducción
Microscopía virtual en el diagnóstico rutinario y la docencia
[18]
Tesis Doctoral. Adela Saco Álvarez
[19]
1. Historia de la microscopía virtual
La microscopía óptica convencional (MC) está ligada a la Anatomía Patológica
desde el inicio de esta especialidad. En los inicios, la evaluación de las muestras se
limitaba al estudio macroscópico e histológico, siendo la MC la base, y prácticamente
la única herramienta para realizar el diagnóstico de rutina de las biopsias. En las
últimas décadas se han incorporado diferentes técnicas útiles para el diagnóstico como
la inmunohistoquímica, la cual también se evalúa usando MC, y otras como la
patología molecular que no requiere ya del uso del microscopio.
Este método de diagnóstico hace que el patólogo dependa totalmente del uso
del microscopio óptico y de la presencia física de las preparaciones histológicas
montadas sobre portaobjetos de cristal. La posibilidad de consultar casos con otros
especialistas solo se puede realizar enviando las preparaciones o los bloques, lo que se
traduce en un importante aumento del tiempo en el diagnóstico y el riesgo de pérdida
o daño del material. El uso del microscopio óptico limita también la visualización
conjunta de casos, práctica especialmente necesaria en la docencia, y requiere el uso
de microscopios con múltiples cabezales. La solución de estos inconvenientes
representaría un gran avance, tanto en el diagnóstico rutinario como en la práctica de
consultas diagnósticas, sesiones y comités conjuntos con otras especialidades, o en la
docencia tanto de pre como de post-grado.
Este escenario empezó a cambiar hace algunas décadas con el desarrollo de la
informática, al aparecer los primeros ordenadores personales [1–3]. Las imágenes
estáticas del material histológico, adquiridas mediante máquinas fotográficas
acopladas al MC, fueron el primer avance digital en aparecer, estando dirigidas
principalmente a la docencia y en menor grado a la teleconsulta. Este método
presentaba grandes inconvenientes como la deficiente calidad de las imágenes y la
imposibilidad de navegar o emplear distintos objetivos, por lo que su uso para el
diagnóstico estaba muy limitado [4]. Posteriormente aparecieron sistemas de
telepatología dinámica en tiempo real con videocámaras integradas al MC. Su uso se
destinó casi exclusivamente a la visualización remota de cortes histológicos obtenidos
de forma convencional o, en la mayoría de los casos en congelación, lo que permitía el
Microscopía virtual en el diagnóstico rutinario y la docencia
[20]
diagnóstico de biopsias peroperatorias aún cuando el patólogo se encontrara situado
en otro centro distante. Esta tecnología resultó también muy útil en hospitales
pequeños en los que el patólogo especializado en algún área concreta no se
encontraba en el mismo centro. La telepatología dinámica en tiempo real permitía
realizar el diagnóstico de biopsias complicadas, y disminuía así la variabilidad
dependiente del centro [5–12]. Sin embargo, a pesar de representar un gran avance,
esta tecnología no permitía la navegación remota, requiriendo personal que se
encargase de mover la platina del microscopio con el portaobjetos y de cambiar el
objetivo óptico; además, la calidad de la imagen continuaba sin ser la óptima para
realizar el diagnóstico primario.
Los rápidos avances informáticos y tecnológicos en las últimas décadas
permitieron el desarrollo de los primeros escáneres capaces de crear una reproducción
digital a partir de una preparación histológica [2,3]. Estos escáneres son la base de la
patología digital o MV, la cual permite la navegación por la preparación histológica a
diferentes objetivos. La calidad de imagen y velocidad de los primeros escáneres
dificultaban el diagnóstico rutinario; además el coste económico era muy elevado, por
lo que esta tecnología se empleaba casi exclusivamente en ciertas áreas, como en
consultas diagnósticas o en docencia, y excluía el diagnóstico rutinario [10,13–17]. A
pesar de todos estos inconvenientes, esta tecnología abría la puerta a la posibilidad
real del diagnóstico virtual de rutina con todos los beneficios que podía conllevar.
En los últimos años han aparecido en el mercado numerosos escáneres con un
coste económico más bajo, con una gran calidad de imagen y con una adecuada
velocidad en la visualización. Se desarrollaron además múltiples programas
informáticos que permiten la visualización de las preparaciones virtuales usando
distintos objetivos y con numerosas herramientas que permiten tomar anotaciones,
poner marcas o realizar medidas [17–22]. Estas mejoras han acelerado la expansión de
esta tecnología y su uso tanto en docencia como en el diagnóstico rutinario en los
servicios de Anatomía Patológica.
A pesar de la gran calidad de imagen de las preparaciones virtuales, el uso de
esta tecnología ha mostrado algunos problemas que dificultan su implementación para
Tesis Doctoral. Adela Saco Álvarez
[21]
el diagnóstico rutinario en los servicios de Anatomía Patológica. La principal de ellas es
la fiabilidad de los diagnósticos emitidos con este método y la necesidad de realizar
validaciones antes de realizar el diagnóstico primario de las biopsias.
2. Ventajas de la microscopía virtual
2.1. Ventajas de la microscopía virtual en el diagnóstico rutinario
La MV presenta múltiples ventajas que resuelven gran parte de los problemas
que plantea el uso de la MC. Entre los principales destaca la imposibilidad de
visualización de preparaciones por un grupo amplio de usuarios, especialmente si
estos se encuentran separados físicamente. La MV permite la navegación y
visualización en una pantalla a tiempo real de las preparaciones, solucionando así este
inconveniente. Esta característica también influye en la presentación de casos en
sesiones o comités multidisciplinares, haciéndola mucho más sencilla y mejorando la
calidad, por lo que, también se favorece la interacción entre especialistas.
Los visores digitales tienen gran número de prestaciones y permiten un rango
mucho más amplio de aumentos a los que se pueden visualizar las preparaciones, lo
que facilita la navegación. En especial el estudio a muy pequeños aumentos (<10x)
puede resultar muy útil en la evaluación de los especímenes quirúrgicos. La navegación
se ve facilitada por la presencia de una imagen del material de pequeño tamaño que
permite conocer la localización exacta del área que se visualiza en la pantalla. La
presencia de un thumbnail permite ver la imagen de la preparación convencional para
asegurar que la totalidad del material se encuentra digitalizado. La MV también
presenta varias herramientas informáticas que hacen posible rotar las imágenes,
realizar fotografías con la simple selección de un área, tomar mediciones precisas y
hacer marcas de áreas de interés o anotaciones.
Una de las ventajas principales de la MV es la visualización simultánea en la
misma pantalla de varias preparaciones, que se pueden mover y cambiar de magnitud
de forma sincrónica. Esto representa una gran mejora a la hora de comparar tinciones
inmunohistoquímicas o de localizar zonas a estudio en las distintas preparaciones.
Microscopía virtual en el diagnóstico rutinario y la docencia
[22]
El desarrollo de la MV ha representado también un gran avance en la
teleconsulta, debido a la facilidad a la hora de compartir casos y realizar consultas
diagnósticas a otros patólogos que se encuentran en centros alejados. El transporte de
las preparaciones deja de ser necesario, lo que conlleva una disminución de riesgo de
ruptura o pérdida del material, así como un ahorro económico al prescindir del servicio
de mensajería. Esto también implica una disminución drástica del tiempo de respuesta
por parte del patólogo consultor, pudiendo llegar a ser de minutos en lugar de días.
La MV permite la visualización no solo desde la pantalla de un ordenador, sino
que posibilita el uso de dispositivos portátiles como smartphones o tablets. Esta
tecnología en constante desarrollo permitirá consultar casos o realizar diagnósticos de
una forma más rápida y sencilla, sin importar la localización del patólogo [23–25].
Existen otras ventajas derivadas del uso de la MV en la rutina diagnóstica, como
la facilidad de almacenamiento y recuperación de biopsias antiguas. Esta característica
hace innecesaria la búsqueda de preparaciones histológicas convencionales en el
archivo; lo que resulta especialmente útil en las patologías crónicas, donde la revisión
de biopsias previas es muy frecuente. A esta ventaja hay que añadir el ahorro de
tiempo y la disminución de la posibilidad de pérdida o daño del material. Además, las
preparaciones digitales mantienen en el tiempo sus características, al contrario que las
convencionales cuyas tinciones pierden color y se deterioran.
Diversos estudios muestran una buena valoración de la MV por parte de los
patólogos, destacando la calidad de imagen, la presencia de herramientas
informáticas, así como la ergonomía de los puestos de trabajo [22]. A pesar de lo cual,
aún existen reticencias al uso de la MV en el diagnóstico de rutina por parte de algunos
facultativos.
La MV también ha permitido la creación de algoritmos diagnósticos para la
cuantificación automática de células positivas con tinciones inmunohistoquímicas; lo
que conlleva resultados más objetivos, al disminuir la variabilidad intra e inter-
observador. Esta característica es extremadamente útil en algunas patologías donde
pequeñas variaciones en estos resultados pueden conllevar un cambio en el
tratamiento o pronóstico del paciente [26–29]. En la actualidad se están desarrollando
Tesis Doctoral. Adela Saco Álvarez
[23]
nuevas herramientas que permiten el reconocimiento de patrones histológicos que
remarcan de forma automática áreas sugestivas de patologías concretas o de
infiltración del estroma.
El flujo de trabajo del personal técnico también se ve modificado con la
introducción de la MV, y aunque la carga y descarga de preparaciones en el escáner
representa un incremento del trabajo técnico, otras tareas se ven reducidas. El reparto
de preparaciones por las distintas subespecialidades deja de ser necesario y la
búsqueda de preparaciones en el archivo se ve reducida debido a que el visor permite
la visualización de preparaciones antiguas.
El archivo de preparaciones de cristal también presenta grandes inconvenientes.
El gran espacio físico necesario y las horas de trabajo del personal técnico destinadas a
esta tarea suponen un gran gasto económico. La MV podría representar una
importante mejora si no fuese necesario el almacenamiento de algunas preparaciones
convencionales al existir la imagen virtual.
2.2. Ventajas de la microscopía virtual en la docencia
La aplicación de la MV presenta grandes ventajas, tanto en la docencia de pre
como de post-grado. En lo que respecta a la docencia de pre-grado, la MV ha supuesto
un gran avance a la hora de realizar prácticas de microscopía en asignaturas como
Histología o Anatomía Patológica. Una de las principales ventajas es que la MV permite
la visualización desde cualquier ordenador, desapareciendo la necesidad de
microscopios convencionales. El hecho de no necesitar aulas de MC hace que se
reduzca el gasto económico dirigido a su mantenimiento. A su vez, también mejora de
forma indirecta la calidad del material docente, pues normalmente los microscopios
ópticos destinados a docencia suelen tener una menor calidad [30,31]. Además, las
imágenes virtuales se encuentran siempre enfocadas y con la iluminación óptima, lo
que contribuye también a mejorar de la calidad del material docente.
A la hora de evaluar el conocimiento adquirido por los estudiantes la MV
también supone un avance, pues el hecho de no usar imágenes estáticas hace que los
Microscopía virtual en el diagnóstico rutinario y la docencia
[24]
alumnos reconozcan las características histológicas y no las de la imagen (forma,
tamaño del tejido, etc), evaluando realmente el conocimiento adquirido [32].
Entre las ventajas de la MV cabe destacar que no es necesario tener
conocimientos previos en el uso del microscopio convencional, lo que resulta
especialmente beneficioso, pues los estudiantes pueden centrarse en las imágenes
histológicas y no en el uso de la MC. Así mismo, la aceptación de la MV es muy buena,
lo cual es debido a que los estudiantes tienen una amplia experiencia en el uso de
dispositivos informáticos.
La navegación por la preparación digital resulta sencilla pues dispone de un
thumbnail y presenta un mayor rango de magnitudes que permite pequeños
aumentos; estas características facilitan la orientación dentro del tejido, lo que es de
especial importancia cuando no se dispone de experiencia previa. La visualización de
varias preparaciones de forma simultánea en una misma pantalla es posible, lo que
facilita la comparación e interpretación de distintas tinciones como por ejemplo de las
técnicas inmunohistoquímicas. La MV presenta también herramientas informáticas
que permite realizar marcas y anotaciones, éstas pueden ser creadas por el profesor
para indicar puntos de interés, o bien por el alumno para señalar dudas; lo cual hace
que el aprendizaje sea mucho más dirigido. Algunos estudios muestran mejores
resultados cuando los docentes realizan marcas con anotaciones sobre puntos clave
para el diagnóstico, respecto a preparaciones sin marcar [33–35].
Los programas para visualizar las preparaciones digitales destinados a la
docencia permiten completar la imagen histológica con datos de la historia clínica,
pruebas de imagen, fotografías macroscópicas o técnicas adicionales como las
tinciones inmunohistoquímicas, FISH o inmunofluorescencia [2]. Esto ofrece una visión
más amplia y completa de los casos, con el fin de acercarse en la mayor medida de lo
posible a la práctica clínica habitual.
La posibilidad de visualización simultanea por parte de varias personas de las
mismas imágenes también ayuda a homogeneizar el aprendizaje de los alumnos. El
hecho de que los alumnos puedan ver la misma imagen a la vez mejora la cooperación
entre ellos y con el profesor, contribuyendo así al proceso de aprendizaje [31,36,37].
Tesis Doctoral. Adela Saco Álvarez
[25]
Con la MC era necesaria la creación de nuevas preparaciones convencionales, las
cuales tenían que ser reemplazadas cada cierto tiempo debido al deterioro por su uso.
Las imágenes generadas con la MV conservan siempre la misma calidad, haciendo
innecesario nuevos cortes histológicos adicionales con la subsecuente pérdida de
material. Gracias a esta característica, es posible incluir dentro del material docente
citologías y biopsias de pequeño tamaño, cuyo estudio se encontraba limitado por el
temor a perder material o a que éste fuese requerido para ampliar las pruebas
diagnósticas del paciente [38]. Lo mismo ocurre con los casos a consulta, en los que se
puede devolver el material una vez digitalizado con fines docentes. También cabe
remarcar el ahorro de tiempo del personal técnico a la hora de archivar y realizar
cortes histológicos destinados a la docencia.
Otra ventaja muy valorada por el alumnado es la posibilidad de acceso a las
preparaciones desde cualquier ordenador y en cualquier momento, facilitando el
estudio y eliminando las restricciones del uso de laboratorios de microscopía después
de las horas lectivas [33–35,39,40].
Por último, la MV permite la confección de series de casos de forma fácil y sin
gastar material de las biopsias. Este material docente resulta muy beneficioso
especialmente en estudios de post-grado y para la formación continuada de los
especialistas en Anatomía Patológica. Los patólogos en formación pueden centrarse en
reconocer patrones histológicos importantes para realizar los distintos diagnósticos. A
su vez, es posible compartir casos de patologías poco frecuentes para ayudar a
homogeneizar aún más el aprendizaje entre los distintos centros.
3. Inconvenientes de la microscopía virtual
3.1. Inconvenientes de la microscopía virtual en el diagnóstico de rutina
La principal limitación de la MV es la inversión económica que conlleva, tanto en
su implementación en un Servicio de Anatomía Patológica como en su posterior
mantenimiento. Aunque en los últimos años los precios de los escáneres han
disminuido, todavía representan un importante gasto económico, a lo que hay que
sumar la creación de puestos de trabajo que contengan un ordenador con un soporte
Microscopía virtual en el diagnóstico rutinario y la docencia
[26]
adecuado para esta tecnología y una pantalla de alta resolución que permita su
correcta visualización [41,42].
El personal técnico es una pieza indispensable en el proceso de digitalización, por
lo que resulta necesario que al menos dos personas tengan conocimientos en MV y
dediquen parte de su tiempo a la carga y descarga de preparaciones en el escáner, así
como a la resolución de incidencias que puedan presentarse durante el proceso. Es
necesario tener en cuenta que el incremento del trabajo del personal técnico se ve en
parte contrarrestado por ciertos beneficios comentados anteriormente como el ahorro
de tiempo a la hora de repartir preparaciones entre las diferentes áreas o en la
búsqueda de preparaciones antiguas en el archivo.
Una vez instaurado todo el sistema, el principal inconveniente es la necesidad de
espacio virtual para el almacenamiento de los archivos. Las imágenes generadas por el
escáner presentan una gran calidad, lo que se traduce en un gran tamaño de archivo,
frecuentemente más de 2 GB por preparación. Son necesarios servidores de gran
capacidad, así como estrategias de reducción de tamaño de los archivos, como la
compresión de archivos o el hecho de escanear con el objetivo de menor magnitud
que permita un correcto diagnóstico [41].
Otro problema añadido es la reticencia de algunos patólogos a abandonar el uso
del MC, aunque las ventajas de la MV son numerosas y cada vez mejor conocidas. La
mayoría de las principales quejas giran en torno al aumento de tiempo empleado en la
visualización de las preparaciones virtuales, respecto a las convencionales. Algunos
estudios han puesto de manifiesto que existe una curva de aprendizaje en el uso de la
MV, mejorando sustancialmente el tiempo cuando el patólogo se familiariza con esta
tecnología [43–48].
3.2. Inconvenientes de la microscopía virtual en la docencia
Nuevamente la mayor desventaja es el coste económico, tanto de implantación
de la MV como de mantenimiento. Una posible solución es el uso de escáneres que ya
estén siendo usados para el diagnóstico rutinario y el uso de un programa informático
de bajo coste [33]. El almacenamiento de archivos también representa una alta
Tesis Doctoral. Adela Saco Álvarez
[27]
inversión económica por lo que es recomendable escanear las preparaciones con un
objetivo máximo de 200x y existe la posibilidad de alquilar terabytes en un servidor
externo.
A nivel de material docente se precisa un aula con ordenadores que dispongan
de una conexión a internet de alta velocidad, y a pesar de que es necesario realizar
mantenimiento, resulta mucho más versátil y económico que un aula con microscopios
ópticos [30,31].
Otra de las principales críticas de la MV en la docencia es que los estudiantes no
aprenden el uso de los microscopios ópticos. Sin embargo, la MV les permite centrarse
en la histología y el reconocimiento de patrones histológicos y no en el uso de la
herramienta, lo que lleva a un mejor aprovechamiento del tiempo dedicado al estudio
[31,49,50].
La interacción entre alumnos y profesor puede verse afectada por el uso de la
MV, ya que la posibilidad de visualización a distancia puede disminuir el contacto entre
ellos. También es necesario realizar controles de calidad periódicos para evaluar el uso
del sistema fuera de horario docente y valorar que imágenes y áreas son las más
visualizadas por los alumnos, con el objetivo de reunir la mayor información posible
sobre el proceso de adquisición de conocimientos por parte de los alumnos [33].
4. Microscopía virtual en el diagnóstico de rutina
4.1. Integración del sistema de microscopía virtual en un servicio de Anatomía
Patológica
La MV requiere la transformación de las preparaciones de cristal convencionales
en imágenes virtuales. En el momento actual existen en el mercado varios escáneres
capaces de realizar este proceso de forma muy eficiente sobre gran número de
preparaciones, permitiendo así procesar toda la actividad de grandes servicios de
Anatomía Patológica [17,18,20,22,51]. El proceso de digitalización de preparaciones
convencionales se realiza de forma automática, incluyendo la selección del área de la
preparación que contiene el tejido, la distribución de los puntos de enfoque y la
Microscopía virtual en el diagnóstico rutinario y la docencia
[28]
calibración. Cuando una preparación contiene secciones seriadas, el sistema escanea
todos los cortes presentes en el cristal. Las preparaciones deben mantener unos
mínimos de calidad en el montaje para poder ser escaneadas correctamente, lo que
también contribuye al mantenimiento del control de calidad.
Para facilitar el diagnóstico, las imágenes virtuales se encuentran ligadas a cada
uno de los pacientes. Esto sucede porque el sistema LIS de los servicios de Anatomía
Patológica se encuentra vinculado al escáner, generando un código QR que el escáner
liga con la información referente a cada caso. Este mecanismo permite ver los datos
del paciente desde el visor, facilitando su reconocimiento y limitando la necesidad de
buscar información adicional en el sistema LIS. A su vez, algunos programas
informáticos encargados del LIS disponen de un acceso directo al visor de MV,
permitiendo la visualización de una preparación virtual de forma directa.
Las preparaciones pueden ser escaneadas a distintas magnitudes; la mayor parte
de los escáneres disponen de objetivos de 20x, 40x y en algunos casos 60x. Esto
permite crear imágenes de gran calidad que pueden ser visualizadas posteriormente a
200x, 400x o 600x sin perder definición. Los programas informáticos empleados como
visores permiten realizar un zoom digital añadido a estas magnitudes, permitiendo una
visualización de hasta 400x cuando la preparación ha sido escaneada a 20x y hasta
600x cuando lo ha sido a 40x.
Los visores de las preparaciones digitales son programas especialmente
diseñados para esta función, imitando en gran medida a los microscopios ópticos
convencionales. Permiten la visualización y magnificación a tiempo real de las
preparaciones con un número de aumentos mucho más amplios que la MC. Las
imágenes se encuentran siempre enfocadas, y presentan un contraste y una
iluminación óptima en todo momento. La mayor parte de estos visores disponen de un
thumbnail de la preparación que permite comprobar que la totalidad del material ha
sido correctamente digitalizado y también presentan una imagen a pequeño aumento
para facilitar la navegación por la preparación digital. Así mismo, incorporan múltiples
herramientas digitales capaces de realizar marcas y anotaciones sobre las imágenes,
tomar medidas precisas, fotografías o incluso realizar cuantificaciones automáticas.
Tesis Doctoral. Adela Saco Álvarez
[29]
Todas estas imágenes son almacenadas en servidores que permiten el acceso a
los casos desde el visor, incluso después de ser validados. Debido al gran tamaño de
los archivos en muchos centros existen dos sistemas de almacenaje: un servidor se
encarga de los casos recientes permitiendo un acceso rápido y otro almacena los casos
más antiguos, con un acceso a las imágenes que tarda algunos segundos más que el
anterior. De esta forma se optimiza el espacio de disco destinado al almacenamiento
de archivos.
Estos escáneres suelen disponer de un módulo de auditoría que registra el
tamaño de cada archivo, el tiempo empleado por el escáner para digitalizar cada caso
en particular y cada uno de los accesos a las imágenes virtuales.
4.2. Validación de la microscopía virtual en el diagnóstico rutinario
El uso de esta nueva tecnología para el diagnóstico primario fue cuestionado por
muchos especialistas, principalmente por las escasas evidencias científicas sobre su
fiabilidad diagnóstica existentes al inicio de su uso. Aunque cada vez existen más
publicaciones demostrando una buena reproducibilidad entre los diagnósticos
realizados con MV y MC, su número sigue sin ser aún elevado y existen
subespecialidades en las que no existen estudios o éstos presentan deficiencias en su
diseño.
Con el objetivo de aclarar estas cuestiones y de normativizar la implementación
de la MV, la American Telemedicine Association, el College of American Pathologist y la
Canadian Association of Pathologist fueron los primeros en realizar una revisión de la
bibliografía existente en ese primer momento y elaboraron la primera guía de
recomendaciones para la realización de una correcta validación de la microscopía
virtual [3,17]. En esta revisión, realizada en 2013, se analizaban los resultados de 767
publicaciones, de las cuales solamente 112 prestaban una metodología adecuada. El
nivel de concordancia entre diagnósticos era muy bueno, variando entre el 73% y el
98% en los distintos estudios; por lo que se concluía que la MV es una tecnología
adecuada para realizar el diagnóstico rutinario. Las discrepancias mayores se
encontraban entre el 3% y el 7%, cifras bastante dependientes de las variaciones en la
metodología de los estudios, por ejemplo si la concordancia entre diagnósticos era
Microscopía virtual en el diagnóstico rutinario y la docencia
[30]
evaluada intra o inter-observador. Estos resultados reflejaban la necesidad de unas
recomendaciones claras sobre el proceso de validación; motivo por el cual crearon las
primeras guías sobre la validación de la MV en el diagnóstico primario. En ellas
recomiendan la realización de una validación interna en cada centro en el que se
introduzca la MV. Este proceso tiene que englobar una variedad de muestras que
refleje la complejidad de la práctica real del centro (congelados, preparaciones teñidas
con H&E e inmunohistoquímica, citologías, etc). Se considera innecesario incluir una
representación de todos los órganos y subespecialidades, pues los resultados de una
subespecialidad pueden ser usados en otra con características similares. Respecto al
número de casos necesarios, la guía sugiere 60 especímenes de cada tipo de muestras
para alcanzar una precisión cercana al 90% y una concordancia del 95%; en los
estudios analizados un incremento en el número de casos no representaba un
aumento sustancial de la precisión en los resultados.
No es necesario realizar la validación de cada uno de los componentes del
sistema de digitalización, sino que se realiza en conjunto. De la misma forma, cuando
se modifica algún componente del sistema, éste tiene que ser reevaluado en su
conjunto.
La forma más adecuada de hallar la concordancia entre los diagnósticos es con el
cálculo de la reproducibilidad intra-observador y se recomienda un periodo mínimo de
descanso entre ambas visualizaciones de dos semanas. El orden de la evaluación con
MV o MC no influye en el resultado final, por lo que puede ser aleatorio.
Dentro de los aspectos técnicos, resulta necesario comprobar que la totalidad
del material de las preparaciones haya sido escaneado; esta tarea se ve facilitada por
la incorporación en muchos de los visores digitales de una imagen en miniatura de la
preparación convencional. A su vez, se debería comprobar que las imágenes generadas
por el escáner son iguales a las recibidas por el especialista, especialmente cuando se
usan sistemas de compresión para reducir el tamaño de éstas, pues la calidad de la
imagen final puede verse comprometida, especialmente cuando se emplean
compresiones del tipo irreversible.
Tesis Doctoral. Adela Saco Álvarez
[31]
Se recomienda que el proceso de validación incluya a todo el personal que
posteriormente se verá involucrado en la digitalización rutinaria, siendo aconsejable
contar con la asistencia de un patólogo con experiencia en el uso de la MV. La
validación se debe llevar a cabo siguiendo los estándares más actualizados en cada
momento y se recomienda generar un documento que recoja todo el proceso por si
resulta necesario efectuar comprobaciones posteriormente.
El Libro Blanco de la Anatomía Patológica en España publicado en 2015 recoge
las recomendaciones de la Sociedad Española de Anatomía Patológica sobre la
validación de la MV. Éstas coinciden con las aconsejadas por las guías de la American
Telemedicine Association, el College of American Pathologists y la Canadian
Association of Pathologists aunque se propone un proceso de validación menos
estricto debido a la gran necesidad de tiempo y recursos requeridos para su
realización, siendo poco compatible con la práctica diaria. Puesto que cada vez existen
más evidencias científicas sobre la buena correlación entre los diagnósticos con MV y
MC, se pueden realizar validaciones menos estrictas o incluso adoptar un sistema
digital validado con anterioridad en otro centro [52]. Sin embargo, es necesario tener
en cuenta que la literatura publicada sobre la validación en el diagnóstico presenta
deficiencias en algunos ámbitos de la patología, por lo que sigue siendo necesario
completar estos estudios para facilitar la implementación de este sistema en los
servicios de Anatomía Patológica.
Otra característica a destacar es la presencia de una curva de aprendizaje en el
uso de la MV, por lo que al inicio todos los patólogos pasan por un periodo de
adaptación en el que resulta necesario compaginar ambas tecnologías para realizar el
diagnóstico, hasta adquirir una mayor seguridad con el uso de la MV.
4.3. Estado la de validación en el momento actual
En los últimos años han aparecido numerosos estudios demostrando muy buena
correlación entre los diagnósticos realizados con MV y MC. Estos buenos resultados
sumados a las numerosas ventajas de la MV, hacen de esta tecnología una excelente
candidata para el diagnóstico de rutina. Sin embargo, cuando se dividen por
subespecialidades aparecen deficiencias en muchos ámbitos de la Anatomía
Microscopía virtual en el diagnóstico rutinario y la docencia
[32]
Patológica, pues existen áreas que no disponen de estudios de validación o éstos
tienen una metodología inadecuada. Antes de utilizar la MV para el diagnóstico
primario de un tipo de biopsias determinadas es fundamental asegurar la existencia de
estudios de validación que engloben dicho tipo de muestras u otras con características
similares. Por esta razón resulta necesario identificar cuáles son las áreas de la
patología en las que aún no existe suficiente evidencia científica sobre el uso de la MV,
para llevar a cabo la realización de nuevos estudios de validación con el objetivo de
que esta tecnología pueda ser empleada en el diagnóstico de rutina.
4.4. Aspectos técnicos de la digitalización de rutina
Tras la instauración de la MV para el diagnóstico rutinario en un Servicio de
Anatomía Patológica es necesario tener en cuenta ciertas consideraciones técnicas que
pueden influir en el buen funcionamiento del sistema.
El uso de la MV en la práctica clínica requiere una digitalización eficiente de las
preparaciones, lo que se traduce en reducir el tiempo desde que las preparaciones
convencionales son montadas hasta que se digitalizan, para ello es necesario
escanearlas de forma frecuente, preferiblemente en más de una ocasión durante la
jornada laboral. Para que el proceso diagnóstico se ralentice lo menos posible es
recomendable que los centros dispongan de más de un escáner, pues al poder dividir
las preparaciones el tiempo se reduce sustancialmente. Además, el segundo escáner
garantiza la continuidad de la MV durante las eventuales averías que puedan surgir;
evitando cambiar el hábito diagnóstico de los patólogos que usan esta tecnología.
También es necesario tener en consideración que existen ciertos casos en los
que, debido a las características de la muestra o al estado del paciente, es
imprescindible realizar un diagnóstico lo más rápido posible, por lo que es conveniente
la creación de estrategias para priorizar biopsias urgentes.
Otra de las cuestiones que han generado controversia desde el comienzo de la
MV es la magnitud a la que se deben de escanear las preparaciones. La imagen
resultante tiene que permitir un correcto diagnóstico, minimizando al máximo el
tamaño del archivo generado para facilitar su almacenamiento. Varios estudios de
Tesis Doctoral. Adela Saco Álvarez
[33]
validación de diferentes subespecialidades señalan al objetivo de 20x como adecuado
para el diagnóstico, aunque existen ciertas excepciones [22,41]. Algunas
subespecialidades como la patología renal, biopsias pequeñas cardiacas o hepáticas,
así como ciertas biopsias dermatológicas necesitan imágenes de una mayor calidad,
pues pequeñas variaciones pueden condicionar un cambio en el tratamiento o
pronóstico del paciente; por lo que, aunque todavía existen escasas publicaciones, la
recomendación más extendida es escanear este tipo de biopsias con un objetivo
mínimo de 40x.
La MV también involucra en gran medida al personal técnico, que ve modificado
su flujo de trabajo. La digitalización rutinaria de preparaciones supone una nueva tarea
de carga y descarga del escáner; así como de comprobación de la correcta
digitalización de las preparaciones, que en ocasiones requiere la modificación o re-
escaneando de aquellas que han presentado algún incidente o problema de enfoque
durante el proceso [3]. Por otro lado, ciertas tareas clásicamente atribuidas al personal
técnico desaparecen o disminuyen en gran medida, como el reparto de preparaciones
por las distintas subespecialidades o la búsqueda de preparaciones antiguas en el
archivo, pues el visor virtual permite la visualización de casos previos.
El diagnóstico digital de preparaciones de material congelado de biopsias
peroperatorias, no solo es posible, sino que presenta una muy buena concordancia con
el diagnóstico con MC según múltiples estudios, con un tiempo total de respuesta que
varía entre 14 y 20 minutos [7,42,53]. El personal técnico y facultativo deben de tener
experiencia en el uso de la MV, con el objetivo de reducir este tiempo al máximo. Una
ventaja de esta tecnología a tener en cuenta en el diagnóstico de biopsias
peroperatorias es la posibilidad de visualización remota, permitiendo el diagnóstico
por parte de un patólogo especialista, aunque éste no se encuentre en el centro en ese
momento.
La MV presenta algunos inconvenientes técnicos, pues generalmente no permite
ajustar el enfoque a distintos planos de la preparación, lo que dificulta su uso para el
diagnóstico en el ámbito de la citología. Algunos escáneres presentan la posibilidad de
escanear usando un plano adicional (plano z) que simula este enfoque en profundidad,
Microscopía virtual en el diagnóstico rutinario y la docencia
[34]
pero conlleva un significativo incremento del tamaño de los archivos, lo que resulta en
un gran inconveniente pues dificulta su almacenamiento.
Las preparaciones convencionales deben presentar unas rigurosas condiciones,
tanto de tinción como de montaje, con el fin de asegurar una correcta digitalización;
motivo por el cual resulta necesario realizar un control de calidad periódico sobre ellas.
Así mismo, todo el sistema de MV también debe de ser sometido a un proceso
de control de calidad para asegurar su correcto funcionamiento. Las principales guías
recogen una serie de recomendaciones muy detallas sobre cómo realizar este proceso
de la forma más adecuada dentro de cada centro [3]. Entre ellas destacan la creación
de un comité específico formado por personal involucrado en el proceso de
digitalización (facultativos, personal técnico y dirección del servicio), con la función de
supervisar este proceso. Una de las funciones de este comité es realizar una
comprobación sistemática de la política sobre el uso de la MV en busca de posibles
actualizaciones. También es necesario crear una guía informativa sobre el uso del
programa informático y el hardware que debe de estar disponible para los usuarios en
todo momento. Así mismo, resulta necesaria la creación de un mecanismo para la
detección y resolución de problemas, con una respuesta lo más rápida posible. Se
recomienda la documentación de parámetros de importancia como el tiempo y la
concordancia en el diagnóstico de biopsias peroperatorias, el porcentaje de casos que
requieren revisión de la preparación convencional durante el diagnóstico y el
porcentaje de re-escaneos. Por último, proponen la revisión de un 10% de los casos
diagnosticados de forma virtual, tanto propios del centro como de consulta
diagnóstica.
Al igual que sucede en la validación de la MV para el diagnóstico, estas
recomendaciones resultan muy estrictas y su completa realización representaría un
gran consumo de tiempo y recursos por parte del personal y del centro. Algunas de
ellas, como la revisión de un 10% de los casos también están recomendadas en el uso
de la MC, pero en la práctica real su aplicación es muy baja [3,52].
Tesis Doctoral. Adela Saco Álvarez
[35]
4.5. Almacenamiento de archivos
Uno de los principales desafíos al que se enfrenta el diagnóstico rutinario con MV
es el gran tamaño de las imágenes, el cual genera una gran necesidad de espacio de
disco para almacenar preparaciones digitales. Cuando este archivo se suma al de
preparaciones convencionales, el gasto económico aumenta significativamente,
representando uno de los principales inconvenientes del uso de la MV en el
diagnóstico rutinario.
Se han propuesto múltiples estrategias con el objetivo de minimizar este gasto,
como la compresión de imágenes, el archivo exclusivo de las preparaciones virtuales
más representativas de cada caso, la eliminación de imágenes o incluso de
preparaciones convencionales tras un tiempo después del diagnóstico; teniendo en
cuenta que en todos los casos los bloques de parafina serían conservados.
En el momento actual no existen apenas estudios que traten este tema, ni
tampoco un marco legal que defina cuál es el tratamiento más correcto para estos
archivos; aunque algunos estudios sugieren un tiempo mínimo de conservación de las
imágenes de 6 meses [22].
5. Microscopía virtual en la docencia
En diversos ámbitos académicos como Medicina, Odontología o Veterinaria el
estudio de la Histología y la Anatomía Patológica tienen una gran relevancia.
Clásicamente la única alternativa para llevarlo a cabo era la utilización de MC y
posteriormente el uso de imágenes estáticas como fotografías de las muestras
histológicas. Este panorama cambió con la aparición de la MV, debido a que ciertas
características intrínsecas a esta tecnología representan numerosas ventajas al ser
aplicadas a la docencia. Algunas de las principales son la posibilidad de visualización de
una preparación por parte de un amplio grupo de personas, el acceso remoto en
cualquier momento del día, la facilidad en su uso y la mejora en la uniformidad del
contenido, pues el material docente es el mismo para todos los alumnos.
Microscopía virtual en el diagnóstico rutinario y la docencia
[36]
Existen numerosas publicaciones que muestran buenos resultados con el uso de
la MV; así como una valoración muy positiva tanto por parte de los docentes como del
alumnado [30,40,54–59].
A pesar de que estas características convierten la MV en una excelente
herramienta para el aprendizaje de la Anatomía Patológica, al igual que ocurre en el
diagnóstico rutinario, resulta necesario confirmar la no inferioridad de esta
herramienta respecto a la MC para asegurar una correcta formación de los alumnos
tanto de pre como de post-grado.
6. Microscopía virtual en comités multidisciplinares
La incorporación de la MV a los servicios de Anatomía Patológica ha
representado un cambio a la hora de la realización de comités multidisciplinares, pues
permite la visualización a tiempo real de las preparaciones histológicas requiriendo
únicamente un ordenador y una pantalla. De esta forma, se disminuyen drásticamente
las barreras entre las distintas especialidades, facilitando la interacción entre
patólogos y clínicos [22,42]. Además, permite un estudio más completo de los
pacientes al incluir las imágenes histológicas a la hora de presentar los casos, al igual
que ocurrió hace algunos años con la incorporación de las imágenes diagnósticas.
Otra ventaja es la posibilidad de participar en comités de otros centros o en
sesiones conjuntas a distancia; lo permite un mayor flujo de información, dado que
posibilita el compartir casos interesantes o realizar consultas a otros especialistas.
Existen estudios cuyos resultados revelan un impacto positivo de la MV a la hora
de preparar los casos para los comités, pues el tiempo empleado por el patólogo
disminuye aproximadamente al 50% (entre 30 minutos y 1 hora semanal de ahorro).
Esto es debido a que las preparaciones se pueden visualizar en tiempo real, por lo que
no resulta necesario realizar fotografías o movilizar material; con el fin de agilizar aún
más la visualización se pueden hacer anotaciones sobre las zonas de interés
diagnóstico [22].
Tesis Doctoral. Adela Saco Álvarez
[37]
7. Microscopía virtual en la teleconsulta
Una de las primeras aplicaciones de la MV fue en la teleconsulta, lo que
representó un gran avance, dado que hasta su aparición estaba basada únicamente en
imágenes estáticas que no permitía la navegación ni el uso de distintas magnitudes
[1,53,60–65]. La MV permite la consulta de casos con dificultad diagnóstica entre
distintos hospitales cuando éstos se encuentran en localizaciones remotas, áreas
rurales o cuando los expertos no están en el centro [14,66]. Además, con la reciente
expansión de la MV existe la posibilidad de establecer redes diagnósticas entre los
hospitales regionales y los centros de referencia, lo que se traduce en una significativa
mejora en el proceso diagnóstico. Estas redes permiten que grupos de expertos
realicen el diagnóstico en casos complicados, incrementando así de forma significativa
la posibilidad de alcanzar un correcto diagnóstico, lo cual es una gran mejora para los
pacientes pues se alcanza la igualdad entre ellos sin importar su localización. Otra
ventaja es que este diagnóstico experto no supone un incremento significativo del
tiempo de respuesta, pues no resulta necesaria la movilización de material, y la
comunicación entre profesionales es mucho más dinámica al poder visualizar los casos
a tiempo real. Algunas publicaciones ponen de manifiesto la gran disminución de este
tiempo de respuesta en las consultas, llegando a ser de hasta 18 minutos en algunos
estudios [66].
Otro beneficio secundario es la disminución de la necesidad de movilizar bloques
de parafina o preparaciones histológicas, evitando así la posible pérdida o daño del
material; a lo que hay que añadir un ahorro del coste de mensajería [14,65–67].
Microscopía virtual en el diagnóstico rutinario y la docencia
[38]
Tesis Doctoral. Adela Saco Álvarez
[39]
II. Hipótesis
Microscopía virtual en el diagnóstico rutinario y la docencia
[40]
Tesis Doctoral. Adela Saco Álvarez
[41]
En los últimos años hemos asistido a una creciente expansión la MV debido sus
claras de ventajas respecto a la MC, tanto en el diagnóstico primario como en la
docencia. Esta tecnología resuelve una gran parte de los inconvenientes intrínsecos al
uso de la MC. Sin embargo, esta tecnología está generando algunas dudas y reticencias
por parte de los patólogos. La mayoría de éstas surgen de la novedad de la
herramienta y de la relativa inseguridad de poder alcanzar un diagnóstico fiable con
ella. Las escasas guías publicadas sobre su uso por el College of American Pathologists,
la Canadian Association of Pathologists y la American Telemedicine Association
proponen la realización de una validación interna en cada uno de los centros en los
que se pretenda implantar esta herramienta. Esta validación podría ser una medida
innecesaria si existiese suficiente evidencia científica de su no inferioridad respecto a
la MC. Sin embargo, existen numerosas subespecialidades en las que las evidencias
existentes sobre su validación son escasas o incluso nulas, por lo que se requieren
estudios de validación adicionales.
Nuestra hipótesis en el estudio número 1 es que existe suficiente evidencia
científica sobre la no inferioridad de la MV con respecto a la MC para el diagnóstico
primario en numerosas áreas, pero que es preciso identificar las áreas de la patología
en las que los estudios de validación sean escasos o inexistentes, en los cuales resulte
necesaria la realización de estudios de validación adicionales antes de su
implementación en el diagnóstico primario. En este estudio inicial se detectaron
algunas áreas en las que la evidencia era escasa o nula. Por ello hemos llevado a cabo
los estudios número 2 y número 3 con el objeto de validar respectivamente la MV
para el diagnóstico primario de biopsias ginecológicas y biopsias hepáticas con aguja.
En ambos casos la principal hipótesis es que el diagnóstico histológico rutinario de
estas biopsias con MV no es inferior al realizado con MC. También planteamos como
hipótesis adicional la existencia de una curva de aprendizaje al inicio del uso de la MV,
pero que rápidamente tanto el diagnóstico como el tiempo empleado por el/la
patólogo/a se equipara con la MC.
Por otro lado, son más numerosas las evidencias de las ventajas que la MV
aporta a la docencia tanto de pre como de post-grado. Sin embargo, resulta también
fundamental la validación de la MV en la docencia. La comparación entre los
Microscopía virtual en el diagnóstico rutinario y la docencia
[42]
resultados de los exámenes tras la realización de las prácticas usando ambas
herramientas es uno de los parámetros más útiles a la hora de asegurar que la MV y la
MC son equiparables. Otro parámetro a tener en cuenta es la valoración por parte del
alumnado y de los docentes; lo que también puede ayudar a conocer cuáles son las
necesidades y la forma de llevar a cabo el proceso educativo en la actualidad. Sobre
estas premisas se realizó el estudio número 4 con la hipótesis de que, al igual que en el
diagnóstico primario, los estudios sobre el uso de la MV en la docencia permiten
confirmar su utilidad para la docencia, tanto en alumnos de pre-grado como en la
formación post-grado. Las conclusiones de este estudio encauzaron la implementación
de la MV en la enseñanza de la Anatomía Patológica en los alumnos de pregrado de la
Facultad de Medicina de la Universidad de Barcelona. El estudio número 5 tiene como
hipótesis principal que los resultados del uso de la MV en docencia no son
equiparables a los alcanzados con la MC y que la valoración por parte de los alumnos
de la MV es positiva.
Tesis Doctoral. Adela Saco Álvarez
[43]
III. Objetivos
Microscopía virtual en el diagnóstico rutinario y la docencia
[44]
Tesis Doctoral. Adela Saco Álvarez
[45]
El objetivo general de la presente tesis es estudiar el uso de la MV en el diagnóstico
rutinario de biopsias en un Servicio de Anatomía Patológica, así como, su uso en la
docencia; con el fin de determinar la utilidad de esta tecnología en ambos ámbitos,
que permita planificar de forma adecuada su implementación en ambos campos.
De forma particular se han planteado los siguientes objetivos específicos:
1. Evaluar en los artículos publicados que la MV y la MC presentan resultados
equiparables en el diagnóstico primario de las distintas subespecialidades de la
Anatomía Patológica (estudio 1)
2. Determinar si existen suficientes estudios de validación que engloben la
totalidad de las diferentes subespecialidades de la Anatomía Patológica, en
especial aquellas con características diferentes (estudio 1)
3. Evaluar la concordancia inter-observador entre los diagnósticos realizados con
MV y con MC en biopsias ginecológicas de rutina, entre dos especialistas con
experiencia en el área (estudio 2)
4. Determinar si existe una curva de aprendizaje en el uso de la MV y sus
características (estudio 2)
5. Establecer cuál es el aumento de escaneo con una mejor relación coste-beneficio
para realizar un diagnóstico adecuado en biopsias ginecológicas (estudio 2)
6. Evaluar la concordancia inter e intra-observador entre los diagnósticos realizados
con MV y con MC en biopsias hepáticas con aguja, tanto provenientes de hígados
nativos como de trasplante (estudio 3)
7. Establecer la estrategia de escaneo más adecuada para alcanzar un diagnóstico
correcto en estas biopsias (estudio 3)
8. Determinar si existe suficiente evidencia sobre la adecuación de la MV en la
docencia de Anatomía Patológica, tanto de pre como de postgrado (estudio 4)
Microscopía virtual en el diagnóstico rutinario y la docencia
[46]
9. Determinar si el paso de la MC a la MV en la asignatura de Anatomía Patológica
tiene impacto en los resultados de los estudiantes de Medicina de la Universidad
de Barcelona (estudio 5)
10. Analizar las impresiones de los estudiantes sobre el uso de la MV y valorar de
forma objetiva cómo influye esta herramienta en el proceso de aprendizaje
(estudio 5)
Tesis Doctoral. Adela Saco Álvarez
[47]
IV. Trabajos
realizados, métodos
y resultados
Microscopía virtual en el diagnóstico rutinario y la docencia
[48]
Tesis Doctoral. Adela Saco Álvarez
[49]
La descripción de las muestras, así como la metodología utilizada en los trabajos
realizados, se encuentran detalladamente descritas en las secciones de “Material y Métodos”
de cada uno de los artículos que constituyen el cuerpo doctrinal de la presente Tesis Doctoral.
Dichos artículos se incluyen a continuación tal y como han sido publicados en la
literatura científica.
Microscopía virtual en el diagnóstico rutinario y la docencia
[50]
Tesis Doctoral. Adela Saco Álvarez
[51]
Estudio número 1
“Validation of Whole-slide Imaging for Histopathological
Diagnosis: Current State”
Adela Saco, José Ramirez, Natalia Rakislova, Aurea Mira, Jaume Ordi
Pathobiology 2016; 83: 89 – 98
Factor de impacto (2016): 1.703
Ranking (2016): 60/193, segundo cuartil
Microscopía virtual en el diagnóstico rutinario y la docencia
[52]
E-Mail [email protected]
Original Paper
Pathobiology 2016;83:89–98 DOI: 10.1159/000442823
Validation of Whole-Slide Imaging for Histolopathogical Diagnosis: Current State
Adela Saco a Jose Ramírez a Natalia Rakislova a Aurea Mira a Jaume Ordi a, b
a Department of Pathology, Hospital Clínic, University of Barcelona School of Medicine, and b ISGlobal, Barcelona Center for International Health Research (CRESIB), Barcelona , Spain
annotations and measurements. WSI can be used from any device and anywhere, thereby providing great opportuni-ties for teleconsultation. New technologies such as the rec-ognition of histopathology patterns using image analysis may facilitate diagnosis and improve the reproducibility among pathologists in the future. © 2016 S. Karger AG, Basel
Introduction and Historical Perspective
For more than a century, conventional light micros-copy (CLM) has been the basic tool for tissue evaluation and has played a pivotal role in pathological diagnosis. Until the incorporation of nonmorphological molecular technologies into routine practice in recent years, the standard of diagnosis for pathologists was morphology and especially CLM-evaluated morphological criteria. In-deed, the evaluation of most specimens submitted to pa-thology laboratories today still relies on the interpretation of images by CLM, complemented by gross examination and a number of ancillary molecular techniques, mostof which [histochemistry and immunohistochemistry (IHC)] are also evaluated with CLM. Asking experts or other colleagues for diagnostic opinions required sending
Key Words
Primary diagnosis · Routine diagnosis · Validation · Virtual microscopy · Whole-slide images
Abstract
Rapid advances in informatics and technological improve-ments have led to the development of high-throughput whole-slide imaging (WSI) scanners able to produce high-quality digital images, which allow achieving a correct diag-nosis of the biopsies using virtual viewers. This technology is currently prepared to be introduced in the departments of pathology for routine diagnosis. The aim of this review is to analyze the current evidence regarding the use of WSI in pri-mary or routine diagnosis in the different subspecialties of pathology. An increasing number of studies have shown al-most perfect inter- and intraobserver agreement between the diagnoses obtained with WSI and the classical diagnoses based on conventional light microscopy. The only exception seems to be cytology, which still requires some technologi-cal development. Although validation studies are needed in some areas of pathology, growing evidence indicates that WSI is a reliable tool for routine diagnosis. Pathologists have a positive perception of the ergonomics of the workstations, the low magnification of WSI and the possibility of making
Published online: April 26, 2016
Jaume Ordi Department of Pathology, Hospital Clínic, University of Barcelona C/Villarroel 170 ES–08036 Barcelona (Spain) E-Mail jordi @ clinic.ub.es
© 2016 S. Karger AG, Basel1015–2008/16/0833–0089$39.50/0
www.karger.com/pat
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Saco/Ramírez/Rakislova/Mira/Ordi
Pathobiology 2016;83:89–98DOI: 10.1159/000442823
90
glass slides or paraffin blocks for examination by CLM. Teaching pathology to undergraduates and residents, and continuing medical education for certified pathologists also depended on the use of CLM.
This scenario slowly started to change a few decades ago [1–3] . Static digital images allowed teaching and, to a certain degree, teleconsultation, but limitations in image quality and, particularly, the inability to navigate and use different optical objectives made the substitution of CLM unfeasible [4] . Dynamic real-time telepathology systems with video cameras integrated into CLM were used for intraoperative frozen biopsies, because they allowed an image to be sent to an expert located remotely. This ca-pacity was extraordinarily useful for small hospitals, as it provided a quick diagnostic approach for difficult cases [5–12] . However, the relatively poor image quality and the impossibility to remotely conduct navigation through a slide made the system inadequate for routine diagnosis.
Rapid advances in informatics as well as technological improvements led to the development of scanners able to create digital reproductions from whole glass slides, which appeared one decade ago [1, 2] . These scanners are the basis of virtual microscopy or whole-slide imaging (WSI), which allows navigation across the virtual slide and visualization at different magnifications, allowing the computer to be used as a CLM. However, the image qual-
ity of the initial scanners was limited, and the costs of implementation of the technology, including the scanner, monitors and suitable computers, were very high, thereby restricting the use of WSI to certain areas, such as teach-ing and teleconsultation, and excluding routine diagnosis [10, 13–17] .
Currently, a number of high-throughput scanners able to produce high-quality images are available on the mar-ket. These scanners allow correct diagnosis of the biopsies using virtual viewers. The cost of implementation of WSI has significantly decreased, and the speed of visualization has notably increased [17–22] . Constant improvements in this technology have led to an important expansion in the use of WSI in routine diagnosis in recent years. The aim of this review is to evaluate the current evidence on the validation of WSI in routine diagnosis.
Advantages and Challenges of WSI for Routine
Diagnosis
Routine histopathological diagnosis can benefit from the multiple advantages of WSI. WSI workstations are more ergonomic ( fig. 1 ). WSI has a much larger field of vision than CLM and allows a wider range of magnifica-tions, thus providing easier navigation. In particular, WSI enables to study very low magnifications (<×100), which is very useful in the evaluation of surgical specimens. The computer tools allow making annotations and measure-ments. WSI viewers can simultaneously show and syn-chronously move several slides of a case, which is particu-larly helpful in the evaluation of IHC-stained slides ( fig. 2 ). Indeed, studies evaluating the opinion of patholo-gists have revealed a positive perception of image quality and stressed the utility of the measurement and annota-tion tools, as well as the ergonomics and usability of the viewer [22] . WSI can be used from any device and any-where, thereby providing great opportunities for telecon-sultation and remote work. Portability is certainly one of the major advantages of WSI, and this will probably be further improved in the near future when the current viewers are fully adapted to portable devices, such as tab-lets and smartphones [23–25] . Moreover, the need for standardization in the diagnosis and evaluation of IHC biomarkers predicting the outcome of specific therapies will probably boost the implementation of WSI.
Finally, WSIs allow for automatic quantification of IHC slides. These diagnostic algorithms facilitate quanti-fication of IHC positivity resulting in a more objective evaluation, which is extremely useful in the evaluation of
Fig. 1. WSI workstations for primary diagnosis typically include two screens, one displaying the WSI viewer and the other the labo-ratory information system and the clinical records or other clinical or imaging information. This physical structure has shown to be highly ergonomic. Additional advantages of WSI viewers are a much larger field of vision than CLM and the possibility of using a very low magnification.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Validation of WSI in Primary Diagnosis Pathobiology 2016;83:89–98DOI: 10.1159/000442823
91
some biological markers. Algorithms for the evaluation of IHC stains are variably used depending on the subspe-cialties and are particularly useful in cases of breast cancer [26–29] .
In contrast with these positive opinions, many pathol-ogists still prefer using CLM. The most criticized feature of WSI is the speed in uploading the image. Indeed, most pathologists feel that more time is required to make a di-agnosis with WSI. However, some studies have shown that although diagnosis with WSI is initially more time-consuming, this time quickly decreases as pathologists become familiar with the use of the WSI viewer [30–35] . Thus, there is a learning curve in the use of WSI and the time required for making a diagnosis, and a recent study conducted at our institution confirmed that the diagnos-tic performance improved with practice [36] . Another limitation of WSI is the relatively high costs of the equip-ment. The basic needs for a WSI system, which is ade-quate for routine diagnosis, include not only high-throughput scanners but also high-resolution monitors [37, 38] . This is a common concern since, despite the reduction in the price of the equipment in the last few years, it still represents a considerably high investment, which has a relatively low added value for many patholo-gists as the basic functions of WSI are already being con-fidently achieved with the old CLM. Finally, WSI re-quires a significant investment in high-capacity servers; the files generated by WSI scanners are huge, with sizes frequently over 2 GB per slide. Thus, strategies to reduce the size of the files, such as scanning at relatively low magnification (×200 instead of ×400 or ×600) are fre-quently used [37] .
The Need for Validation Studies
The number of studies aimed at validating WSI in pri-mary or routine diagnosis is rapidly increasing. However, whereas relatively abundant information is available in some areas, validation studies are very scant in several sub-specialties and completely absent in others. Some valida-tion studies include biopsies from several subspecialties in-stead of analyzing biopsies with similar characteristics [33, 39–43] . This relative absence of validation studies has led to reluctance in the implementation of WSI in routine clin-ical practice. Nevertheless, the number of centers imple-menting this technology is increasing due to the positive experiences reported in many departments [41, 42, 44, 45] .
Below, we review the current evidence on the valida-tion of WSI versus CLM in the different subspecialties of pathology.
Breast Pathology
WSI has been validated in the diagnosis of breast pa-thology in a number of studies conducted by different groups. Most of these studies analyzed a relatively small number of routine biopsies (between 100 and 150), in-cluding either only needle biopsies or both needle and surgical specimens [32, 46, 47] . Although scanning at ×400 was recommended in one of the studies [32] , in two of the studies a scanning magnification of ×200 was con-sidered as sufficient [46, 47] .
The intra- and interobserver agreement between CLM and WSI is excellent in all the studies, with values ranging
Fig. 2. WSI viewers may simultaneously show and synchronously move several slides of a case, which is particularly helpful in the evaluation of IHC-stained slides.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Saco/Ramírez/Rakislova/Mira/Ordi
Pathobiology 2016;83:89–98DOI: 10.1159/000442823
92
between 90 and 99%. Most of the discrepancies detected did not have clinical repercussion. Interestingly, in two of the reports, the WSI diagnosis was more frequently con-sidered as correct compared to the diagnosis performed with CLM [32, 46] . A study specifically dealing with the distinction between hyperplasia and cancer reported in-terobserver concordance in the diagnosis of 90.2%. Major discrepancies appeared in 2.3% of the cases, which, in most cases, were solved with IHC stains [48] .
A major advantage of digitization in breast pathology is the possibility to use image analysis to improve the ac-curacy and reproducibility of HER-2, estrogen and pro-gesterone receptors, and Ki67 scoring, which have a cru-cial role in the planning of treatment strategies [27–29, 49] . Moreover, the evaluation may be improved with the use of automatic quantification algorithms ( fig. 3 ).
Cytopathology
The use of WSI in cytopathology has shown someadvantages in second opinions, quality assurance, slide archiving, proficiency testing and education. However, a number of significant weaknesses of the current WSI scanners, such as the difficulties in focusing at different z-axes, are a major limitation for the introduction of this technology in routine diagnosis [50, 51] . Improvements in informatics may allow multiplane focusing using the z-axis, but they still need to be validated [21, 52, 53] .
Indeed, the current evidence of validation in cytology is almost limited to real-time dynamic digital microscopy using a video camera connected to the optical microscope
and not to WSI. The intraobserver agreement of this ap-proach with the final diagnosis is high (92%) [54] , and, in some studies, it is better than with CLM [53–55] . One study evaluating 192 liquid-based cervical cytology slides showed good intraobserver concordance (89–97%), but the interobserver concordance was better for CLM than for WSI (94 vs. 82%) [52] .
Dermatopathology
Only two studies have focused on the validation of skin biopsies evaluating routine specimens. Although both studies included a small number of cases (100 and 79, respectively), the intraobserver agreement was high (94% for WSI and 96% for CLM, respectively) [30, 56] . A study limited to tumor and tumor-like skin lesions showed agreement in the diagnosis by WSI and CLM, with a κ value of 0.93 for both methods [57] . Another study evaluated inflammatory and melanocytic lesions, with good agreement between CLM and WSI (only 1 dis-cordant diagnosis in the inflammatory biopsies and 100% concordance in the melanocytic specimens), but the number of patients included was very limited (24 cas-es). In this study, it was concluded that in most cases scanning at ×200 is sufficient to achieve a correct diag-nosis [56] .
Interestingly, WSI has shown to be suitable for tele-consultation in skin biopsies and may reduce the time of response in expert diagnosis from 5–10 days to a few hours or even minutes [57] .
Fig. 3. A major advantage of digitization in breast pathology is the possibility to use image analysis in improving the accuracy and reliability of HER-2, estrogen and pro-gesterone receptors and Ki67 scoring, which have a crucial role in the planning of treatment strategies.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Validation of WSI in Primary Diagnosis Pathobiology 2016;83:89–98DOI: 10.1159/000442823
93
Gastrointestinal Pathology
A few studies have shown that the diagnosis of gastro-intestinal biopsies using WSI or CLM provides compa-rable results [58, 59] . Two reports analyzed consecutive routine biopsies, but one was limited to gastric and co-lonic biopsies [59] . The intraobserver concordance be-tween WSI and CLM was 95% in both studies, and scan-ning at ×200 was considered as adequate. One study com-pared WSI and CLM in the evaluation of polyps in surgical specimens. Although the intra- and interobserv-er agreement was excellent for both methods in terms of diagnosis, WSI facilitated the quantification of the polyps due to the very low magnification that allows a panoram-ic view of the complete sample [60] . A study focused on Barrett’s dysplasia and neoplasia showed good diagnostic agreement between WSI and CLM, but the consensus neoplasia score was lower using WSI and the time spent in making the diagnosis was longer. These results were probably due to the lack of confidence and experience in the manipulation of the WSI viewer and seemed to im-prove with familiarity and practice [34] .
Genitourinary Pathology
Prostatic biopsies, particularly needle biopsies, are good candidates for digitization for a number of reasons: the tissue size is small and the images generated are light-er; multiple measurements are frequently required and informatics tools can facilitate these, and WSI allows a global view to more easily establish the Gleason score ( fig. 4 ) [61] . An additional advantage of WSI is the pos-
sibility to synchronize hematoxylin-eosin stains and p63 IHC in the same screen, thereby allowing the comparison of the two images and facilitating the diagnostic and teaching process [62] .
Thus, the current evidence on the validation of WSI in the diagnosis of prostatic biopsies is more extensive than in other areas. A number of studies (from 50 to over 800 cases) have been focused on the evaluation of the Gleason score in needle biopsies. Scanning at ×200 was considered sufficient to make the diagnosis. The κ values for diagno-sis ranged between 0.586 and 0.813 [63–65] , and one of the reports included only difficult biopsies with a border-line Gleason score. Concordance between WSI and CLM seems to be higher for primary (κ values 0.65–0.96) than for secondary Gleason scores (κ values 0.53–0.75), and most discordances have no impact on patient manage-ment [66] . Tumor size is better evaluated with WSI, and other parameters such as perineural invasion show simi-lar values with WSI and CLM [66] .
Two additional studies focused on genitourinary biop-sies included a mixture of prostatic and urinary tract bi-opsies and showed good intraobserver concordance (90 and 87.5%, respectively) [67, 68] .
Gynecological Pathology
Studies on the validation of WSI in gynecological bi-opsies are scant. Only one study conducted at our institu-tion analyzed interobserver agreement in 452 routine gy-necological specimens showing a κ index of 0.914 (almost perfect concordance). Interestingly, the agreement be-tween WSI and CLM increased in this study in parallel
Fig. 4. Prostatic biopsies often require mul-tiple measurements. The tools of WSI viewers make these measurements easy. WSI allows a global view to more adequate-ly establish the Gleason score.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Saco/Ramírez/Rakislova/Mira/Ordi
Pathobiology 2016;83:89–98DOI: 10.1159/000442823
94
with time, suggesting that there is a learning curve in the use of WSI and that experience in the use of WSI viewers improves the results obtained. Major discrepancies were found in only 2% of the cases, and none was related to poor image quality. Most discrepancies in this study were observed in biopsies of premalignant lesions of the uter-ine cervix, an area which has shown high inter- and in-traobserver variability rates using CLM [36] . The magni-fication used in the study was ×200, and higher magnifi-cation did not seem to be required.
A second study described the usefulness of WSI in the evaluation of 52 frozen ovarian sections showing 96% in-terobserver agreement. Interestingly, in this study, no clinical information was provided to the pathologists, and the time spent per case was 3–5 min [7] .
Head and Neck Pathology
To date, no validation study including the complete spectrum of samples of this subspecialty is available. Only one study on premalignant laryngeal lesions has been published. This study concluded that WSI is a valid alter-native to CLM. Although the correlation with the final diagnosis was slightly lower with WSI than with CLM, the differences were not statistically significant [69] .
Neuropathology
Validation studies of neuropathology are limited to in-traoperative biopsies and smears [5, 8, 70] . Agreement between the diagnosis with WSI and the final diagnosis using CLM is very good, even with low scan magnifica-tion (×100). The studies conclude that ×200 magnifica-tion is sufficient to obtain a diagnosis. In one study, the diagnosis achieved with WSI was concordant with CLM in 29 of the 30 cases evaluated, and the discordant diag-nosis did not lead to changes in the management of the patient [8] . A second study included 126 frozen sections that were evaluated by four different pathologists. The diagnosis was discordant with the final report in only 8 cases. In this study, the diagnosis of frozen sections scanned and diagnosed using WSI was compared with the final diagnosis obtained in formalin-fixed, paraffin-embedded tissue [70] .
Algorithms are currently being developed to identify the hot spots in Ki67-stained sections to automatically quantify the proliferative activity in tumors of the central nervous system [71, 72] .
Pediatric Pathology
Two studies have validated the use of WSI in pediatric pathology. One included 80 routine biopsies of patients under 18 years of age and 20 placentas. The intraobserver concordance between the diagnoses with WSI and CLM was 90% in pediatric biopsies and 93% in placental speci-mens. Major discrepancies were observed in only 2% of the cases. A scanning magnification of ×200 generated an image quality allowing correct diagnosis except for the identification of nucleated red blood cells, which is very difficult even when the slides are scanned at a magnifica-tion of ×400 [73] .
The second study evaluated WSI in 60 cases selected to include the whole spectrum of the diagnostic complexity of pediatric biopsies. The surgical specimens were digitized at ×200 magnification, whereas small biopsies and cyto-logical samples were digitized at ×400. The intraobserver agreement was almost perfect with only 1 discordant case. The scanning process of two cytological smears was unsat-isfactory because the material was very scanty [74] .
Pulmonary Pathology
One study validating WSI in the diagnosis of intraop-erative pulmonary specimens included a variety of sam-ples, with 114 frozen sections from tumors, lymph nodes and bronchial margins, 174 fine-needle aspiration slides, 3 exfoliative smears and 13 small biopsies. This study evaluated both a dynamic real-time telepathology system and WSI, and found very good agreement, which was bet-ter for WSI than for the real-time telepathology system (100% in consultation and frozen biopsies) [75] . A second study analyzed the use of WSI in 20 tumor biopsies and surgical specimens sent for consultation. Complete in-terobserver agreement was achieved in 85% of the cases, even at a scanning magnification of ×100 [75] .
Renal Pathology
Validation studies of WSI in the diagnosis of renal pa-thology biopsies are scarce and include few cases. A re-port including 50 routine renal biopsies showed complete intraobserver agreement in 84% of the cases. Five major discrepancies (with clinical repercussion for the patient) were found and in 2 cases the correct diagnosis was made with WSI. In this study, renal transplant biopsies showed significantly more discrepancies at a magnification of
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Validation of WSI in Primary Diagnosis Pathobiology 2016;83:89–98DOI: 10.1159/000442823
95
×200 [67] . Another study (using a magnification of ×400) reported good agreement in renal transplant biopsies, but the time spent for obtaining the diagnosis was longer with WSI than with CLM [76] . Finally, one study evaluated the concordance between 96 pathologists in the diagnosis of 12 renal biopsies using WSI and CLM and found no sig-nificant differences between both methods [77] .
Intraoperative Diagnosis of Frozen Sections
A number of studies have evaluated dynamic real-time telepathology in intraoperative sections, showing a good correlation with CLM diagnosis. They emphasize the learning curve in the use of the WSI technology, which typically involves a longer diagnostic time window at the beginning but rapid improvement with practice [78] . A validation study using WSI in frozen intraoperative sec-tions from different anatomical sites has shown almost perfect agreement with a κ index of 0.85. The mean time spent on diagnosis was 2 min 50 s per case. The quality of the image was considered excellent in 98% of cases [9] . Studies using WSI in frozen intraoperative sections from specific specialties have been discussed above. In another study evaluating 67 consecutive frozen intraoperative sections, diagnosis was obtained by viewing the virtual slides in a portable device (iPad tablet). The slides were scanned at ×200, and all cases were shown together with the clinical information. The concordance between the diagnoses achieved with WSI and CLM was good with a κ value of 0.85. The mean time to achieving a diagnosis using WSI was 2 min and 46 s [79] .
Surgical Pathology
A number of studies have evaluated a variety of differ-ent specimens from the routine practice in the depart-ment of pathology, including between 25 and 607 samples [22, 39–41, 80, 81] . Inter- and intraobserver agreement between WSI and CLM varied from 75 to 97.7% depend-ing on the study. Most studies concluded that a magnifi-cation of ×200 provides images with adequate quality for diagnosis [22, 39–41, 80, 81] . The interobserver agree-ment between WSI and CLM was 95%, and all discrepan-cies were minor. However, although the general opinion of the pathologists was positive, some felt that the WSI system was slower than CLM, and most of the patholo-gists interviewed were reluctant to completely move from CLM to WSI in routine diagnosis [33] . One study sug-
gested that the interobserver agreement was better for neoplastic than for nonneoplastic diseases [16] . It has been suggested that a scanning magnification of ×200 may not be sufficient to allow correct diagnosis in inflam-matory lesions [82] .
Finally, two studies included only consultation biop-sies of different organs. The interobserver agreement be-tween WSI and CLM diagnoses in these studies was great-er than 91%, and most of the discrepancies were due to the intrinsic difficulty in diagnosing some cases [15, 16, 82] .
Current Recommendations for WSI Validation
Validation of WSI at each institution has been recom-mended before its implementation in routine diagnosis. Several professional associations have developed guide-lines with recommendations for the introduction of WSI in routine diagnosis in a department of pathology. The first guidelines were developed by the College of Ameri-can Pathologists and the American Telemedicine Asso-ciation, and include some recommendations and sugges-tions to be followed before using WSI for diagnosis [2, 10, 17] . It is recommended to include a variety of different biopsies representative of the complexity of the surgical specimens usually analyzed in the center. The guidelines state that it is not necessary to validate each subspecialty because the results from one specialty can be extrapolated to others with similar features. Each specific type of spec-imen with significant differences requires an internal val-idation. The guidelines recommend measuring intraob-server agreement between WSI and CLM, using a ‘wash-out period’ of 2 weeks. Finally, it is recommended that a pathologist with experience in WSI should be involved in the process of validation.
Conclusions
In conclusion, independently of the subspecialty, all the validation studies published show a very good correla-tion between diagnoses achieved with WSI and CLM. Thus, WSI seems to be an adequate tool for histological diagnosis in routine practice and has several advantages over CLM. However, although good evidence demon-strating that WSI can be reliably used for routine diagno-sis has been provided for several specialties, there are a number of areas of pathology, such as hematopathology and liver, endocrine, bone and soft-tissue pathology, for which no study has yet been published. Although some of
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Saco/Ramírez/Rakislova/Mira/Ordi
Pathobiology 2016;83:89–98DOI: 10.1159/000442823
96
these areas may be considered similar to others already validated, specific validation studies are needed in areas with many differences such as liver biopsies or hematopa-thology. These validations are necessary before the use of WSI can be extended to these subspecialties with the aim of going fully digital in pathological services in the future.
Notwithstanding, as with many other new tools, the use of WSI has a learning curve, and the time spent on the diagnosis and, to a lesser extent, inter- and intraobserver agreement may be suboptimal in the initial phases of its use. Cytology seems to be an exception; the application of WSI in this area is more controversial due to the impos-sibility of focusing on different planes.
However, the introduction of WSI in routine diagnosis faces some difficulties, mainly related to the reluctance of pathologists to abandon CLM and to the costs associated with the acquisition of the equipment and the storage of the images generated. New technologies that allow creat-ing 3D reconstruction from 2D biopsies may help to im-prove the understanding of the growth patterns and the spatial arrangement of diseased cells [21, 83] . Another area that will markedly expand in the next few years is that of histopathology pattern recognition using image analy-sis, which can facilitate the diagnostic tasks and improve the reproducibility among pathologists in many subspe-cialties [65, 84–90] .
References
1 Al-Janabi S, Huisman A, van Diest PJ: Digital pathology: current status and future perspec-tives. Histopathology 2012; 61: 1–9.
2 Pantanowitz L, Sinard JH, Henricks WH, Fa-theree LA, Carter AB, Contis L, Beckwith BA, Evans AJ, Lal A, Parwani AV: Validating whole slide imaging for diagnostic purposes in pathology: guideline from the College of American Pathologists Pathology and Labo-ratory Quality Center. Arch Pathol Lab Med 2013; 137: 1710–1722.
3 Weinstein RS: Prospects for telepathology. Hum Pathol 1986; 17: 433–434.
4 Cross SS, Dennis T, Start RD: Telepathology: current status and future prospects in diag-nostic histopathology. Histopathology 2002; 41: 91–109.
5 Evans AJ, Kiehl TR, Croul S: Frequently asked questions concerning the use of whole-slide imaging telepathology for neuropathology frozen sections. Semin Diagn Pathol 2010; 27: 160–166.
6 Evans AJ, Chetty R, Clarke BA, Croul S, Gha-zarian DM, Kiehl TR, Ordonez BP, Ilaalagan S, Asa SL: Primary frozen section diagnosis by robotic microscopy and virtual slide telepa-thology: the University Health Network expe-rience. Semin Diagn Pathol 2009; 26: 165–176.
7 Fallon MA, Wilbur DC, Prasad M: Ovarian frozen section diagnosis: use of whole-slide imaging shows excellent correlation between virtual slide and original interpretations in a large series of cases. Arch Pathol Lab Med 2010; 134: 1020–1023.
8 Gould PV, Saikali S: A comparison of digi-tized frozen section and smear preparations for intraoperative neurotelepathology. Anal Cell Pathol (Amst) 2012; 35: 85–91.
9 Kaplan KJ, Burgess JR, Sandberg GD, Myers CP, Bigott TR, Greenspan RB: Use of robotic telepathology for frozen-section diagnosis: a retrospective trial of a telepathology system for intraoperative consultation. Mod Pathol 2002; 15: 1197–1204.
10 Pantanowitz L, Dickinson K, Evans AJ, Has-sell LA, Henricks WH, Lennerz JK, Lowe A, Parwani AV, Riben M, Smith CD, Tuthill JM, Weinstein RS, Wilbur DC, Krupinski EA, Bernard J: American Telemedicine Associa-tion clinical guidelines for telepathology. J Pathol Inform 2014; 5: 39.
11 Slodkowska J, Pankowski J, Siemiatkowska K, Chyczewski L: Use of the virtual slide and the dynamic real-time telepathology systems for a consultation and the frozen section intra-operative diagnosis in thoracic/pulmonary pathology. Folia Histochem Cytobiol 2009; 47: 679–684.
12 Wilbur DC: Digital pathology: get on board – the train is leaving the station. Cancer Cyto-pathol 2014; 122: 791–795.
13 Ayad E: Virtual telepathology in Egypt, appli-cations of WSI in Cairo University. Diagn Pathol 2011; 6(suppl 1):S1.
14 Romero LG, Cable W, Lesniak A, Tseytlin E, McHugh J, Parwani A, Pantanowitz L: Digital pathology consultations – a new era in digital imaging, challenges and practical applica-tions. J Digit Imaging 2013; 26: 668–677.
15 Wienert S, Beil M, Saeger K, Hufnagl P, Schrader T: Integration and acceleration of virtual microscopy as the key to successful im-plementation into the routine diagnostic pro-cess. Diagn Pathol 2009; 4: 3.
16 Wilbur DC, Madi K, Colvin RB, Duncan LM, Faquin WC, Ferry JA, Frosch MP, Houser SL, Kradin RL, Lauwers GY, Louis DN, Mark EJ, Mino-Kenudson M, Misdraji J, Nielsen GP, Pitman MB, Rosenberg AE, Smith RN, Sohani AR, Stone JR, Tambouret RH, Wu CL, Young RH, Zembowicz A, Klietmann W: Whole-slide imaging digital pathology as a platform for teleconsultation: a pilot study using paired subspecialist correlations. Arch Pathol Lab Med 2009; 133: 1949–1953.
17 Bernard C, Chandrakanth SA, Cornell IS, Dalton J, Evans A, Garcia BM, Godin C, God-lewski M, Jansen GH, Kabani A, Louahlia S, Manning L, Maung R, Moore L, Philley J, Slat-nik J, Srigley J, Thibault A, Picard DD, Cra-cower H, Tetu B: Guidelines from the Cana-dian Association of Pathologists for establish-ing a telepathology service for anatomic pathology using whole-slide imaging. J Pathol Inform 2014; 5: 15.
18 Hedvat CV: Digital microscopy: past, present, and future. Arch Pathol Lab Med 2010; 134: 1666–1670.
19 Ho J, Ahlers SM, Stratman C, Aridor O, Pan-tanowitz L, Fine JL, Kuzmishin JA, Montalto MC, Parwani AV: Can digital pathology result in cost savings? A financial projection for dig-ital pathology implementation at a large inte-grated health care organization. J Pathol In-form 2014; 5: 33.
20 Isaacs M, Lennerz JK, Yates S, Clermont W, Rossi J, Pfeifer JD: Implementation of whole slide imaging in surgical pathology: a value added approach. J Pathol Inform 2011; 2: 39.
21 Pantanowitz L: Digital images and the future of digital pathology. J Pathol Inform 2010; 1: 15.
22 Thorstenson S, Molin J, Lundstrom C: Imple-mentation of large-scale routine diagnostics using whole slide imaging in Sweden: digital pathology experiences 2006–2013. J Pathol Inform 2014; 5: 14.
23 Hartman DJ, Parwani AV, Cable B, Cucoranu IC, McHugh JS, Kolowitz BJ, Yousem SA, Pa-lat V, Reden AV, Sloka S, Lauro GR, Ahmed I, Pantanowitz L: Pocket pathologist: a mobile application for rapid diagnostic surgical pa-thology consultation. J Pathol Inform 2014; 5: 10.
24 Roy S, Pantanowitz L, Amin M, Seethala RR, Ishtiaque A, Yousem SA, Parwani AV, Cuco-ranu I, Hartman DJ: Smartphone adapters for digital photomicrography. J Pathol Inform 2014; 5: 24.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Validation of WSI in Primary Diagnosis Pathobiology 2016;83:89–98DOI: 10.1159/000442823
97
25 Speiser JJ, Hughes I, Mehta V, Wojcik EM, Hutchens KA: Mobile teledermatopathology: using a tablet PC as a novel and cost-efficient method to remotely diagnose dermatopathol-ogy cases. Am J Dermatopathol 2014; 36: 54–57.
26 Gavrielides MA, Conway C, O’Flaherty N, Gallas BD, Hewitt SM: Observer performance in the use of digital and optical microscopy for the interpretation of tissue-based biomarkers. Anal Cell Pathol (Amst) 2014; 2014: 157308.
27 Nassar A, Cohen C, Agersborg SS, Zhou W, Lynch KA, Barker EA, Vanderbilt BL, Thompson J, Heyman ER, Olson A, Lange H, Siddiqui MT: A multisite performance study comparing the reading of immunohisto-chemical slides on a computer monitor with conventional manual microscopy for estro-gen and progesterone receptor analysis. Am J Clin Pathol 2011; 135: 461–467.
28 Micsik T, Kiszler G, Szabo D, Krecsak L, He-gedus C, Tibor K, Molnar B: Computer aided semi-automated evaluation of HER2 immu-nodetection – a robust solution for support-ing the accuracy of anti HER2 therapy. Pathol Oncol Res 2015; 21: 1005–1011.
29 Krenacs T, Zsakovics I, Diczhazi C, Ficsor L, Varga VS, Molnar B: The potential of digital microscopy in breast pathology. Pathol Oncol Res 2009; 15: 55–58.
30 Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJ, Ten Kate FJ, van Dijk MR, van Diest PJ: Whole slide images for primary di-agnostics in dermatopathology: a feasibility study. J Clin Pathol 2012; 65: 152–158.
31 Randell R, Ruddle RA, Mello-Thoms C, Thomas RG, Quirke P, Treanor D: Virtual re-ality microscope versus conventional micro-scope regarding time to diagnosis: an experi-mental study. Histopathology 2013; 62: 351–358.
32 Krishnamurthy S, Mathews K, McClure S, Murray M, Gilcrease M, Albarracin C, Spino-sa J, Chang B, Ho J, Holt J, Cohen A, Giri D, Garg K, Bassett RL Jr, Liang K: Multi-institu-tional comparison of whole slide digital imag-ing and optical microscopy for interpretation of hematoxylin-eosin-stained breast tissue sections. Arch Pathol Lab Med 2013; 137: 1733–1739.
33 Houghton JP, Ervine AJ, Kenny SL, Kelly PJ, Napier SS, McCluggage WG, Walsh MY, Hamilton PW: Concordance between digital pathology and light microscopy in general surgical pathology: a pilot study of 100 cases. J Clin Pathol 2014; 67: 1052–1055.
34 Sanders DS, Grabsch H, Harrison R, Bateman A, Going J, Goldin R, Mapstone N, Novelli M, Walker MM, Jankowski J: Comparing virtual with conventional microscopy for the con-sensus diagnosis of Barrett’s neoplasia in the AspECT Barrett’s chemoprevention trial pa-thology audit. Histopathology 2012; 61: 795–800.
35 Randell R, Ruddle RA, Thomas RG, Mello-Thoms C, Treanor D: Diagnosis of major can-cer resection specimens with virtual slides: impact of a novel digital pathology worksta-tion. Hum Pathol 2014; 45: 2101–2106.
36 Ordi J, Castillo P, Saco A, Del Pino M, Ordi O, Rodriguez-Carunchio L, Ramirez J: Vali-dation of whole slide imaging in the primary diagnosis of gynaecological pathology in a university hospital. J Clin Pathol 2015; 68: 33–39.
37 Cornish TC, Swapp RE, Kaplan KJ: Whole-slide imaging: routine pathologic diagnosis. Adv Anat Pathol 2012; 19: 152–159.
38 Pantanowitz L, Valenstein PN, Evans AJ,Kaplan KJ, Pfeifer JD, Wilbur DC, Collins LC, Colgan TJ: Review of the current state of whole slide imaging in pathology. J Pathol In-form 2011; 2: 36.
39 Gilbertson JR, Ho J, Anthony L, Jukic DM, Yagi Y, Parwani AV: Primary histologic diag-nosis using automated whole slide imaging: a validation study. BMC Clin Pathol 2006; 6: 4.
40 Jukic DM, Drogowski LM, Martina J, Parwani AV: Clinical examination and validation of primary diagnosis in anatomic pathology us-ing whole slide digital images. Arch Pathol Lab Med 2011; 135: 372–378.
41 Li X, Liu J, Xu H, Gong E, McNutt MA, Li F, Anderson VM, Gu J: A feasibility study of vir-tual slides in surgical pathology in China. Hum Pathol 2007; 38: 1842–1848.
42 Stathonikos N, Veta M, Huisman A, van Diest PJ: Going fully digital: perspective of a Dutch academic pathology lab. J Pathol Inform 2013; 4: 15.
43 Bauer TW, Schoenfield L, Slaw RJ, Yerian L, Sun Z, Henricks WH: Validation of whole slide imaging for primary diagnosis in surgi-cal pathology. Arch Pathol Lab Med 2013; 137: 518–524.
44 Al-Janabi S, Huisman A, Nap M, Clarijs R, van Diest PJ: Whole slide images as a platform for initial diagnostics in histopathology in a medium-sized routine laboratory. J Clin Pathol 2012; 65: 1107–1111.
45 Pantanowitz L, Wiley CA, Demetris A, Les-niak A, Ahmed I, Cable W, Contis L, Parwani AV: Experience with multimodality telepa-thology at the University of Pittsburgh Medi-cal Center. J Pathol Inform 2012; 3: 45.
46 Al-Janabi S, Huisman A, Willems SM, vanDiest PJ: Digital slide images for primary di-agnostics in breast pathology: a feasibility study. Hum Pathol 2012; 43: 2318–2325.
47 Reyes C, Ikpatt OF, Nadji M, Cote RJ: Intra-observer reproducibility of whole slide imag-ing for the primary diagnosis of breast needle biopsies. J Pathol Inform 2014; 5: 5.
48 Lopez AM, Graham AR, Barker GP, Richter LC, Krupinski EA, Lian F, Grasso LL, Miller A, Kreykes LN, Henderson JT, Bhattacharyya AK, Weinstein RS: Virtual slide telepathology enables an innovative telehealth rapid breast care clinic. Semin Diagn Pathol 2009; 26: 177–186.
49 Nassar A, Cohen C, Albitar M, Agersborg SS, Zhou W, Lynch KA, Heyman ER, Lange H, Siddiqui MT: Reading immunohistochemical slides on a computer monitor – a multisite performance study using 180 HER2-stained breast carcinomas. Appl Immunohistochem Mol Morphol 2011; 19: 212–217.
50 Khurana KK: Telecytology and its evolving role in cytopathology. Diagn Cytopathol 2012; 40: 498–502.
51 Thrall M, Pantanowitz L, Khalbuss W: Tele-cytology: clinical applications, current chal-lenges, and future benefits. J Pathol Inform 2011; 2: 51.
52 Donnelly AD, Mukherjee MS, Lyden ER, Bridge JA, Lele SM, Wright N, McGaughey MF, Culberson AM, Horn AJ, Wedel WR, Ra-dio SJ: Optimal z-axis scanning parameters for gynecologic cytology specimens. J Pathol Inform 2013; 4: 38.
53 Wilbur DC: Digital cytology: current state of the art and prospects for the future. Acta Cy-tol 2011; 55: 227–238.
54 Buxbaum JL, Eloubeidi MA, Lane CJ, Varada-rajulu S, Linder A, Crowe AE, Jhala D, Jhala NC, Crowe DR, Eltoum IA: Dynamic telecy-tology compares favorably to rapid onsite evaluation of endoscopic ultrasound fine nee-dle aspirates. Dig Dis Sci 2012; 57: 3092–3097.
55 Collins BT: Telepathology in cytopathology: challenges and opportunities. Acta Cytol 2013; 57: 221–232.
56 Al Habeeb HA, Evans A, Ghazarian D: Vir-tual microscopy using whole-slide imaging as an enabler for teledermatopathology: a paired consultant validation study. J Pathol Inform 2012; 3: 2.
57 Nielsen PS, Lindebjerg J, Rasmussen J, Starklint H, Waldstrom M, Nielsen B: Virtual microscopy: an evaluation of its validity and diagnostic performance in routine histologic diagnosis of skin tumors. Hum Pathol 2010; 41: 1770–1776.
58 Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJ, Ten Kate FJ, van Diest PJ: Whole slide images for primary diagnostics of gastrointestinal tract pathology: a feasibility study. Hum Pathol 2012; 43: 702–707.
59 Molnar B, Berczi L, Diczhazy C, Tagscherer A, Varga SV, Szende B, Tulassay Z: Digital slide and virtual microscopy based routine and telepathology evaluation of routine gas-trointestinal biopsy specimens. J Clin Pathol 2003; 56: 433–438.
60 Risio M, Bussolati G, Senore C, Vigna S, Fran-gipane E, Segnan N, Cassoni P: Virtual mi-croscopy for histology quality assurance of screen-detected polyps. J Clin Pathol 2010; 63: 916–920.
61 Camparo P, Egevad L, Algaba F, Berney DM, Boccon-Gibod L, Comperat E, Evans AJ, Grobholz R, Kristiansen G, Langner C, Lo-pez-Beltran A, Montironi R, Oliveira P, Vain-er B, Varma M: Utility of whole slide imaging and virtual microscopy in prostate pathology. APMIS 2012; 120: 298–304.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Saco/Ramírez/Rakislova/Mira/Ordi
Pathobiology 2016;83:89–98DOI: 10.1159/000442823
98
62 Helin HO, Lundin ME, Laakso M, Lundin J, Helin HJ, Isola J: Virtual microscopy in pros-tate histopathology: simultaneous viewing of biopsies stained sequentially with hematoxy-lin and eosin, and alpha-methylacyl-coen-zyme A racemase/p63 immunohistochemis-try. J Urol 2006; 175: 495–499.
63 Chargari C, Comperat E, Magne N, Vedrine L, Houlgatte A, Egevad L, Camparo P: Pros-tate needle biopsy examination by means of virtual microscopy. Pathol Res Pract 2011; 207: 366–369.
64 Fine JL, Grzybicki DM, Silowash R, Ho J, Gil-bertson JR, Anthony L, Wilson R, Parwani AV, Bastacky SI, Epstein JI, Jukic DM: Evalu-ation of whole slide image immunohisto-chemistry interpretation in challenging pros-tate needle biopsies. Hum Pathol 2008; 39: 564–572.
65 Helin H, Lundin M, Lundin J, Martikainen P, Tammela T, Helin H, van der Kwast T, Isola J: Web-based virtual microscopy in teaching and standardizing Gleason grading. Hum Pathol 2005; 36: 381–386.
66 Rodriguez-Urrego PA, Cronin AM, Al-Ah-madie HA, Gopalan A, Tickoo SK, Reuter VE, Fine SW: Interobserver and intraobserver re-producibility in digital and routine micro-scopic assessment of prostate needle biopsies. Hum Pathol 2011; 42: 68–74.
67 Al-Janabi S, Huisman A, Jonges GN, Ten Kate FJ, Goldschmeding R, van Diest PJ: Whole slide images for primary diagnostics of uri-nary system pathology: a feasibility study. J Renal Inj Prev 2014; 3: 91–96.
68 Ho J, Parwani AV, Jukic DM, Yagi Y, Antho-ny L, Gilbertson JR: Use of whole slide imag-ing in surgical pathology quality assurance: design and pilot validation studies. Hum Pathol 2006; 37: 322–331.
69 Sturm B, Fleskens SJ, Bot FJ, van Velthuysen ML, Speel EJ, Slootweg PJ, van der Laak JA: Virtual microscopy is a valid alternative for the diagnostic assessment of laryngeal prema-lignancies. Histopathology 2014; 64: 602–604.
70 Wiley CA, Murdoch G, Parwani A, Cudahy T, Wilson D, Payner T, Springer K, Lewis T: In-terinstitutional and interstate teleneuropa-thology. J Pathol Inform 2011; 2: 21.
71 Grala B, Markiewicz T, Kozlowski W, Osow-ski S, Slodkowska J, Papierz W: New automat-ed image analysis method for the assessment of Ki-67 labeling index in meningiomas. Folia Histochem Cytobiol 2009; 47: 587–592.
72 Alomari YM, Sheikh Abdullah SN, MdZin RR, Omar K: Adaptive localization of focus point regions via random patch probabilistic density from whole-slide, Ki-67-stained brain tumor tissue. Comput Math Methods Med 2015; 2015: 673658.
73 Al-Janabi S, Huisman A, Nikkels PG, Ten Kate FJ, van Diest PJ: Whole slide images for primary diagnostics of paediatric pathology specimens: a feasibility study. J Clin Pathol 2013; 66: 218–223.
74 Arnold MA, Chenever E, Baker PB, Boue DR, Fung B, Hammond S, Hendrickson BW, Kah-wash SB, Pierson CR, Prasad V, Nicol KK, Barr T: The College of American Pathologists guidelines for whole slide imaging validation are feasible for pediatric pathology: a pediat-ric pathology practice experience. Pediatr Dev Pathol 2015; 18: 109–116.
75 Slodkowska J, Chyczewski L, Wojciechowski M: Virtual slides: application in pulmonary pathology consultations. Folia Histochem Cytobiol 2008; 46: 121–124.
76 Jen KY, Olson JL, Brodsky S, Zhou XJ, Na-dasdy T, Laszik ZG: Reliability of whole slide images as a diagnostic modality for renal al-lograft biopsies. Hum Pathol 2013; 44: 888–894.
77 Furness P: A randomized controlled trial of the diagnostic accuracy of internet-based tele-pathology compared with conventional mi-croscopy. Histopathology 2007; 50: 266–273.
78 Baak JP, van Diest PJ, Meijer GA: Experience with a dynamic inexpensive video-conferenc-ing system for frozen section telepathology. Anal Cell Pathol 2000; 21: 169–175.
79 Ramey J, Fung KM, Hassell LA: Use of mobile high-resolution device for remote frozen sec-tion evaluation of whole slide images. J Pathol Inform 2011; 2: 41.
80 Buck TP, Dilorio R, Havrilla L, O’Neill DG: Validation of a whole slide imaging system for primary diagnosis in surgical pathology: a community hospital experience. J Pathol In-form 2014; 5: 43.
81 Campbell WS, Lele SM, West WW, Lazenby AJ, Smith LM, Hinrichs SH: Concordance be-tween whole-slide imaging and light micros-copy for routine surgical pathology. Hum Pathol 2012; 43: 1739–1744.
82 Dangott B, Parwani A: Whole slide imaging for teleconsultation and clinical use. J Pathol Inform 2010; 1: 7.
83 Song Y, Treanor D, Bulpitt AJ, Magee DR: 3D reconstruction of multiple stained histology images. J Pathol Inform 2013; 4:S7.
84 Caie PD, Turnbull AK, Farrington SM, Onis-cu A, Harrison DJ: Quantification of tumour budding, lymphatic vessel density and inva-sion through image analysis in colorectal can-cer. J Transl Med 2014; 12: 156.
85 Murakami Y, Abe T, Hashiguchi A, Yamagu-chi M, Saito A, Sakamoto M: Color correction for automatic fibrosis quantification in liver biopsy specimens. J Pathol Inform 2013; 4: 36.
86 Neil DA, Roberts IS, Bellamy CO, Wigmore SJ, Neuberger JM: Improved access to histo-pathology using a digital system could in-crease the organ donor pool and improve al-location. Transpl Int 2014; 27: 759–764.
87 Neltner JH, Abner EL, Schmitt FA, Denison SK, Anderson S, Patel E, Nelson PT: Digital pathology and image analysis for robust high-throughput quantitative assessment of Alz-heimer disease neuropathologic changes. J Neuropathol Exp Neurol 2012; 71: 1075–1085.
88 Riber-Hansen R, Vainer B, Steiniche T: Digi-tal image analysis: a review of reproducibility, stability and basic requirements for optimal results. APMIS 2012; 120: 276–289.
89 Webster JD, Michalowski AM, Dwyer JE, Corps KN, Wei BR, Juopperi T, Hoover SB, Simpson RM: Investigation into diagnostic agreement using automated computer-assist-ed histopathology pattern recognition image analysis. J Pathol Inform 2012; 3: 18.
90 Webster JD, Dunstan RW: Whole-slide imag-ing and automated image analysis: consider-ations and opportunities in the practice of pa-thology. Vet Pathol 2014; 51: 211–223.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
12:
01:4
3 P
M
Tesis Doctoral. Adela Saco Álvarez
[63]
Estudio número 2
“Validation of Whole-Slide Imaging in the Primary Diagnosis of
Gynecological Pathology in a University Hospital”
Jaume Ordi, Paola Castillo, Adela Saco, Marta del Pino, Oriol Ordi,
Leonardo Rodríguez-Carunchio, Jose Ramírez
Journal of Clinical Pathology 2015 Jan; 68 (1): 33 – 9
Factor de impacto (2016): 2.687
Ranking (2016): 33/193, primer cuartil
Microscopía virtual en el diagnóstico rutinario y la docencia
[64]
Validation of whole slide imaging in the primarydiagnosis of gynaecological pathology in aUniversity HospitalJaume Ordi,1,2,3 Paola Castillo,1,3 Adela Saco,1 Marta del Pino,4 Oriol Ordi,2
Leonardo Rodríguez-Carunchio,1 Jose Ramírez1,2
▸ Additional material ispublished online only. To viewplease visit the journal online(http://dx.doi.org/10.1136/jclinpath-2014-202524).1Department of Pathology,Hospital Clínic, Barcelona,Spain2University of Barcelona,School of Medicine, Barcelona,Spain3Centre de Recerca en SalutInternacional de Barcelona(CRESIB), Barcelona, Spain4Institute of Gynecology,Obstetrics and Neonatology,Hospital Clínic—Institut d´Investigacions BiomèdiquesAugust Pi I Sunyer (IDIBAPS),Faculty of Medicine-Universityof Barcelona, Barcelona, Spain
Correspondence toProfessor Jaume Ordi,Department of Pathology,CRESIB (Centre de Recerca enSalut Internacional deBarcelona) -Hospital Clínic,University of Barcelona Facultyof Medicine, Barcelona, Spain,C/Villarroel 170, 08036-Barcelona, Spain;[email protected]
Presented in part at the annualmeeting of the USA andCanadian Academy ofPathology; March 2014; SanDiego, California, USA.
Received 18 July 2014Revised 23 September 2014Accepted 8 October 2014Published Online First29 October 2014
To cite: Ordi J, Castillo P,Saco A, et al. J Clin Pathol2015;68:33–39.
ABSTRACTAims Experience in the use of whole slide imaging(WSI) for primary diagnosis in pathology is very limited.We aimed to determine the accuracy of interpretation ofWSI compared with conventional light microscopy (CLM)in the diagnosis of routine gynaecological biopsies.Methods All gynaecological specimens (n=452)received over a 2-month period at the Department ofPathology of the Hospital Clinic of Barcelona wereanalysed blindly by two gynaecological pathologists, oneusing CLM and the other WSI. All slides were digitisedin a Ventana iScan HT (Roche diagnostics) at 200×.All discrepant diagnoses were reviewed, and a finalconsensus diagnosis was established. The results wereevaluated by weighted κ statistics for two observers.Results The level of interobserver agreement betweenWSI and CLM evaluations was almost perfect (κ value:0.914; 95% CI 0.879 to 0.949) and increased duringthe study period: κ value 0.890; 95% CI 0.835 to0.945 in the first period and 0.941; 95%; CI 0.899 to0.983 in the second period. Major discrepancies(differences in clinical management or prognosis) wereobserved in 9 cases (2.0%). All discrepancies consistedof small lesions (8 high grade squamous intraepitheliallesions of the uterine cervix, one lymph nodemicrometastasis of an ovarian carcinoma)underdiagnosed or missed in the WSI or the CLMevaluation. Discrepancies with no or minor clinicalrelevance were identified in 3.8% of the biopsies. Nodiscrepancy was related to the poor quality of the WSIimage.Conclusions Diagnosis of gynaecological specimens byWSI is accurate and may be introduced into routinediagnosis.
INTRODUCTIONWhole slide imaging (WSI), also referred to asvirtual microscopy or digital pathology allows digit-isation of entire glass slides to achieve the diagnosisof pathological specimens. WSI scanners create adigital slide of the tissue section which, with theuse of specific software, can be viewed and magni-fied in real time across the web very much like theuse of a conventional light microscope (CLM).1–3
WSI has been shown to have many practical appli-cations including education,4–6 teleconsultation forsecond opinions,7–10 intraoperative frozen sectionconsultation10 11 and quality assurance.12 13
Potential additional benefits of WSI includeimprovement of the workflow by eliminating thetask of slide distribution and facilitating slide filingand retrieval.1–3
The rapid advances in this technology and itsmany potential benefits will probably result in aprogressive shift from conventional to virtualmicroscopy in the routine diagnosis in pathology.Currently, several commercially available systemsare able to digitise glass slides containing tissue sec-tions and produce virtual slides of excellent quality.However, although WSI has been around for morethan a decade, its widespread application inprimary histological diagnosis still awaits validationas opposed to traditional CLM. Moreover, despiteseveral pilot studies suggesting that WSI is as usefulas CLM for diagnostic purposes,14–22 WSI-baseddiagnosis has yet to be integrated in routine patho-logical studies, with a very small number of excep-tions. The use of WSI in the routine practice ofpathology laboratories is still not common becauseof difficulties in the integration between the WSIsoftware and the laboratory information systems(LIS), insufficient scanning speed and robustness,and limitations in storage capacity and/or excessivecosts of file storage. The lack of systematic valid-ation studies on their use in primary diagnosis isalso a major concern that hampers the introductionof this technique.1 3
The College of American Pathologists Pathologyand Laboratory Quality Center has recently pub-lished the guidelines for validating the use of WSIfor diagnostic purposes.3 According to these guide-lines all laboratories should conduct their own val-idation studies, and the validation should include atleast 60 samples. However, very few large valid-ation studies have focused on pathology subspecial-ties. Indeed, only one study has been published ongynaecological disease, evaluating the correlationbetween WSI and CLM in the assessment of thediagnoses of frozen sections of 52 ovarianlesions.11 No previous studies have evaluated theaccuracy of WSI diagnosis in the routine practice ofgynaecological pathology. In the present study weaimed to determine the accuracy of interpretationusing WSI as compared with CLM in a series ofconsecutive gynaecological specimens, representa-tive of the spectrum of specimen types and diagno-ses encountered in the routine practice of a largeacademic department.
MATERIALS AND METHODSCharacteristics of the institutionThis study was performed at the Department ofPathology, Hospital Clinic (Barcelona, Spain), alarge academic department composed of 15 staffpathologists, 8 residents and additional fellows.
Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524 33
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
There are 14 subspecialties, and most pathologists limit theirpractice to one or two subspecialty areas. In 2013 theDepartment handled 41 928 specimens with 83 619 blocks. Thenumber of gynaecological specimens analysed during 2013 was4909, with 16 805 blocks of gynaecological specimens and witha median number of slides per case of 1.
Sample size calculationBased on previous reports,15 23 the major rate of discrepancybetween the original diagnosis by CLM and that by WSI wascalculated to be 3%, with a non-inferiority margin for WSIreview of 4%. A one-sided binomial test was used for compari-son at a level of significance of 0.05. The power to be achievedwas 80%, and the level of significance was 0.05. Based on theseassumptions, it was calculated that 450 cases would need to bereviewed to establish non-inferiority.
Specimens included in the studyAll gynaecological specimens consecutively received over a2-month period ( July–August 2013) were included in the studyat the Department of Pathology of the Hospital Clinic ofBarcelona (n=452). This represented 9.21% (452/4909) of thetotal number of gynaecological specimens evaluated in 2013:353/452 (78.1%) specimens were evaluated in the 1st month( July) and 99/452 (21.9%) in the 2nd month (August). Table 1shows a summary of the location and type (either biopsy or resec-tion) of the specimens, and the number of cases included foreach type. The number of blocks per case ranged from 1 to 30(median 1, IQR 1–3). The overall number of slides scanned was1253.
To evaluate possible changes associated with increasingexperience in the use of WSI over the study period and theagreement between WSI and CLM observers, the 456 specimenswere divided into two periods: one including the first 226 speci-mens and the second including the last 226 specimens.
The Department of Gynecology of our hospital has a veryactive referral Colposcopy Clinic.24 Consequently, the seriesincluded 157 biopsies or excisions of the uterine cervix frompatients referred to colposcopy because of an abnormal Papsmear. In addition to evaluation in the general analysis, thesespecimens were evaluated independently.
The study was approved by the Hospital Clinic InstitutionalEthical Review Board.
Scanning process and characteristics of the WSI displayIn July and August 2013 all the slides of gynaecological path-ology were scanned daily after diagnosis by light microscopy.Scanning was performed on a Ventana iScan HT (RocheDiagnostics, Sant Cugat, Spain) at a magnification of 200×. Thesystem creates high-resolution digital images of the tissue sec-tions. The whole scanning process runs automatically (includingselection of the area of the slide that contains tissue, placingfocus points, calibration, etc). In cases with step sections of asample on a single slide the system scans all the sections. Nospecific quality control of the slides scanned was made by thetechnicians prior to evaluation by the pathologist. The WSI pro-duced are stored in a dedicated mass storage environment andlinked to the pathology report, based on the recognition of aquick response (QR) code on the slide label. Although WSI canbe accessed through the pathology LIS software (Novopath,Vitrosoft, Sevilla, Spain), for the purposes of this study theaccession to the WSI was made through the viewer.
The images are viewed in the Virtuoso viewer (Roche), whichworks as a web browser and simulates a conventional micro-scope. The images are shown using the same structure providedby the LIS. No specific software installation is required to visual-ise the WSI. The images scanned can be viewed up to a realmagnification of 200× and up to 400× with digital zoom, arealways in focus, with optimised contrast and adjusted illumin-ation. The viewer shows a thumbnail of the whole slide, whichallows confirmation that all the material present on the glassslide has been included in the digital image and helps in thenavigation through the slide. Figure 1 shows the appearance ofthe virtual microscope display.
The WSI were displayed on a 30” Coronis fusion MDC4130monitor which has a resolution of four megapixels (BarcoElectronic Systems, Barcelona, Spain).
WSI and CLM diagnosisAll cases were analysed blindly by two gynaecological patholo-gists, one using CLM and the other WSI. The pathologist doingthe WSI evaluations had previously had a 1-week trainingcourse on the use of WSI. WSI were presented to the patholo-gist per case, together with the original clinical information inorder to emulate the real clinical environment, and blinded tothe original report based on CLM. For the purposes of thisstudy only the H&E slides were evaluated.
The original CLM and the WSI-based diagnoses were com-pared by an independent gynaecologist, who judged the con-cordance of the two diagnoses as: (A) complete agreementbetween the original diagnosis and that determined with WSI;(B) minor discrepancy (mild differences which would not have
Table 1 Location and type (either small biopsies or surgicalresections) of the specimens evaluated in the study
Location Number (%) Biopsies Surgical specimens
Vulva 19 (4) 13 6Vagina 12 (3) 12 1Uterine cervix 125 (28) 93 32Endocervix 46 (10) 46 0Endometrium 60 (13) 52 8Uterus 69 (15) 0 69Fallopian tube 12 (3) 0 12Ovary 36 (8) 0 36Lymph nodes 44 (10) 0 44Peritoneum 14 (3) 0 14Other * 15 (3) 0 15
*Includes biopsies of the abdominal wall (1), large bowel (9), urinary bladder (4) andureter (1). Figure 1 Screenshot of the virtual microscope display.
34 Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
any clinical or prognostic implications); and (C) major discrep-ancy (differences with clinical and/or prognostic implications forthe patient).
Final gold standard diagnosisThe gold standard was considered as the concordant diagnosisin all cases with complete agreement in both evaluations. Eachcase with a discrepant result was reviewed by the two patholo-gists involved in the study. The revision was made using CLM,and a final consensus diagnosis was established. In this finaladjudication process, the H&E slides, as well as the immunohis-tochemical stains (when necessary) were used to achieve thefinal diagnosis.
Statistical analysisThe SPSS statistical programme (SPSS TM140, V.18, Chicago,Illinois, USA) was used for statistical analysis. The results for cat-egorical variables are expressed as absolute numbers and percen-tages and 95% CIs. The χ2 or the Fisher’s exact tests were usedto compare the variables. The results were evaluated byweighted κ statistics for two raters. This measure calculates thedegree of agreement in classification over that which would beexpected by chance and is scored as a number between 0 and1. Following the Landis-Koch benchmarks the strength of agree-ment of the κ values is: 0 poor; 0–0.20 slight; 0.21–0.40 fair;0.41–0.60 moderate; 0.61–0.80 substantial; 0.81–1.00 almostperfect.25 For the purposes of weighted κ calculation the diag-noses were categorised from normal to cancer as 1: normaltissue or reactive lesions; 2: benign tumours; 3: low-grade pre-malignant lesions; 4: high-grade premalignant lesions; and 5:malignant tumours.
RESULTSFinal diagnosesOverall, 218/452 specimens (48.2%) were evaluated as beingcomposed of normal tissue or showing reactive lesions; 130(28.8%) were benign tumours, 18 (4.0%) were low-grade pre-malignant lesions, 48 (10.6%) high grade premalignant lesionsand 38 (8.4%) showed malignant tumours. Table 2 shows thedistribution of these diagnoses for each specific site.
Agreement between WSI and CLM evaluationsInterpretations by WSI and CLM completely agreed in 94.2%of the biopsies (95% CI 91.7 to 96.0) while major discrepancieswere observed in 9/452 (2.0%) and minor discrepancies wereidentified in 17/452 (3.8%) of the biopsies.
The final consensus diagnosis achieved after the adjudicationmeeting was in agreement with the CLM evaluation in 7/9(77.8%) major discrepancies and in 11/17 (64.7%) minor dis-crepancies. Discrepancy in interpretations between WSI andCLM evaluations occurred in only two settings. Eight out of thenine major discrepancies (88.9%) observed in the study wererelated to the diagnosis of a high-grade squamous intraepitheliallesion (H-SIL) of the uterine cervix, as a low-grade squamousintraepithelial lesion (L-SIL) or as negative (four cases each,figure 2). The consensus diagnosis was in keeping with theCLM evaluation in six of eight cases and with the WSI evalu-ation in two of eight cases. The last case was a small lymphnode metastasis of an ovarian carcinoma missed in the WSIevaluation (figure 3). In this latter case a small lipogranulomawas identified close to the small metastatic nest missed in theWSI evaluation. Thirteen out of the 17 minor discrepancieswere related to overdiagnosis or underdiagnosis of L-SIL. In 10cases the L-SIL lesions were missed in the evaluation (8 in the
WSI and 2 in the CLM evaluation). Three cases were reactivechanges in the uterine cervix overdiagnosed as L-SIL (threebiopsies, two overdiagnosed in the WSI and one in the CLMevaluation). Two cases of endometrial polyps were missed (onecase missed in the WSI evaluation) or overdiagnosed (one case,overdiagnosed in the CLM evaluation). The other two minordiscrepancies were two small foci of endometriosis (one in theovary, one in the Fallopian tube) missed in the CLM evaluation.None of the discrepancies was related to the poor quality of theWSI image or to insufficient magnification.
Overall the level of interobserver agreement between the WSIand CLM evaluations was almost perfect (κ value: 0.914; 95%CI 0.879 to 0.949).
Concordance in biopsies of the uterine cervixand in other samplesIn the subset of 157 biopsies or excisions of the uterine cervixfrom patients referred to colposcopy because of abnormal Papsmear, complete agreement was observed between the WSI andCLM interpretations in 86.6% (95% CI 80.3 to 91.5) of thebiopsies. Major discrepancies were observed in 8/157 (5.1%)and minor discrepancies in 13/157 (8.3%) of the samples. Theκ value for this subset of samples was 0.832 (95% CI 0.757 to0.906).
In the 295 gynaecological specimens representing tissues otherthan cervical biopsies and excisions, complete agreement betweenWSI and CLM was observed in 98.3% (95% CI 96.1 to 99.4) ofthe biopsies. Major discrepancies were observed in 1/295 (0.3%)and minor discrepancies in 4/295 (1.4%) of the samples. The κvalue for this subset of samples was 0.976 (95% CI 0.950 to 1).
κ Analysis and discrepant diagnoses in thetwo study periodsInterobserver agreement increased during the study period,κ value: 0.890 (95% CI 0.835 to 0.945) in the first period, andκ value: 0.941 (95% CI 0.899 to 0.983) in the second period.In the first period of the study 5/226 (2.2%) major discrepanciesand 12/226 (5.3%) minor discrepancies were detected. Thenumber of major and minor discrepancies in the second periodwas 4/226 (1.80%) and 5/226 (2.2%), respectively. Interestingly,whereas the consensus gold standard diagnosis was in keepingwith the CLM diagnosis in 14/17 (82.4%) discrepanciesobserved in the initial period, the consensus was in keepingwith the WSI evaluation in 5/9 discrepancies (55.5%) observedin the second period (p=0.078).
DISCUSSIONThe results of our study show a high concordance between WSIand CLM evaluations (over 94%) in the diagnosis of a largeseries of routine gynaecological specimens. The κ value, consid-ered as a measure of the level of agreement among observerscorrected by chance, was at the almost perfect level (0. 914).Thus, our results confirm that WSI may safely be used for per-forming primary histological diagnoses of gynaecologicalspecimens.
The results of our study are comparable with other valid-ation studies conducted on skin,5 19 breast,14 26 prostate,17 27
urinary bladder,18 gastrointestinal8 20 and paediatric path-ology,28 which show a high rate of concordance between WSIand CLM-based diagnoses. However, no previous studies haveevaluated the accuracy of WSI diagnosis in the routine practiceof gynaecological pathology, and neither are there any dataabout intraobserver or interobserver agreement in the evalu-ation of routine gynaecological specimens using CLM. The rate
Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524 35
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
of discrepancies observed in our study is within the range ofgenerally observed intraobserver variability in pathology.27 29
Interestingly, the final consensus diagnosis was in agreementwith the WSI evaluation in 22.2% of the major discrepanciesand in 35.3% of the minor discrepancies. None of the
discrepancies was related to the poor quality of the WSI imageor to insufficient magnification, but rather were mostly asso-ciated with different interpretations of difficult or borderlinecases or with the presence of small lesions overlooked in theevaluation.
Table 2 Absolute numbers and percentages (in parenthesis) of specimens showing normal/reactive lesions, benign tumours, low-gradepremalignant, high-grade premalignant and malignant tumours in each location
Location Normal/reactive Benign tumoursLow-gradepremalignant lesions
High-gradepremalignant lesions Malignant tumours
Vulva 7 (36.8) 6 (31.6) 0 (0) 5 (26.3) 1 (5.3)Vagina 8 (66.7) 0 (0) 0 (0) 0 (0) 4 (33.3)Uterine cervix 64 (51.2) 3 (2.4) 16 (12.8) 37 (29.6) 5 (4.0)Endocervix 30 (65.2) 9 (19.6) 2 (4.3) 3 (6.6) 2 (4.3)Endometrium 29 (48.3) 28 (46.7) 0 (0) 1 (1.7) 2 (3.3)Uterus 13 (18.8) 46 (66.7) 0 (0) 2 (2.9) 8 (11.6)Fallopian tube 7 (58.3) 5 (41.7) 0 (0) 0 (0) 0 (0)Ovary 8 (22.2) 24 (66.7) 0 (0) 0 (0) 4 (11.1)Lymph nodes 41 (93.2) 0 (0) 0 (0) 0 (0) 3 (6.8)Peritoneum 6 (42.9) 0 (0) 0 (0) 0 (0) 8 (57.1)Other * 5 (33.3) 9 (60.0) 0 (0) 0 (0) 1 (6.7)
*Includes biopsies of the abdominal wall (1), large bowel (9), urinary bladder (4) and ureter (1).
Figure 2 One of the major discrepancies identified in the study: a small area of high-grade squamous intraepithelial lesion (H-SIL) involving thesquamous epithelium of the uterine cervix diagnosed as reactive changes in the whole slide imaging (WSI) evaluation. In the consensus meeting ap16 staining was requested, which confirmed the diagnosis of H-SIL rendered by conventional light microscopy (A and C, H&E; B and D p16immunostaining; screenshots of the WSI image).
36 Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
Eight out of the nine discrepancies (88.9%) observed in thestudy were related to the diagnosis of H-SIL as L-SIL or as nega-tive or reactive changes in the uterine cervix (four cases each).Similarly, 13 out of the 17 minor discrepancies (76.5%) wererelated to discrepancies in the diagnosis of L-SIL lesions versusnormal/reactive cervical epithelium. Consequently, the κ valuefor the subset of cervical biopsies or excisions from patientsreferred because of abnormal Pap smear was 0.832, clearly lowerthan the general value. A number of studies using CLM haveshown that there is a substantial variation between andwithin-observers in the interpretation of squamous intraepitheliallesions on H&E-stained tissue sections. Indeed, κ coefficients aretypically found within the range of 0.45–0.50,29–36 indicatingmoderate agreement. Estimates for the reproducibility of histo-logical cervical specimen interpretations performed in the courseof the ASCUS/L-SIL Triage Study comparing the diagnosticresults of the original clinical centre pathologists with the resultsfrom a quality control review35 showed that the reproducibilityof histological interpretations of biopsy specimens was moderate(κ<0.5). In the latter study, the lack of reproducibility was sub-stantially higher for punch biopsy specimens than loop excisionprocedure specimens, and variability was more prominent in thelow-grade abnormalities, similar to what was observed in our
study. The p16 immunohistochemical stain, strongly expressed inalmost all H-SIL and not in reactive lesions,36 has recently beenrecommended by the College of American Pathologists to reduceinterobserver variability, particularly in cases of professional dis-agreement.37 This technique was used in our series to achieve thefinal consensus diagnosis in all disagreements between CLM andWSI detected in biopsies of the uterine cervix.
The pathologist who performed the WSI evaluation had pre-viously had very little experience in the use of WSI, althoughthis did not severely affect the reproducibility, even in the initialperiod of the study. Nevertheless, a clear increase in the rate ofreproducibility was observed during the study period. This sug-gests that minor difficulties may arise in the initial periods ofusing this new tool and that increasing experience with WSIincreases the accuracy of the diagnosis. Moreover, whereas theconsensus diagnosis reached in the discrepant cases was inkeeping with the CLM diagnosis in most discrepancies observedin the initial period (82.4%), in the second period there was atendency towards a more balanced situation and in 55.5% ofthe cases the gold standard diagnosis was in keeping with theWSI evaluation.
The pathologist working with WSI did not report difficultiesin rendering the diagnosis at the magnification of 200× applied,
Figure 3 Screenshots of the only major discrepancy observed in the study not related to cervical pathology; a small lymph node metastasis of anovarian carcinoma, which was missed in the whole slide imaging (WSI) evaluation. (A) Low power magnification showing several lymph nodesincluded in the slide; (B) higher magnification showing the two areas of interest located in the centre of the largest node (green squares); (C) highermagnification of the area identified with the large green square in B, showing a lipogranuloma with a large fatty vacuole and giant macrophages inthe periphery; (D) higher magnification of the area identified in the small green square in B, showing a small nest of metastatic carcinoma cells,which were missed in the WSI evaluation.
Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524 37
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
indicating that a higher magnification does not seem to be rele-vant for most cases. This scanning strategy which is used inmost validation studies,15 17 19–21 23 28 significantly saves scan-ning time and storage requirements. However, when WSI is rou-tinely used for primary diagnosis it is very likely that thepathologist would require an increase in the scanning magnifica-tion in a small percentage of cases to safely establish the diagno-sis. Although no formal timing was performed, the pathologistdoing the WSI evaluation perceived the diagnostic process to bea little slower.
The main strength of our study is that it is the largest valid-ation study focused on gynaecological pathology that includes asufficient number of cases to allow robust statistical power.Indeed, only one published study has analysed the correlationbetween WSI and CLM in the evaluation of frozen section diag-nosis of 52 ovarian lesions, showing that, as observed in ourstudy, the correlation between CLM and WSI diagnoses is verygood.11 The second strength of our study is that the participat-ing pathologists were aware of the clinical information whichcould affect the diagnostic outcome.
The main limitation of this study is that intraobserver variabil-ity, which is considered the best design to confirm the reprodu-cibility of the results obtained with two different techniques,was not evaluated.38 However, the very good interobserverreproducibility observed in our series suggests that very similardata would be obtained when cases are evaluated by the sameobserver.
A potential advantage of WSI is that it allows performingimage analysis. This may assist in the objective evaluation of thesize of the tumours and the depth of the invasion, which arerelevant information for tumours of the vulva, cervix and endo-metrium. Eventually, this may even permit computer-assisteddiagnosis that could help improve the diagnosis and decreaseinterobserver variability. Several legal issues have arisen from theuse of WSI for primary diagnosis related to image quality, imagepresentation (monitor quality), storage space, adequate backup,document transfer, patient confidentiality and the confidence ofthe pathologist to sign out a pathology report depending onWSI. Most of these issues will probably be settled in the nearfuture. Several digital pathology vendors are currently seekingapproval from the US Food and Drug Administration to useWSI in primary diagnosis, which will definitely encourage thegeneral use of this technique.
In conclusion, the diagnosis of gynaecological specimensusing WSI shows a high concordance with the results of CLMevaluation. Our results confirm that WSI may safely be used forperforming primary histological diagnosis of gynaecologicalspecimens.
Acknowledgements The authors are grateful to Rosana Millan, MargaritaMainar, Berta Coloma, Ingrid Rubio, Olga Ten, Gemma Laguna, Silvia Moya, andArantxa Sánchez, members of the technical and administrative staff of thedepartment of Pathology, for their support in the scanning of the slides. The authorsthank Antonio Teruel, Anna Rubi, Jaume Barderi, Jose Antonio Collados and EnricVidal, members of the staff of Roche for their technical support. The authors thankDonna Pringle for English revision of the manuscript.
Contributors JO: conception, design and conduct of the study, analysis andwriting of the manuscript, approval of the final version of the manuscript. PC, AS,LR-C: conduct of the study, writing of the manuscript, approval of the final versionof the manuscript. OO: conception, design and writing of the manuscript, approvalof the final version of the manuscript. MdP: conception of the study, the statisticalanalysis and writing of the manuscript, approval of the final version of themanuscript. JR: design of the study, writing of the manuscript, approval of the finalversion of the manuscript.
Competing interests None.
Ethics approval Ethical committee of clinical research of Hospital Clinic ofBarcelona.
Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES1 Al-Janabi S, Huisman A, Van Diest PJ. Digital pathology: current status and future
perspectives. Histopathology 2012;61:1–9.2 Brachtel E, Yagi Y. Digital imaging in pathology—current applications and
challenges. J Biophotonics 2012;5:327–35.3 Pantanowitz L, Valenstein PN, Evans AJ, et al. Review of the current state of whole
slide imaging in pathology. J Pathol Inform 2011;2:36.4 Pantanowitz L, Szymas J, Yagi Y, et al. Whole slide imaging for educational
purposes. J Pathol Inform 2012;3:46.5 Brick KE, Comfere NI, Broeren MD, et al. The application of virtual microscopy in a
dermatopathology educational setting: assessment of attitudes amongdermatopathologists. Int J Dermatol 2014;53:224–7.
6 Carlson AM, McPhail ED, Rodriguez V, et al. A prospective, randomized crossoverstudy comparing direct inspection by light microscopy versus projected images forteaching of hematopathology to medical students. Anat Sci Educ 2014;7:130–4.
7 Leong FJ, McGee JO. Automated complete slide digitization: a medium forsimultaneous viewing by multiple pathologists. J Pathol 2001;195:508–14.
8 Singson RP, Natarajan S, Greenson JK, et al. Virtual microscopy and the Internet astelepathology consultation tools. A study of gastrointestinal biopsy specimens. Am JClin Pathol 1999;111:792–5.
9 Wilbur DC, Madi K, Colvin RB, et al. Whole-slide imaging digital pathology as aplatform for teleconsultation: a pilot study using paired subspecialist correlations.Arch Pathol Lab Med 2009;133:1949–53.
10 Winokur TS, McClellan S, Siegal GP, et al. A prospective trial of telepathology forintraoperative consultation (frozen sections). Hum Pathol 2000;31:781–5.
11 Fallon MA, Wilbur DC, Prasad M. Ovarian frozen section diagnosis: use ofwhole-slide imaging shows excellent correlation between virtual slide and originalinterpretations in a large series of cases. Arch Pathol Lab Med 2010;134:1020–3.
12 Harnden P, Coleman D, Moss S, et al. Evaluation of the use of digital images for anational prostate core external quality assurance scheme. Histopathology2011;59:703–9.
13 Ho J, Parwani AV, Jukic DM, et al. Use of whole slide imaging in surgical pathologyquality assurance: design and pilot validation studies. Hum Pathol 2006;37:322–31.
14 Reyes C, Ikpatt OF, Nadji M, et al. Intra-observer reproducibility of whole slideimaging for the primary diagnosis of breast needle biopsies. J Pathol Inform2014;5:5.
15 Bauer TW, Schoenfield L, Slaw RJ, et al. Validation of whole slide imaging forprimary diagnosis in surgical pathology. Arch Pathol Lab Med 2013;137:518–24.
16 Onega T, Weaver D, Geller B, et al. Digitized whole slides for breast pathologyinterpretation: current practices and perceptions. J Digit Imaging 2014;27:642–8.
17 Camparo P, Egevad L, Algaba F, et al. Utility of whole slide imaging and virtualmicroscopy in prostate pathology. APMIS 2012;120:298–304.
18 Comperat E, Egevad L, Lopez-Beltran A, et al. An interobserver reproducibility studyon invasiveness of bladder cancer using virtual microscopy and heatmaps.Histopathology 2013;63:756–66.
19 Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primary diagnosticsin dermatopathology: a feasibility study. J Clin Pathol 2012;65:152–8.
Take home messages
▸ Interobserver agreement between whole slide imaging (WSI)and conventional light microscopy (CLM) evaluations is verygood, with κ values of over 0.91.
▸ Although the accuracy of WSI diagnosis is good even inusers with limited experience, the reproducibility betweenWSI and CLM improves over time, indicating that increasingexperience with WSI increases the accuracy of the diagnosis.
▸ WSI may safely be used for performing primary histologicaldiagnosis of gynaecological specimens using currentscanning technology and viewing interfaces.
▸ Scanning slides at the magnification of 20× is sufficient toachieve a correct diagnosis in most gynaecological biopsies.This scanning strategy significantly saves scanning time andstorage requirements.
38 Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
20 Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primarydiagnostics of gastrointestinal tract pathology: a feasibility study. Hum Pathol2012;43:702–7.
21 Gilbertson JR, Ho J, Anthony L, et al. Primary histologic diagnosis using automatedwhole slide imaging: a validation study. BMC Clin Pathol 2006;6:4.
22 Cornish TC, Swapp RE, Kaplan KJ. Whole-slide imaging: routine pathologicdiagnosis. Adv Anat Pathol 2012;19:152–9.
23 Bauer TW, Slaw RJ. Validating Whole-Slide Imaging for Consultation Diagnoses inSurgical Pathology. Arch Pathol Lab Med Published Online First: 9 May 2014.
24 Ordi J, Sagasta A, Munmany M, et al. Usefulness of p16/Ki67 immunostainingin the triage of women referred to colposcopy. Cancer Cytopathol2014;122:227–35.
25 Landis JR, Koch GG. The measurement of observer agreement for categorical data.Biometrics 1977;33:159–74.
26 Krishnamurthy S, Mathews K, McClure S, et al. Multi-institutional comparison ofwhole slide digital imaging and optical microscopy for interpretation ofhematoxylin-eosin-stained breast tissue sections. Arch Pathol Lab Med2013;137:1733–9.
27 Rodriguez-Urrego PA, Cronin AM, Al-Ahmadie HA, et al. Interobserver andintraobserver reproducibility in digital and routine microscopic assessment ofprostate needle biopsies. Hum Pathol 2011;42:68–74.
28 Al-Janabi S, Huisman A, Nikkels PG, et al. Whole slide images for primarydiagnostics of paediatric pathology specimens: a feasibility study. J Clin Pathol2013;66:218–23.
29 Nelson D, Ziv A, Bandali KS. Going glass to digital: virtual microscopy as asimulation-based revolution in pathology and laboratory science. J Clin Pathol2012;65:877–81.
30 Creagh T, Bridger JE, Kupek E, et al. Pathologist variation in reporting cervicalborderline epithelial abnormalities and cervical intraepithelial neoplasia. J ClinPathol 1995;48:59–60.
31 de Vet HC, Koudstaal J, Kwee WS, et al. Efforts to improve interobserver agreementin histopathological grading. J Clin Epidemiol 1995;48:869–73.
32 McCluggage WG, Walsh MY, Thornton CM, et al. Inter- and intra-observer variationin the histopathological reporting of cervical squamous intraepithelial lesions usinga modified Bethesda grading system. Br J Obstet Gynaecol 1998;105:206–10.
33 McCluggage WG, Bharucha H, Caughley LM, et al. Interobserver variation in thereporting of cervical colposcopic biopsy specimens: comparison of grading systems.J Clin Pathol 1996;49:833–5.
34 Robertson AJ, Anderson JM, Beck JS, et al. Observer variability in histopathologicalreporting of cervical biopsy specimens. J Clin Pathol 1989;42:231–8.
35 Stoler MH, Schiffman M. Interobserver reproducibility of cervical cytologic andhistologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study.JAMA 2001;285:1500–5.
36 Bergeron C, Ordi J, Schmidt D, et al. Conjunctive p16INK4a testing significantlyincreases accuracy in diagnosing high-grade cervical intraepithelial neoplasia. Am JClin Pathol 2010;133:395–406.
37 Darragh TM, Colgan TJ, Thomas CJ, et al. The Lower Anogenital Squamous TerminologyStandardization project for HPV-associated lesions: background and consensusrecommendations from the College of American Pathologists and the American Societyfor Colposcopy and Cervical Pathology. Int J Gynecol Pathol 2013;32:76–115.
38 Pantanowitz L, Sinard JH, Henricks WH, et al. Validating whole slide imaging fordiagnostic purposes in pathology: guideline from the College of AmericanPathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med2013;137:1710–22.
Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524 39
Original article
group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from
Microscopía virtual en el diagnóstico rutinario y la docencia
[72]
Tesis Doctoral. Adela Saco Álvarez
[73]
Estudio número 3
“Validation of Whole-Slide Imaging in the Primary Diagnosis of
Liver Biopsies in a University Hospital”
Adela Saco, Alba Diaz, Monica Hernandez, Daniel Martinez, Carla
Montironi, Paola Castillo, Natalia Rakislova, Marta del Pino, Antonio
Martinez, Jaume Ordi
Dig Liver Dis. 2017 Jul 19. pii: S1590-8658(17)30977-5.
doi: 10.1016/j.dld.2017.07.002. [Epub ahead of print]
Factor de impacto (2016): 2.875
Ranking (2016): 35/134, Segundo cuartil
Microscopía virtual en el diagnóstico rutinario y la docencia
[74]
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
Digestive and Liver Disease xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Digestive and Liver Disease
journa l homepage: www.e lsev ier .com/ locate /d ld
Liver, Pancreas and Biliary Tract
Validation of whole-slide imaging in the primary diagnosis of liverbiopsies in a University Hospital
Adela Saco a,1, Alba Diaz a,1, Monica Hernandez a, Daniel Martinez a, Carla Montironi a,Paola Castillo a,b, Natalia Rakislova a, Marta del Pino b,d, Antonio Martinez a,b,Jaume Ordi a,b,c,∗
a Department of Pathology, Hospital Clínic, Barcelona, Spainb ISGlobal, Barcelona Ctr. Int. Health Res. (CRESIB), Hospital Clínic – Universitat de Barcelona, Barcelona, Spainc University of Barcelona, School of Medicine, Barcelona, Spaind Institute of Gynecology, Obstetrics and Neonatology, Hospital Clínic – Institut dıInvestigacions Biomèdiques August Pi I Sunyer (IDIBAPS), Faculty ofMedicine, University of Barcelona, Spain
a r t i c l e i n f o
Article history:Received 30 January 2017Received in revised form 11 June 2017Accepted 11 July 2017Available online xxx
Keywords:Digital pathologyHepatic needle biopsiesIntra-observer agreementLiver pathology
a b s t r a c t
Background: Experience in the use of whole slide imaging (WSI) for primary diagnosis is limited and thereare no comprehensive reports evaluating this technology in liver biopsy specimens.Aims: To determine the accuracy of interpretation of WSI compared with conventional light microscopy(CLM) in the diagnosis of needle liver biopsies.Methods: Two experienced liver pathologists blindly analyzed 176 consecutive biopsies from the Pathol-ogy Department at the Hospital Clinic of Barcelona. One of the observers performed the initial evaluationwith CLM, and the second evaluation with WSI, whereas the second observer performed the first eval-uation with WSI and the second with CLM. All slides were digitized in a Ventana iScan HT at 400× andevaluated with the Virtuoso viewer (Roche diagnostics). We used kappa statistics (�) for two observations.Results: Intra-observer agreement between WSI and CLM evaluations was almost perfect (96.6%, � = 0.9;95% confidence interval [95% CI]: 0.9–1 for observer 1, and 90.3%, � = 0.9; 95%CI: 0.8-0.9 for observer 2).Both native and transplantation biopsies showed an almost perfect concordance in the diagnosis.Conclusion: Diagnosis of needle liver biopsy specimens using WSI is accurate. This technology can reliablybe introduced in routine diagnosis.
© 2017 Published by Elsevier Ltd on behalf of Editrice Gastroenterologica Italiana S.r.l.
1. Introduction
Conventional light microscopy (CLM) has been the basic and,until recently, the only tool for the histological diagnosis of biopsyspecimens. The development of the whole-slide imaging (WSI)technology has started to change this picture in the last few years.
The basis of the WSI technology is the use of high throughputscanners able to create high quality digital reproductions of glassslides containing a complete histological section and WSI viewersthat allow navigation across the virtual slide. These tools enable the
∗ Corresponding author at: Department of Pathology, Hospital Clínic – ISGlobal,Barcelona Ctr. Int. Health Res. (CRESIB), University of Barcelona Faculty of Medicine,Barcelona, Spain, C/Villarroel 170, 08036, Barcelona, Spain.
E-mail address: [email protected] (J. Ordi).1 These authors equally contributed to the work, and should share co-primary
authorship.
use of the computer as a CLM. WSI has many practical applicationsthat include education and teleconsultation [1–4]. In the last fewyears the medical community has shown increasing interest in theuse of WSI for routine primary diagnosis [5–7].
Indeed, routine pathological diagnosis can benefit from theadvantages of this technology. The WSI workstations are moreergonomic and facilitate a more efficient sign-out process. WSIallows viewing several slides at the same time on the same screen,which is particularly helpful for the evaluation of immuno- or his-tochemically stained slides that can be analyzed together withhematoxylin-eosin (H&E) staining (Fig. 1). The digital viewersincorporate tools that enable making annotations, rotating theimages and making precise measurements [8]. WSI has a muchlarger field of vision than CLM and a wider range of magnifica-tions, including very low magnifications that are very useful forthe evaluation of surgical specimens. WSI facilitates sharing imagesand information with clinicians and other pathologists. This is not
http://dx.doi.org/10.1016/j.dld.2017.07.0021590-8658/© 2017 Published by Elsevier Ltd on behalf of Editrice Gastroenterologica Italiana S.r.l.
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
2 A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx
Fig. 1. The WSI viewer may simultaneously show and synchronously move several slides of a case. This is particularly helpful in the evaluation of liver biopsy specimenssince it allows the analysis of an H&E stained slide together with histochemically and/or immunohistochemically stained slides.
only extremely useful in tumors boards, but also allows expert tele-consultation of difficult cases and frozen section intra-operativebiopsies [9,10]. Finally, with WSI algorithms can be used for theevaluation and quantification of immuhistochemical stains, result-ing in a more objective evaluation [11–14]. This tool is likely tobecome essential to achieve standardized diagnoses in the nearfuture.
Although WSI is considered to be comparable to CLM, adequatecorrelation between WSI and CLM diagnoses should be confirmedbefore this technology is used for primary diagnosis. The num-ber of studies aimed at validating WSI in the routine diagnosis ofthe different areas of pathology is rapidly growing [15]. However,whereas relatively abundant information is available in some areasof pathology, validation studies are very scant or even absent inother areas, such as liver biopsy. Indeed, while a few studies haveused this tool in research and automated image analysis [16–24],there is a complete absence of studies validating the use of WSI inneedle liver biopsies, which may lead to reluctance in implement-ing this technology in routine diagnosis.
2. Materials and methods
2.1. Characteristics of the institution
This study was performed at the Department of Pathology inthe Hospital Clinic (Barcelona, Spain). This department is com-posed of 16 pathologists, 8 residents and a variable number offellows. The specimens are divided into 14 subspecialties, and thepathologists limit their practice to one or two areas. In 2015 theDepartment handled 43,678 specimens with 11,081 paraffin blocks.The number of liver needle biopsy specimens during this year was230. The study was approved by the institutional ethics reviewboard/HCB/2014/0514.
2.2. Sample size calculation
The highest rate of discrepancy between the original diagnosisby CLM and that by WSI was calculated to be 3%, with a non-inferiority margin for WSI review of 5%. A 1-sided binomial testwas used for comparison at a level of significance of .05. The powerto be achieved was 70%, and the level of significance was .05. Basedon these assumptions, it was calculated that 100 cases would needto be reviewed to establish non-inferiority [25].
2.3. Specimens included in the study
All consecutive needle liver biopsy specimens received at theDepartment of Pathology of the Hospital Clinic in a 9-monthperiod (February–October 2015) and assigned to the same expertpathologist were included in the study (n = 176). This represented76.5% of the total number of liver biopsies evaluated in 2015. Allcases had a single paraffin block, containing one to five speci-mens (median 1). All specimens were routinely stained with H&E,Masson’s trichrome and reticulin stain. Additionally, immunohis-tochemical stains were used for specific cases after the request ofthe pathologist. The total number of scanned slides was 1286. Thebiopsies included both native and transplanted livers (n = 112 andn = 64, respectively). The median age of the patients was 57 years(range 18–91).
2.4. Scanning process and characteristics of the WSI display
All the needle liver biopsies were scanned daily after CLM diag-nosis. The scanning of the histological slides was performed on aVentana iScan HT (Ventana Medical Systems, Tucson, AZ, USA) ata magnification of 400x. The scanning process run automatically,and includes the selection of the area that contains the tissue, thedetermination of the focus points, the calibration, and the scanning.
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx 3
When more than one section are mounted on a single slide, the sys-tem scans all the sections. No specific quality control of the slidesscanned was made prior to evaluation by the pathologist. The WSIproduced are stored in a dedicated mass storage environment andlinked to the pathology report, based on the recognition of a QRcode on the slide label. Although WSI can be accessed through thepathology laboratory information system (LIS) software (Novopath,Vitrosoft SL, Sevilla, Spain), for the purposes of this study access tothe WSI was made through the viewer.
The images were viewed with the Virtuoso viewer (Ventana),which works as a web browser and simulates a CLM. The imagesare shown using the same structure provided by the LIS. No specificsoftware installation is required to visualize the WSI. The scannedimages can be viewed up to a real magnification of 400× and upto 600× with a digital zoom, are always in focus, and have anoptimized contrast and adjusted illumination. The viewer showsa thumbnail of the whole slide, which allows confirmation that allthe material present on the glass slide has been included in the dig-ital image and helps in the navigation through the slide. The WSIare displayed on a 30′′ Coronis fusion MDC4130 monitor which hasa resolution of 4 Megapixels (Barco Electronic Systems, Barcelona,Spain).
2.5. CLM and WSI diagnosis
Two experts in liver pathology analyzed all cases. The firstobserver performed the initial evaluation with CLM, which wasconsidered the reference for diagnostic attribution, and the secondobservation with WSI, whereas the second observer performed theinitial evaluation with WSI and the second with CLM. In order toavoid interference with the first diagnosis, the minimum washoutperiod between the two observations was 1.5 months (range 1.5–4months). All the histological slides of each case were scanned andevaluated. A summary with the basic original clinical informationwas provided to the pathologist in both evaluations in order toemulate the real clinical environment, and the pathologist couldalso request additional information. When performing the WSI, thepathologist was blinded to the original diagnostic report, as well asto the evaluation made by the other pathologist. In all cases, a maindiagnosis was rendered in both evaluations. Additionally, in somespecimens additional secondary diagnoses were also provided.
2.6. Concordance between CLM and WSI diagnosis and featuresevaluated
An independent pathologist not involved with the evaluationcompared the original CLM and the WSI-based evaluations andjudged the concordance of the two diagnoses. Concordance wasclassified as: a) complete agreement; b) minor discrepancy (slightdifferences which would not have any clinical or prognostic impli-cations); and c) major discrepancy (differences with clinical and/orprognostic implications for the patient).
Some histological features were routinely evaluated in all thespecimens: portal fibrosis (using a 0–4 scale), presence or absenceof Mallory-Denk bodies, steatosis (using a 0–3 scale), and liver cellballooning [26,27]. Portal inflammation, cholangitis and endothe-litis were estimated using a 0–3 scale in all the acute rejectionbiopsies [28]. In the cases of cirrhosis and chronic hepatitis,necro-inflammatory activity (portal/periportal and lobular) wasevaluated with a 0–3 scale [29–32]. Other characteristics wererecorded when present.
2.7. Statistical analysis
The SPSS (SPSS IncTM140, Version 18, Chicago, IL, USA) wasused for statistical analyses. The results for categorical variables
are expressed as absolute numbers and percentages and 95% con-fidence intervals (95% CI). The Chi-squared or the Fisher’s exacttests were used to compare the variables. The results were eval-uated by unweighted Kappa statistics for two observations. Thismeasure calculates the degree of agreement in classification overthat which would be expected by chance and is scored as a num-ber between 0 and 1. According to the Landis-defined categoriesthe strength of agreement of the Kappa values (�) is: 0 none,beyond chance; 0–0.20 slight; 0.21–0.40 fair; 0.41–0.60 moder-ate; 0.61–0.80 substantial; 0.81–1.00 almost perfect. For the kappavalue calculation, the main diagnoses were grouped into nine cat-egories for the native livers and six categories for the transplantedlivers. The diagnostic categories for the native livers included: a)slight changes (including isolated steatosis), b) venous congestion,c) autoimmune diseases (autoimmune hepatitis and primary bil-iary cirrhosis), d) steatohepatitis, e) acute hepatitis, f) chronic viralhepatitis g) cirrhosis, h) malignant tumors (primary or metastatic),and i) other diseases. The diagnostic categories for the transplantedlivers included: a) slight changes; b) autoimmune hepatitis, c)steatohepatitis, d) hepatitis C virus infection, e) acute cellular rejec-tion, and f) chronic rejection.
3. Results
3.1. Intra-observer and inter-observer agreement
The overall intra-observer agreement between the CLM and theWSI diagnoses was 96.6% (� = 0.9; 95% CI: 0.9–1) for the observer 1and 90.3% (� = 0.9; 95% CI: 0.8–0.9) for the observer 2. There werefour minor discrepancies between the CLM and the WSI diagnosesfor observer 1 and 14 for observer 2. None of the discrepancieswere related to a poor quality of the WSI image or to insufficientmagnification. The diagnoses of carcinoma showed 100% concor-dance in all four evaluations. The overall inter-observer agreementbetween the CLM diagnoses performed by observer 1 and 2 was92.6% (� = 0.9; 95% CI: 0.9–1) and 96.6% (� = 1; 95% CI: 0.9–1) forthe WSI diagnoses.
The intra-observer agreement between CLM and WSI for thenative liver biopsies (n = 112) was 95.5% (� = 0.9; 95% CI: 0.9–1) forobserver 1 and 90.2% (� = 0.9; 95% CI: 0.8–0.9) for observer 2. Theinter-observer agreement between the CLM diagnoses performedby observer 1 and 2 in this group of native liver biopsies was 92.9%(� = 0.9; 95% CI: 0.9–1) and 87.5% (� = 0.9; 95% CI: 0.8–0.9) for theWSI diagnoses. Table 1 shows the intra-observer (WSI vs. CLM)for the two observers and the inter-observer agreement for CLMand WSI for each specific diagnostic group for the 112 native liverspecimens.
Table 2 shows the intra-observer (WSI vs. CLM) for the twoobservers and the inter-observer agreement for CLM and WSI foreach specific diagnostic group for the 64 transplantation biopsies.The inter-observer agreement between the CLM diagnoses per-formed by observer 1 and 2 in this group of transplanted liverbiopsies was 89.1% (� = 0.9; 95% CI: 0.8–1) for observer 1 and 87.5%(� = 0.8; 95% CI: 0.7–0.9) for observer 2. The inter-observer agree-ment between the CLM diagnoses performed by observer 1 and 2was in this group of transplanted liver biopsies was 89.1% (� = 0.8;95% CI: 0.7–1) and 87.5% (� = 0.8; 95% CI: 0.7–0.9) for the WSI diag-noses.
3.2. Agreement between WSI and CLM in relevant liver changes
Table 3 shows the intra-observer (WSI vs. CLM) and the inter-observer (WSI vs. WSI and CLM vs. CLM) agreement in theevaluation of relevant histological features in the native livers(steatosis, liver cell ballooning, presence or absence or Mallory-
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
4 A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx
Table 1Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the diagnosis ofnative liver specimens (n = 112).
Diagnosis n Intra-observeragreement
Kappa value(95% CI)
Intra-observeragreement
Kappa value(95% CI)
Inter-observeragreement CLM
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Observer 1 Observer 2 WSI
Mild changesa 22 98.3 0.9 (0.9–1) 94.9 0.9 (0.8–0.9) 96.0 0.9 (0.8–1) 96.0 0.9 (0.8–1)Venouscongestion
4 100 1 (NA) 98.9 0.9 (0.5–1) 98.3 0.7 (0.4–1) 99.4 0.9 (0.7–1)
Autoimmunediseasesb
15 99.4 1 (0.9–1.0) 98.3 0.9 (0.8–1) 98.9 0.9 (0.8–1) 97.7 0.9 (0.7–1)
Steatohepatitis 12 98.3 0.9 (0.8–1) 97.7 0.8 (0.7–1) 97.7 0.8 (0.7–1) 96.0 0.7 (0.6–0.9)Acutehepatitisc
7 99.4 0.9 (0.8–1) 98.3 0.8 (0.6–1) 99.4 0.9 (0.8–1) 98.3 0.8 (0.6–1)
Chronichepatitisd
14 97.7 0.9 (0.9–1) 94.9 0.8 (0.7–0.9) 97.7 0.9 (0.9–1) 93.7 0.8 (0.7–0.9)
Cirrhosis 19 100 1 (NA) 100 1 (NA) 98.9 0.9 (0.8–1) 98.9 0.9 (0.8–1)Tumorse 14 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)Otherf 5 100 1 (NA) 98.9 0.8 (0.6–1) 99.4 0.9 (0.8–1) 99.4 0.9 (0.8–1)
95% CI: 95% confidence interval.NA: not applicable.
a Includes mild to moderate macrovesicular steatosis (n = 15) and mild non-specific changes (n = 7).b Includes autoimmune hepatitis (7 cases) and primary biliary cirrhosis (8 cases).c Includes four toxic acute hepatitis, two acute B hepatitis and one hepatitis of unknown origin.d Includes 10 chronic hepatitis C, two chronic hepatits B and two drug induced hepatitis.e Includes one cholangiocarcinoma, six metastatic carcinomas, six hepatocellular carcinomas and one hemangioma.f Includes one case each of schistosomiasis, cystic fibrosis, graft versus host disease, sclerosing cholangitis and nodular regenerative hyperplasia.
Table 2Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the diagnosis ofliver transplantation biopsies (n = 64).
Diagnosis n Intra-observeragreement
Kappa value(95% CI)
Intra-observeragreement
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Observer 1 Observer 2 CLM WSI
Mild changes 23 95.3 0.9 (0.8–1) 92.2 0.8 (0.7–1) 93.7 0.9 (0.7–1) 90.6 0.8 (0.6–1)Autoimmunehepatitis
2 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)
Steatohepatitis 1 100 1 (NA) 100 1 (NA) 98.4 7 (0.0–1) 98.4 0.7 (0.0–1)Chronichepatitisa
20 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)
Acute cellularrejection
15 95.3 0.9 (0.7–1) 89.1 0.7 (0.5–0.9) 90.6 0.7 (0.6–0.9) 93.7 0.8 (0.7–1)
Chronicrejection
1 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)
Other lesionsb 2 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)
95%CI: 95% confidence interval; NA: not applicable.a Hepatitis C virus reinfection.b Includes one case of preservation injury and one insufficient biopsy.
Denk bodies, portal/peri-portal inflammatory activity and necrosisand lobular necrosis and inflammatory activity). Table 4 shows theintra-observer (WSI vs. CLM) and the inter-observer (WSI vs. WSIand CLM vs. CLM) agreement in the evaluation of major histologi-cal features in transplanted livers (portal inflammation, cholangitisand endothelitis).
4. Discussion
This is the first study evaluating the accuracy of WSI diagnosis inthe routine practice of needle liver biopsies. Our results show a highintra-observer concordance between the CLM and the WSI evalua-tions (over 90% for both obseervers) in the diagnoses of a large seriesof routine needle liver specimens. The kappa value, considered asa measure of the level of intra- and inter-observer agreement cor-rected by chance, was almost perfect (0.9 for both observers) and0.9–1 for the inter-oberver comparisons of the CLM and the WSIevaluations. The percentage of discrepancies between the CLM andWSI diagnoses observed in our study was below 10%, and onlyminor discrepancies were identified. Neither had an impact onpatient management. More importantly, none of the discrepancies
was related to a poor quality of the WSI image or to insufficientmagnification. All the discrepancies observed were associated witheither the small size of the material or to the intrinsic difficulty ofthe case. Thus, our results confirm that WSI may confidently beused for primary histological diagnosis of liver biopsies.
A number of studies have shown that there is a substantialvariation between and within observers in the evaluation of liverbiopsy specimens. These studies are limited to specific diseasessuch as non-alcoholic steatohepatitis and chronic viral hepatitis[33–37]. In an intra-observer concordance study including 50 biop-sies oriented as non-alcoholic steatohepatitis Kleiner et al. reporteda kappa value of 0.61 [34]. Our study showed a higher rate of con-cordance in the evaluation of steatohepatitis with a kappa valueranging from 0.7 to 0.9 in the different comparisons, althoughthe number of cases with this diagnosis was much lower andincluded both alcoholic and non-alcoholic steatohepatitis. Threestudies have evaluated intra-observer concordance in the diagno-sis of chronic viral hepatitis. The evaluation of fibrosis grade andstage in these studies showed kappa values ranging from 0.72 and1 [35–37], which were comparable with the concordance ratesobserved in our study (0.7–0.9). These discrepancies have mainlybeen attributed to the inherent intra-observer variability in the
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx 5
Table 3Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the evaluationof major histological changes in the native livers (n = 112).
Histologicalfeature
Intra-observeragreement
Kappa value(95% CI)
Intra-observeragreement
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Observer 1 Observer 2 CLM WSI
Fibrosisa 90.9 0.8 (0.8–0.9) 92.6 0.9 (0.8–0.9) 82.9 0.7 (0.6–0.9) 90.3 0.8 (0.8–0.9)Steatosisb 91.1 0.9 (0.8–0.9) 99.1 1 (NA) 92.9 0.9 (0.8–1) 92.0 0.9 (0.8–0.9)Liver cellballooningc
98.2 0.9 (0.9–1) 100 1 (NA) 96.4 0.9 (0.8–1) 98.2 0.9 (0.9–1)
Mallory-Denkbodiesc
98.2 0.9 (0.8–1) 99.5 1.0 (0.9–1.0) 97.3 0.9 (0.7–1) 100 1 (NA)
Portal/peri-portalinflammatoryactivity andnecrosis c,d
92.9 0.8 (0.6–0.9) 92.0 0.8 (0.7–0.9) 78.6 0.5 (0.3–0.6) 84.8 0.6 (0.5–0.8)
Lobularnecrosis andinflammatoryactivity c,d
96.4 0.9 (0.8–1) 93.7 0.9 (0.7–1.0) 87.5 0.7 (0.5–0.8) 86.6 0.7 (0.5–0.8)
95% CI: 95% confidence interval.a Graded on a 0–4 scale.b Graded on a 0–3 scale.c Evaluated as absent or present.d Portal/peri-portal inflammatory activity and necrosis and lobular necrosis and inflammatory activity were evaluated only in the cases with a diagnosis of cirrhosis (n = 19)
and chronic hepatitis (14 in native livers and 20 in transplanted livers).
Table 4Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the evaluationof the main histological features in the transplanted livers (n = 64).
Histologicalfeature
Intra-observeragreement
Kappa value(95% CI)
Intra-observeragreement
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Inter-observeragreement
Kappa value(95% CI)
Observer 1 Observer 2 CLM WSI
Portalinflammation
100 1 (NA) 90.6 0.8 (0.6–0.9) 85.9 0.7 (0.5–0.9) 95.3 0.9 (0.8–1)
Cholangitis 96.9 0.9 (0.8–1) 98.4 1 (0.9–1) 95.3 0.9 (0.7–1) 92.2 0.8 (0.6–1)Endothelitis 96.9 0.9 (0.8–1) 93.7 0.8 (0.8–1) 93.7 0.8 (0.7–1) 95.3 0.9 (0.7–1)
95% CI: 95% confidence interval; All features were graded on a 0–3 scale.
diagnosis of needle liver biopsy specimens. Interestingly, some ofthese studies analyzed a number of histological features separately,showing high concordance rates for steatosis (� = 0.79), periportalnecrosis (� = 0.74) and fibrosis (� = 0.86) and lower values for lobu-lar necrosis (� = 042) [33,34]. In the present study, the concordanceobserved for all these histological findings showed even betterresults. A possible limitation of our study is the lower number ofcases of each particular disease included in the analysis comparedto previous reports [33–37]. However, our study was designed toevaluate the reliability of the WSI tool for the diagnosis of any liverlesion and not specifically for a single disease.
Interestingly, the specific analysis of liver transplantation spec-imens (n = 64) showed a high intra-observer concordance thatremained almost perfect (93.7%; � = 0.9 for observer 1, 87.5, � = 0.8for observer 2). There were no differences in the diagnosis of rejec-tion.
The results obtained in our study with the liver biopsies arecomparable to other validation studies conducted in other areas ofpathology, such as breast [38], skin [39], gastrointestinal [40,41],prostate [42–46], gynecological [25], renal [46,47] or pediatricpathology [48,49] which show similar high rates of concordancebetween CLM and WSI diagnoses. Thus, the results of all thesestudies indicate that WSI should be considered as a validated tool,almost equivalent to the CLM. In keeping with this assumption,the guidelines and recommendations of the College of Ameri-can Pathologists, the Canadian Association of Pathologists and theAmerican Telemedicine Association for adequate validation of WSIbefore its use in routine diagnosis do not require a validation foreach specific area [1,4]. These recommendations indicate that only
60 samples per pathologist should be evaluated in order to ensurethe familiarity of the pathologist with the new tool. Indeed, as withany other tool, there is a learning curve for WSI [25,50–54].
Remarkably, the pathologist did not report any difficulty in ren-dering the diagnosis at the magnification used in this study (400×).A 200× magnification is considered as appropriate to achieve acorrect diagnosis in most previously published studies evaluatingother areas of pathology [25,42,48,50,55–57]. However, this scan-ning magnification may not be sufficient for some areas, such asthe liver due to the small size of the specimens and the need toevaluate subtle changes that frequently require the use of highmagnification.
The introduction of the WSI technology may significantlyimprove the diagnosis of routine needle liver biopsy specimens tak-ing into account the advantage of the possibility of viewing multipleslides at the same time with this technique. Indeed, this advan-tage can be very useful in liver pathology since several stains areoften used and WSI facilitates tele-consultation. Finally, the futuredevelopment of computer-assisted diagnostic algorithms is likelyto help reduce intra- and inter-observer variability. However, manyissues should be addressed to make this implementation feasibleand cost-efficient, such as the cost of the scanners [4,8,58–61], thecosts associated with the maintenance of the system and the stor-age of the images and legal issues related to the use of WSI forprimary diagnosis, including image storage and patient confiden-tiality. Approval is currently being sought from the US Food andDrug Administration (FDA) for the use of WSI in primary diagnosis.In the meantime, WSI is being increasingly used in several centersaround the world.
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
6 A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx
In conclusion, the diagnosis of needle liver biopsies using WSIhas high intra-observer concordance with the results of CLM eval-uation. Our results confirm that WSI can be safely used for primaryhistological diagnosis of liver biopsies, including native and trans-plantation specimens.
Conflict of interest statementAll the authors have read and approved the manuscript, take publicresponsibility for its content, and consent the title, the authorship,and the contents of the paper. Neither the article nor any part of ithas been published or submitted elsewhere. There are no conflict ofinterest or commercial interests with this work. Should the paperbe accepted for publication, we authorize transferring the copyrightto Digestive and Liver Disease.
References
[1] Pantanowitz L, Sinard JH, Henricks WH, et al. Validating whole slide imagingfor diagnostic purposes in pathology: guideline from the College of AmericanPathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med2013;137:1710–22.
[2] Wilbur DC, Madi K, Colvin RB, et al. Whole-slide imaging digital pathology as aplatform for teleconsultation: a pilot study using paired subspecialist correla-tions. Arch Pathol Lab Med 2009;133:1949–53.
[3] Saco A, Bombi JA, Garcia A, et al. Current status of whole-slide imaging ineducation. Pathobiology 2016;83:79–88.
[4] Bernard C, Chandrakanth SA, Cornell IS, et al. Guidelines from the CanadianAssociation of Pathologists for establishing a telepathology service for anatomicpathology using whole-slide imaging. J Pathol Inform 2014;5:15.
[5] Al-Janabi S, Huisman A, Van Diest PJ. Digital pathology: current status andfuture perspectives. Histopathology 2012;61:1–9.
[6] Brachtel E, Yagi Y. Digital imaging in pathology–current applications and chal-lenges. J Biophotonics 2012;5:327–35.
[7] Pantanowitz L, Valenstein PN, Evans AJ, et al. Review of the current state ofwhole slide imaging in pathology. J Pathol Inform 2011;2:36.
[8] Thorstenson S, Molin J, Lundstrom C. Implementation of large-scale routinediagnostics using whole slide imaging in Sweden: digital pathology experi-ences 2006-2013. J Pathol Inform 2014;5:14.
[9] Hartman DJ, Parwani AV, Cable B, et al. Pocket pathologist: a mobile applicationfor rapid diagnostic surgical pathology consultation. J Pathol Inform 2014;5:10.
[10] Speiser JJ, Hughes I, Mehta V, et al. Mobile teledermatopathology: using a tabletPC as a novel and cost-efficient method to remotely diagnose dermatopathol-ogy cases. Am J Dermatopathol 2014;36:54–7.
[11] Gavrielides MA, Conway C, O’Flaherty N, et al. Observer performance in theuse of digital and optical microscopy for the interpretation of tissue-basedbiomarkers. Anal Cell Pathol (Amst) 2014;2014:157308.
[12] Nassar A, Cohen C, Agersborg SS, et al. A multisite performance study comparingthe reading of immunohistochemical slides on a computer monitor with con-ventional manual microscopy for estrogen and progesterone receptor analysis.Am J Clin Pathol 2011;135:461–7.
[13] Micsik T, Kiszler G, Szabo D, et al. Computer aided semi-automated evaluationof HER2 immunodetection–a robust solution for supporting the accuracy ofanti HER2 therapy. Pathol Oncol Res 2015;21:1005–11.
[14] Krenacs T, Zsakovics I, Diczhazi C, et al. The potential of digital microscopy inbreast pathology. Pathol Oncol Res 2009;15:55–8.
[15] Saco A, Ramirez J, Rakislova N, et al. Validation of whole-slide imaging forhistolopathogical diagnosis: current state. Pathobiology 2016;83:89–98.
[16] Atupelage C, Nagahashi H, Kimura F, et al. Computational hepatocellular car-cinoma tumor grading based on cell nuclei classification. J Med Imaging(Bellingham, Wash) 2014;1:34501.
[17] Bejnordi BE, Litjens G, Timofeeva N, et al. Stain specific standardiza-tion of whole-slide histopathological images. IEEE Trans Med Imaging2016;35:404–15.
[18] Isse K, Grama K, Abbott IM, et al. Adding value to liver (and allograft) biopsyevaluation using a combination of multiplex quantum dot immunostaining,high-resolution whole-slide digital imaging, and automated image analysis.Clin Liver Dis 2010;14:669–85.
[19] Abe T, Hashiguchi A, Yamazaki K, et al. Quantification of collagen and elas-tic fibers using whole-slide images of liver biopsy specimens. Pathol Int2013;63:305–10.
[20] Liang Y, Wang F, Treanor D, et al. A framework for 3D vessel analysis using wholeslide images of liver tissue sections. Int J Comput Biol Drug Des 2016;9:102–19.
[21] Liang Y, Wang F, Treanor D, et al. Liver whole slide image analysis for 3D vesselreconstruction. Proc IEEE Int Symp Biomed Imaging 2015;2015:182–5.
[22] Nagase A, Takahashi M, Nakano M. Automatic calculation and visualization ofnuclear density in whole slide images of hepatic histological sections. BiomedMater Eng 2015;26(Suppl. 1):S1335–44.
[23] Rawlins SR, El-Zammar O, Zinkievich JM, et al. Digital quantification is moreprecise than traditional semiquantitation of hepatic steatosis: correlation withfibrosis in 220 treatment-naive patients with chronic hepatitis C. Dig Dis Sci2010;55:2049–57.
[24] Hall AR, Dhillon AP, Green AC, et al. Hepatic steatosis estimated microscopicallyversus digital image analysis. Liver Int 2013;33:926–35.
[25] Ordi J, Castillo P, Saco A, et al. Validation of whole slide imaging in the primarydiagnosis of gynaecological pathology in a University Hospital. J Clin Pathol2015;68:33–9.
[26] Scheuer PJ. Classification of chronic viral hepatitis: a need for reassessment. JHepatol 1991;13:372–4.
[27] Brunt EM, Janney CG, Di Bisceglie AM, et al. Nonalcoholic steatohepatitis: aproposal for grading and staging the histological lesions. Am J Gastroenterol1999;94:2467–74.
[28] Demetris AJ, Batts KP, Dhilion AP, et al. Banff schema for grading liver allograftrejection: an international consensus document. Hepatology 1997;25:658–63.
[29] Brunt EM. Grading and staging the histopathological lesions of chronichepatitis: the Knodell histology activity index and beyond. Hepatology2000;31:241–6.
[30] Batts KP, Ludwig J. Chronic hepatitis. An update on terminology and reporting.Am J Surg Pathol 1995;19:1409–17.
[31] Desmet VJ, Knodell RG, Ishak KG, et al. Formulation and application of a numer-ical scoring system for assessing histological activity in asymptomatic chronicactive hepatitis [Hepatology 1981;1:431-435]. J Hepatol 2003;38:382–6.
[32] Hubscher SG. Histological grading and staging in chronic hepatitis: clinicalapplications and problems. J Hepatol 1998;29:1015–22.
[33] Rousselet M-C, Michalak S, Dupre F, et al. Sources of variability in histologicalscoring of chronic viral hepatitis. Hepatology 2005;41:257–64.
[34] Kleiner DE, Brunt EM, Van Natta M, et al. Design and validation of a his-tological scoring system for nonalcoholic fatty liver disease. Hepatology2005;41:1313–21.
[35] Regev A, Berho M, Jeffers LJ, et al. Sampling error and intraobserver varia-tion in liver biopsy in patients with chronic HCV infection. Am J Gastroenterol2002;97:2614–8.
[36] Robert M, Sofair AN, Thomas A, et al. A comparison of hepatopathologists’ andcommunity pathologists’ review of liver biopsy specimens from patients withhepatitis C. Clin Gastroenterol Hepatol 2009;7:335–8.
[37] Skripenova S, Trainer TD, Krawitt EL, et al. Variability of grade and stage insimultaneous paired liver biopsies in patients with hepatitis C. J Clin Pathol2007;60:321–4.
[38] Al-Janabi S, Huisman A, Willems SM, et al. Digital slide images for primary diag-nostics in breast pathology: a feasibility study. Hum Pathol 2012;43:2318–25.
[39] Al Habeeb A, Evans A, Ghazarian D. Virtual microscopy using whole-slide imag-ing as an enabler for teledermatopathology: a paired consultant validationstudy. J Pathol Inform 2012;3:2.
[40] Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primary diag-nostics of gastrointestinal tract pathology: a feasibility study. Hum Pathol2012;43:702–7.
[41] Molnar B, Berczi L, Diczhazy C, et al. Digital slide and virtual microscopy basedroutine and telepathology evaluation of routine gastrointestinal biopsy speci-mens. J Clin Pathol 2003;56:433–8.
[42] Camparo P, Egevad L, Algaba F, et al. Utility of whole slide imaging and virtualmicroscopy in prostate pathology. APMIS 2012;120:298–304.
[43] Chargari C, Comperat E, Magne N, et al. Prostate needle biopsy examination bymeans of virtual microscopy. Pathol Res Pract 2011;207:366–9.
[44] Fine JL, Grzybicki DM, Silowash R, et al. Evaluation of whole slide imageimmunohistochemistry interpretation in challenging prostate needle biopsies.Hum Pathol 2008;39:564–72.
[45] Helin H, Lundin M, Lundin J, et al. Web-based virtual microscopy in teachingand standardizing Gleason grading. Hum Pathol 2005;36:381–6.
[46] Al-Janabi S, Huisman A, Jonges GN, et al. Whole slide images for primarydiagnostics of urinary system pathology: a feasibility study. J Ren Inj Prev2014;3:91–6.
[47] Furness P. A randomized controlled trial of the diagnostic accuracy of internet-based telepathology compared with conventional microscopy. Histopathology2007;50:266–73.
[48] Al-Janabi S, Huisman A, Nikkels PGJ, et al. Whole slide images for primarydiagnostics of paediatric pathology specimens: a feasibility study. J Clin Pathol2013;66:218–23.
[49] Arnold MA, Chenever E, Baker PB, et al. The College of American Patholo-gists guidelines for whole slide imaging validation are feasible for pediatricpathology: a pediatric pathology practice experience. Pediatr Dev Pathol2015;18:109–16.
[50] Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primary diagnosticsin dermatopathology: a feasibility study. J Clin Pathol 2012;65:152–8.
[51] Randell R, Ruddle RA, Mello-Thoms C, et al. Virtual reality microscope versusconventional microscope regarding time to diagnosis: an experimental study.Histopathology 2013;62:351–8.
[52] Krishnamurthy S, Mathews K, McClure S, et al. Multi-institutional compari-son of whole slide digital imaging and optical microscopy for interpretationof hematoxylin-eosin-stained breast tissue sections. Arch Pathol Lab Med2013;137:1733–9.
[53] Houghton JP, Ervine AJ, Kenny SL, et al. Concordance between digital pathologyand light microscopy in general surgical pathology: a pilot study of 100 cases.J Clin Pathol 2014;67:1052–5.
[54] Randell R, Ruddle RA, Thomas RG, et al. Diagnosis of major cancer resectionspecimens with virtual slides: impact of a novel digital pathology workstation.Hum Pathol 2014;45:2101–6.
[55] Bauer TW, Schoenfield L, Slaw RJ, et al. Validation of whole slide imaging for pri-mary diagnosis in surgical pathology. Arch Pathol Lab Med 2013;137:518–24.
Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002
ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7
A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx 7
[56] Gilbertson JR, Ho J, Anthony L, et al. Primary histologic diagnosis using auto-mated whole slide imaging: a validation study. BMC Clin Pathol 2006;6:4.
[57] Bauer TW, Slaw RJ. Validating whole-slide imaging for consultation diagnosesin surgical pathology. Arch Pathol Lab Med 2014;138:1459–65.
[58] Hedvat CV. Digital microscopy: past, present, and future. Arch Pathol Lab Med2010;134:1666–70.
[59] Ho J, Ahlers SM, Stratman C, et al. Can digital pathology result in cost savings? Afinancial projection for digital pathology implementation at a large integratedhealth care organization. J Pathol Inform 2014;5:33.
[60] Isaacs M, Lennerz JK, Yates S, et al. Implementation of whole slide imaging insurgical pathology: a value added approach. J Pathol Inform 2011;2:39.
[61] Pantanowitz L. Digital images and the future of digital pathology. J PatholInform 2010;1.
Microscopía virtual en el diagnóstico rutinario y la docencia
[82]
Tesis Doctoral. Adela Saco Álvarez
[83]
Estudio número 4
“Current Status of Whole-Slide Imaging in Education”
Adela Saco, Josep Antoni Bombi, Adriana Garcia, José Ramirez, Jaume Ordi
Pathobiology 2016; 83:79 – 88
Factor de impacto (2016): 1.703
Ranking (2016): 60/193, segundo cuartil
Microscopía virtual en el diagnóstico rutinario y la docencia
[84]
E-Mail [email protected]
Original Paper
Pathobiology 2016;83:79–88 DOI: 10.1159/000442391
Current Status of Whole-Slide Imaging in Education
Adela Saco a Jose Antoni Bombi a Adriana Garcia a Jose Ramírez a
Jaume Ordi a, b
a Department of Pathology, Hospital Clínic, University of Barcelona School of Medicine, and b ISGlobal, Barcelona Center for International Health Research (CRESIB), Barcelona , Spain
shown to be an extremely useful tool for undergraduate ed-ucation (medical, dental and veterinary schools), for the training of residents of pathology, tele-education and in tu-mor boards. © 2016 S. Karger AG, Basel
Introduction and Historical Perspective
Basic skills in histology and pathology are an essential component of the education of undergraduate students, not only at medical schools, but also at schools of den-tistry, veterinary medicine and biology. Residents and fel-lows in pathology demand more advanced training, and certified pathologists require continuing education. This training has to be provided more and more often to par-ticipants located in distant sites. Until a few years ago, the only tool that fulfilled these needs was conventional light microscopy (CLM). CLM was introduced as a diagnostic and teaching tool nearly two centuries ago and became the basis for teaching histology and pathology [1, 2] . However, CLM has many limitations, not only in con-ducting the classes but also in assessing the skills of the students. One of the main disadvantages of using CLM is that it does not allow simultaneous viewing of the slides
Key Words
Medical education · Teaching · Virtual microscopy · Whole-slide imaging
Abstract
Conventional light microscopy (CLM) has classically been the basic tool to teach histology and pathology. In recent years, whole-slide imaging (WSI), which consists of generat-ing a high-magnification digital image of an entire histolog-ical glass slide, has emerged as a useful alternative to CLM offering a myriad of opportunities for education. Navigation through the digitized slides closely simulates viewing glass slides with a microscope and is also referred to as virtual mi-croscopy. WSI has many advantages for education. Students feel more comfortable with its use, and it can be used in any classroom as it only requires a computer with Internet access and it allows remote access from anywhere and from any device. WSI can be used simultaneously by a large number of people, stimulating cooperation between students and improving the interaction with the teachers. It allows mak-ing marks and annotations on specific fields, which enable specific directed questions to the teacher. Finally, WSI sup-ports are cost-effective compared with CLM. Consequently, WSI has begun to replace CLM in many institutions. WSI has
Published online: April 26, 2016
Jaume Ordi Department of Pathology, Hospital Clínic University of Barcelona, C/Villarroel 170 ES–08036 Barcelona (Spain) E-Mail jordi @ clinic.ub.es
© 2016 S. Karger AG, Basel1015–2008/16/0833–0079$39.50/0
www.karger.com/pat
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
Saco/Bombi/Garcia/Ramírez/Ordi
Pathobiology 2016;83:79–88DOI: 10.1159/000442391
80
by multiple students. Secondly, rooms with multiple mi-croscopes are expensive and require continuous mainte-nance. A third important disadvantage of CLM is the need to generate glass slides and store them, which entails significant economic costs and the loss of specimen mate-rial.
In the early 20th century, the emergence of projectors facilitated the training of students because it allowed dis-playing the histological images to a number of partici-pants at the same time. The introduction and develop-ment of initially analogic and later on digital video cam-eras boosted the use of CLM in multiple teaching settings. However, these devices allowed moving the slide only to one person and, consequently, served as a teaching sup-plement but did not allow the replacement of CLM.
In the 1980s, the first digital images were generated from histological slides, although it was not until later, with the advent of personal computers with sufficient memory capacity, that digital microscopy progressed rapidly to the technology we know today. A few years lat-er, imaging converter programs and servers appeared which allowed uploading the virtual slides to the web and permitted viewing the images and zooming [1] .
Whole slide imaging (WSI), also referred to as virtual or digital pathology, consists of generating a virtual image of the entire histological glass slide. The process is per-formed by WSI scanners, robotic microscopes capable to automatically generate digital images from the glass slide. Current scanners are able to use different optical objec-tives (×10, ×20, ×40 and ×60) depending on the specific needs and allow scanning the whole slide or a particular portion of the tissue. Specific software allows viewing dig-ital slides online from computers without the need for a CLM. At present, there are numerous systems prepared to generate good-quality virtual images of histological sections obtained from paraffin blocks or frozen tissue as well as from cytological smears. Image browsing is per-formed with a mouse or a joystick, which allows moving through the different areas of the slide and permits image zooming, thus simulating the optical objectives of a CLM
[3] . Software viewers have multiple tools to make mea-surements and annotations on the images, and may con-tain additional patient information, thus providing a complete view of the cases [4] . The information is ade-quately protected because, although students and resi-dents can access the virtual slides from any computer, ac-cess is controlled by the use of a specific password and registration in the system.
WSI is an extremely useful teaching tool because it al-lows displaying the histological slides on a computer monitor and solves many of the problems of CLM. In-deed, WSI has been used for many different educational activities ( table 1 ). This technology is rapidly expanding as a teaching tool, either as a complement to or a substi-tute of CLM. In many centers, the transition from CLM to WSI has already occurred either gradually or suddenly. This review focuses on the applications of WSI (or vir-tual microscopy) in education.
Advantages of WSI
WSI has brought about an important change in the way of understanding teaching in histology and pathol-ogy, allowing the introduction of some actions and ca-pacities that were previously not possible. The main ad-vantages of WSI are summarized in table 2 .
Some of the advantages of WSI are related to the change in the physical tool itself. WSI allows any com-puter to work as a CLM, and consequently reduces or eliminates the need for conventional microscopes. It has been clearly shown that students feel more comfortable with the use of WSI because they all have prior knowl-edge of computers (almost all students are currently dig-ital natives) [5–10] . An additional advantage of WSI is that the slides are always in focus. Thus, the students get used to the tool very quickly and can immediately con-centrate on the histological features of the slides and do not have to become acquainted with the microscope [6, 9, 10] . WSI can be used in any classroom, because it only requires a computer with Internet access [1, 11] . More-over, WSI allows access from any device, either in or outside the facilities of the institution. This implies that the students can review the histological slides at any time and from anywhere, thereby facilitating their study and eliminating the restrictions of access to the labora-tory of microscopy after class hours [1, 6, 10, 12–14] . As a consequence, laboratories of microscopy, which are very expensive and costly in maintenance, become un-necessary, leading to a significant reduction in costs [9,
Table 1. Major educational uses of WSI
Undergraduate teaching (medicine, veterinary medicine, biologyand dentistry)
Pathology training (residency and fellowship programs)Schools of cytotechnologyTumor boardsTele-education, e-learning and virtual workshops
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
WSI in Education Pathobiology 2016;83:79–88DOI: 10.1159/000442391
81
15] . With an appropriate bandwidth, virtual micro-scopes can be used simultaneously by a large number of people, thereby surpassing the restrictions related to a limited number of CLMs in a laboratory [4, 9, 12, 16, 17] and stimulating cooperation between students [9, 12, 15, 17–19] . Moreover, the interaction between teachers and students is improved by viewing the same image at the same time in the classroom, which allows bringing up questions easily and improves the learning curve [15, 18, 19] .
The second group of advantages of WSI is related to the characteristics of the viewer. The presence of a thumb-nail indicating the area shown on the screen ( fig. 1 ) pro-motes better orientation when the student browses the slide [12] . Digital viewers allow seeing the slide at a very low magnification, which also helps the student to be ori-entated in the tissue. Teaching tools based on WSI allow completing the information of the histological slides with clinical data, imaging studies (conventional radiology, CT, ultrasound and MRI), macroscopic images, as well as
Table 2. Advantages of WSI for teaching purposes
Related to the equipmentStudents feel more comfortable with the use of the virtual microscope because they have prior computer knowledgeIt can be used in any classroom: it only requires a computer with Internet accessIt allows remote access anywhere and from any deviceIt can be simultaneously used by an unlimited number of peopleSeveral students can use the same computer at the same time, stimulating cooperation between studentsIt improves the interaction between teachers and students by viewing the same image at the same timeEliminates the need for investing in the creation and maintenance of microscopic laboratories
Related to the viewerThe thumbnail in the viewer facilitates better orientation when the students browse through the slideThe information of the histological slide can be completed with macroscopic images, immunohistochemical stains, radiologicalimages and clinical dataSeveral slides can be displayed simultaneously on the same screen, facilitating the interpretation of immunohistochemical techniquesMarks and annotations can be made on specific fields, which facilitate specific directed questions to the teacher
Related to the digital slidesNo deterioration in digital slides with time and slides do not have to be replacedHomogeneity in the quality of images available to studentsAdditional histological sections are not neededCases with scant tissue or consultation material can also be studiedFISH or immunofluorescence digital images do not lose fluorescenceOld cases are immediately accessible, without the help of technical staff searching for the files in the archive
Fig. 1. The viewer displays a thumbnail that indicates the area shown on the screen. This promotes better orientation when the student browses the slide.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
Saco/Bombi/Garcia/Ramírez/Ordi
Pathobiology 2016;83:79–88DOI: 10.1159/000442391
82
histochemical or immunohistochemical stains ( fig. 2 ), which, in turn, allow a more complete picture of the study cases [16, 17] . Many viewers allow simultaneously dis-playing several slides on the screen, thus facilitating the interpretation of immunohistochemical techniques by comparing them with the conventional hematoxylin-eo-sin stain ( fig. 3 ) [1] . Another feature of WSI viewers that is particularly helpful for teaching purposes is the possi-bility to make marks and annotations on specific fields of the digital slides to show the key features that allow the recognition of the lesions. This tool facilitates solving spe-cific doubts and improves the interaction between the teacher and the student. With the annotations made by the teacher, the students can easily identify the key areas on a slide, allowing them to focus on the diagnostic clues of a specific lesion. Annotations and marks made by the students allow them to ask the teacher about specific doubts. Indeed, annotations seem to improve the final results of the students, and several studies have shown
that the students who had notes on the slides have better scores than those who do not make any annotation [1, 12, 14, 20–22] .
The third set of advantages of WSI is related to the his-tological slides. Digital slides always have the same qual-ity. They never break, get lost or deteriorate with time. Thus, it is not necessary to replace them by recutting and staining new slides every certain period of time, with the subsequent impact on costs and preservation of the fre-quently highly valuable tissue from the paraffin block [17, 20] . WSI homogenizes the material available to students. All students have exactly the same digital slide, which eliminates the variability in quality among different glass slides. Creating selections of interesting cases to students becomes extremely easy because there is no need for ad-ditional histological sections [10, 17, 22] . The easiness to prepare collections of cases with a high number of slides from each organ or pathology allows randomly showing many different cases to the students. This makes the stu-
Fig. 2. Teaching tools based on WSI allow completing the information of the histological slides with clinical in-formation, imaging studies (conventional radiology, CT, ultrasound and MRI) macroscopic images, as well as histochemical or immunohistochemical stains.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
WSI in Education Pathobiology 2016;83:79–88DOI: 10.1159/000442391
83
dent learn to recognize the histological characteristics of the organ or the lesion, and not to remember the slide because of other features, such as the size or shape of the tissue [17] . These features enhance learning and are espe-cially useful during the evaluation test because the results are totally dependent on the knowledge of the student [23] .
Using WSI, cases with very scant material and consul-tation cases can be investigated, even when the material has to be returned to the referral center. WSI enables im-mediate accessibility to old cases, without the need for technical staff to retrieve the slides or paraffin blocks from the archive. WSI also allows using cytological im-ages while preserving the original slide [22] . Finally, WSI permits introducing FISH or immunofluorescence slides, which previously had to be shown on static images be-cause of the deterioration and loss of fluorescence.
Weaknesses of WSI
The main disadvantage of WSI is the initial economic investment for the acquisition of the scanner and the complete WSI system [1, 12] . The high-resolution WSI images are associated with files of very large size, which means significant needs in terms of disk memory for their storage. Thus, high-capacity servers to store and distrib-ute the information are needed, and regular maintenance of the servers is required. Proper functioning of the sys-tem requires high-speed Internet connection [4] .
However, the tool can be used for different subjects and, thus, be shared by different departments at the school (histology, pathology or medical specialties). A possible
solution to this problem is the use of scanners already working at the institution for pathological diagnosis and using free or low-cost software to view the virtual slides [1] . Several vendors offer renting systems that include the digitalization of a set (or a few sets) of slides, the use of teaching software and a number of terabytes in a server. This option eliminates the investment in the equipment and reduces the costs associated with servers. In any case, computer classrooms are much more versatile than labo-ratories of microscopy, and all the initial economic disad-vantages of WSI are rewarded with time, because the in-vestment in computer equipment and maintenance is much less than that of the costs associated with a labora-tory with optical microscopes [9, 15] .
Another possible disadvantage is that in the centers that only use WSI, the students do not learn how to use a conventional light microscope. Nonetheless, this is a mi-nor problem since most medical students almost never use the microscope after finishing their training. Indeed, it is more important for them to recognize histological patterns and lesions than to learn how to use the tool [8, 15, 24] .
Finally, the use of these computer-based tools may re-sult in a dramatic reduction in the personal contact be-tween teachers and students. Actions promoting face-to-face meetings should be considered to avoid the deperson-alization associated with the spread in the use of computers. On the other hand, periodic quality control of WSI is high-ly recommended. This should include a comparison be-tween virtual and glass slides, an evaluation of the use of the system outside the course schedule and the registration of what virtual images and which areas in these virtual im-ages are most often seen by the students [1] .
Fig. 3. The viewer allows displaying simul-taneously several slides on the screen, thus facilitating the analysis and interpretation of immunohistochemical techniques.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
Saco/Bombi/Garcia/Ramírez/Ordi
Pathobiology 2016;83:79–88DOI: 10.1159/000442391
84
WSI in Undergraduate Teaching
The number of centers using WSI in undergraduate teaching, either as a complement to or as a substitute of CLM, has markedly increased over the last few years. Many studies have shown excellent results with WSI, not only for medical students but also for students of biology, dentistry, parasitology and veterinary medicine [2, 7, 8, 23, 25–30] . Many of these centers have evaluated the
opinion of students after the introduction of WSI, and all have reported very positive feedback [9, 26, 30, 31] . One of the most valued advantages is the improved accessibil-ity to the slides with WSI, allowing the student to access the slides at any time and from any place [10, 14, 19] . Sev-eral studies have shown that this is one of the most ap-preciated features of WSI, and this was the most prized feature in a study conducted at our institution [10] . In-deed, data obtained from the audit of accesses to the nav-
400
350
300
250
200
150
100
50
0
Acce
sses
(n)
140
120
100
80
60
40
20
00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Acce
sses
(n)
–30 –28 –19 –15 –13 –11 –9 –7 –5 –3 –1 ExamDay
Time of day (h)
a
b
Fig. 4. a Number of accesses to the virtual slides from the opening of the website to the day of the examination. The x-axis shows the day slides were accessed in rela-tion to the examination (day 0). The y-axis shows the absolute number of accesses. Red columns identify accesses on week-ends/holidays (Saturday, Sunday or other holidays). Blue columns indicate work-days. b Time virtual slides were accessed during the day. The x-axis shows the time of the day and the y-axis shows the absolute number of accesses.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
WSI in Education Pathobiology 2016;83:79–88DOI: 10.1159/000442391
85
igator at our institution showed that over half of the ac-cesses are made on holidays and over one third after working hours ( fig. 4 ).
Students consider the use of WSI to be easier than that of CLM. WSI allows them to concentrate on the tissues and lesions and not on handling the microscope [8, 15, 32] . All studies agree that WSI improves collaboration between students and self-learning [6, 9, 16, 24, 33, 34] . Teachers also positively evaluate WSI because, although there is initially a significant increase in the time related to the preparation of the material, in the end it results in a significant reduction in the time spent in the prepara-tion of the lessons [19, 35] . WSI provides a more complete approach to the cases by adding clinical information, ra-diological imaging, macroscopic images, immunohisto-chemical stains and molecular data to the virtual histo-logical images [33, 36] . This can more accurately simulate the actual diagnostic practice and seems to be associated with better final results. Most studies agree that the image quality is better with WSI than with CLM, because the microscopes used for undergraduate teaching are, in gen-eral, of poor quality [20] . Images generated at a ×200 magnification are of sufficient quality for undergraduate teaching.
The negative evaluations are basically related to tech-nical aspects, such as the speed in loading the images or compatibility problems with some computer models. A high-speed Internet connection and a server with suffi-cient capacity are required, especially when many stu-dents view slides at the same time [12] . Some centers have solved the problem of storage space by giving the virtual images on a DVD support or scanning only the portion of tissue they want to view, thus saving space in relation with the whole slide [37] . There is still some controversy about whether WSI can totally replace CLM, because stu-dents using WSI do not learn how to handle the CLM. The students’ opinions about this issue are divided. Most studies reveal that the teachers think that it is more im-portant for the students to use their time in learning his-tology or pathology than in getting familiar with the tech-nical aspects of the management of the microscope [2] .
Finally, the evaluation process using WSI offers choos-ing between numerous slides and allows the homogeniza-tion of the test, because all the students can visualize the same slide. In addition, in contrast to CLM, only the his-topathological knowledge of the student is evaluated us-ing WSI, obviating the influence of the students’ ability to manage the microscope [32] . Students positively evaluate the use of WSI in the tests as long as the practices are car-ried out in the same way [30] . Indeed, it is strongly rec-
ommended that the same system be used to perform the practical lessons and the examinations. There is major reluctance to completely abandon CLM when lessons are taken with WSI and the tests are performed with glass slides. In contrast, if WSI is also used for the evaluation, the students do not feel the need to use CLM during the course [30] . Comparing the final results of students using WSI with those using CLM (either in the same year or in previous years), either no differences have been observed or results were better among those using WSI. Moreover, students using WSI seem to recognize histological pat-terns better [5, 31, 38, 39] .
In summary, studies assessing the use of WSI by stu-dents have shown very positive results. WSI does not seem to affect the final knowledge or may even improve it [10, 40–42] .
WSI in Postgraduate Training (Residency and
Fellowship Programs)
One of the first uses of WSI was to create series of in-teresting cases for residents or fellows in pathology and other specialties that require histological recognition of normal and pathological tissues [17] . Preparing virtual slides prevents the loss of biopsy material in making sec-tions for teaching, thus making a collection of interesting cases much easier. Many residents can view the same im-age at the same time, with the possibility of working out-side the center facilities and at any hour. These features provide an optimization of time and generate more com-fort for residents and fellows.
The presence of annotations on the virtual slides also represents a major improvement because it facilitates learning. In a study comparing results between residents of dermatology and pathology who visualized the same virtual images supplemented with or without annotations made by teachers, residents whose preparations had an-notations showed a better score, because the learning was more directed towards the most characteristic histologi-cal changes that led to the correct diagnosis [21] .
WSI has also successfully been used to measure the learning level reached by residents. Several initiatives, such as the European Association of Pathology Chairs and Residency Program Directors, aim to homogenize the knowledge of residents all over Europe evaluating them by performing a test containing virtual slides [17] . WSI is well evaluated by residents in the studies, although the final results comparing abilities reached with WSI and CLM differ from one study to another [43–45] . The
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
Saco/Bombi/Garcia/Ramírez/Ordi
Pathobiology 2016;83:79–88DOI: 10.1159/000442391
86
presence of clinical and radiological information and previous practice using WSI seem to influence the results [36] .
Some studies point out the relevance of establishing the steps involved in the diagnostic process, stressing that residents should learn not only what the correct diagnosis is, but also what the logical sequence to achieve this diag-nosis is. These studies use eye-tracking cameras that pick up eye movement during the diagnostic process. The dif-ference between senior pathologists and residents are es-tablished by examining which fields are more often viewed, the time spent in these fields and the order in which the slide is observed. Residents spend much more time evaluating the slides than experienced pathologists. However, experienced pathologists are slower in choos-ing the field in where they will zoom and they do this more frequently in areas outside central (foveal) vision than residents. Thus, there are two kinds of slide evalua-tion: one more dispersed and time-consuming and an-other more targeted and effective [17, 46] .
Although WSI is currently mainly a complement to CLM, in the future it is expected to almost completely re-place CLM as the use of WSI in routine diagnosis expands in laboratories of pathology. Studies to determine the im-pact of training pathologists without exposure to actual glass slides are warranted.
WSI in Schools of Cytopathology
The homogeneity in the quality of the images dis-played to the students is a highly appreciated feature of WSI and represents a major advantage in cytopathology, because it allows many students to use the same slide si-multaneously without the risk of breaking or losing the slide. In addition, annotated digitized slides are very use-ful for teaching cytomorphology to cytotechnologists. The use of annotations for evaluation has excellent re-sults, allowing comparison between expert cytologists and students [47] .
However, the implications of the relatively low resolu-tion of some WSI systems at low (screening) magnifica-tion still need to be solved. Moreover, viewing through a CLM may provide a different perceived field width than what is seen on a monitor. WSI has more difficulty in gen-erating appropriate images due to the difficulties in focus-ing the images at different levels [48] . This problem can be solved with the use of software packages that allow focusing the different levels of the virtual slide thereby more closely simulating the daily practice of cytology.
In conclusion, although WSI has significant educa-tional advantages, a number of technical problems should be solved before it can be confidently used to teach cytol-ogy.
Tumor Boards
Many hospitals have tumor boards where clinicians meet for multidisciplinary case presentations. Patholo-gists are often required to present the pathology findings at the board presentations. WSI is currently successfully used for this purpose at several institutional tumor boards [49, 50] . The use of WSI in tumor boards and interdisci-plinary sessions also helps to bring histology knowledge to residents of medical or surgical specialties who do not have much contact with the laboratory [22] .
Tele-Education and e-Learning
The use of e-learning and tele-education is expanding extraordinarily because it allows providing continuous medical education in a practical and convenient way, markedly reducing costs. Virtual workshops avoid the need to travel to meetings, providing significant econom-ic savings and more flexibility. WSI allows easy visualiza-tion of the cases and eliminates the requirement of send-ing the glass slides, saving time and the costs of courier service. Moreover, the system avoids the risks of loss of or damage to the glass slides [22] . Different online services permit the access to online teaching with the objective of sharing virtual slides (e.g. PathXchange, vMic Pathorama or Slide2Go) [51] .
The use of this technology is markedly spreading, and numerous institutions promote web-based learning. The United States and Canadian Academy of Pathology (USCAP), the American Society of Cytopathology and the International Academy of Cytopathology, among others, have virtual atlases that include numerous cases with educational purposes [52] . Currently, there are sev-eral collections of cases aimed at teaching residents such as the Dr. Juan Rosai collection (www.rosaicollection.org; approx. 20,000 cases from 1945 onwards), the col-lection of the Pathological Society of Great Britain (www.pathsoc.org) or other collections intended for the general public (www.virtualpathology.leeds.ac.uk) [17, 22, 53] . Virtual slides and seminars are offered by the USCAP online academy (>100 virtual slides from differ-ent organ systems). The number of journals that allow
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
WSI in Education Pathobiology 2016;83:79–88DOI: 10.1159/000442391
87
access to WSI examples to illustrate the articles, thus improving the content of the publication, is also on the rise.
Perspectives for the Future
CLM has been used in pathology departments, as well as in the laboratories of the schools of medicine, biology, dentistry and veterinary medicine, for decades. Its re-placement by new technologies represents a major chal-lenge. However, increasing needs of education and grow-ing evidence indicating the very positive effect of WSI on teaching undergraduates and pathologists are resulting in a significant expansion of this tool. Comfort in its use and
the ability to display the same slide anywhere and at any time by several people simultaneously makes this tech-nology much more convenient than CLM. However, the high maintenance costs of the laboratories of microscopy and optimization of the time of teachers and students with the use of WSI is likely to push most training centers towards the replacement of CLM by WSI. WSI is also ex-traordinarily expanding as a tool for tele-education and e-learning. Further investigation is necessary to improve the existing WSI applications specifically designed for ed-ucation and to develop ergonomic tools that improve the navigation of virtual slides. It is possible that in the near future textbooks will have the option to visualize WSI, which will be facilitated with the use of portable devices like tablets or smartphones [22] .
References
1 Paulsen FP, Eichhorn M, Brauer L: Virtual microscopy – the future of teaching histology in the medical curriculum? Ann Anat 2010; 192: 378–382.
2 Blake CA, Lavoie HA, Millette CF: Teaching medical histology at the University of South Carolina School of Medicine: transition to virtual slides and virtual microscopes. Anat Rec B New Anat 2003; 275: 196–206.
3 Pantanowitz L, Valenstein PN, Evans AJ,Kaplan KJ, Pfeifer JD, Wilbur DC, Collins LC, Colgan TJ: Review of the current state of whole slide imaging in pathology. J Pathol In-form 2011; 2: 36.
4 Al-Janabi S, Huisman A, Van Diest PJ: Digital pathology: current status and future perspec-tives. Histopathology 2012; 61: 1–9.
5 Anyanwu GE, Agu AU, Anyaehie UB: En-hancing learning objectives by use of simple virtual microscopic slides in cellular physiol-ogy and histology: impact and attitudes. Adv Physiol Educ 2012; 36: 158–163.
6 Husmann PR, O’Loughlin VD, Braun MW: Quantitative and qualitative changes in teach-ing histology by means of virtual microscopy in an introductory course in human anatomy. Anat Sci Educ 2009; 2: 218–226.
7 Farah CS, Maybury TS: The e-evolution of microscopy in dental education. J Dent Educ 2009; 73: 942–949.
8 Fonseca FP, Santos-Silva AR, Lopes MA, Al-meida OP, Vargas PA: Transition from glass to digital slide microscopy in the teaching of oral pathology in a Brazilian dental school. Med Oral Patol Oral Cir Bucal 2015; 20:e17–e22.
9 Boutonnat J, Paulin C, Faure C, Colle PE, Ronot X, Seigneurin D: A pilot study in two French medical schools for teaching histology using virtual microscopy. Morphologie 2006; 90: 21–25.
10 Ordi O, Bombi JA, Martinez A, Ramirez J, Alos L, Saco A, Ribalta T, Fernandez PL, Campo E, Ordi J: Virtual microscopy in the undergraduate teaching of pathology. J Pathol Inform 2015; 6: 1.
11 Romer DJ, Suster S: Use of virtual microscopy for didactic live-audience presentation in an-atomic pathology. Ann Diagn Pathol 2003; 7: 67–72.
12 Foster K: Medical education in the digital age: digital whole slide imaging as an e-learning tool. J Pathol Inform 2010; 1: 14.
13 Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F: Comparison of a virtual mi-croscope laboratory to a regular microscope laboratory for teaching histology. Anat Rec 2001; 265: 10–14.
14 Merk M, Knuechel R, Perez-Bouza A: Web-based virtual microscopy at the RWTH Aachen University: didactic concept, meth-ods and analysis of acceptance by the stu-dents. Ann Anat 2010; 192: 383–387.
15 Braun MW, Kearns KD: Improved learning efficiency and increased student collabora-tion through use of virtual microscopy in the teaching of human pathology. Anat Sci Educ 2008; 1: 240–246.
16 Craig FE, McGee JB, Mahoney JF, Roth CG: The Virtual Pathology Instructor: a medical student teaching tool developed using patient simulator software. Hum Pathol 2014; 45: 1985–1994.
17 Hamilton PW, Wang Y, McCullough SJ: Vir-tual microscopy and digital pathology in training and education. APMIS 2012; 120: 305–315.
18 Collier L, Dunham S, Braun MW, O’Loughlin VD: Optical versus virtual: teaching assistant perceptions of the use of virtual microscopy in an undergraduate human anatomy course. Anat Sci Educ 2012; 5: 10–19.
19 Szymas J, Lundin M: Five years of experience teaching pathology to dental students using the WebMicroscope. Diagn Pathol 2011; 6(suppl 1):S13.
20 Helin H, Lundin M, Lundin J, Martikainen P, Tammela T, Helin H, van der Kwast T, Isola J: Web-based virtual microscopy in teaching and standardizing Gleason grading. Hum Pathol 2005; 36: 381–386.
21 Marsch AF, Espiritu B, Groth J, Hutchens KA: The effectiveness of annotated (vs. non-anno-tated) digital pathology slides as a teaching tool during dermatology and pathology resi-dencies. J Cutan Pathol 2014; 41: 513–518.
22 Pantanowitz L, Szymas J, Yagi Y, Wilbur D: Whole slide imaging for educational purpos-es. J Pathol Inform 2012; 3: 46.
23 Linder E, Lundin M, Thors C, Lebbad M, Winiecka-Krusnell J, Helin H, Leiva B, Isola J, Lundin J: Web-based virtual microscopy for parasitology: a novel tool for education and quality assurance. PLoS Negl Trop Dis 2008; 2:e315.
24 Kumar RK, Velan GM, Korell SO, Kandara M, Dee FR, Wakefield D: Virtual microscopy for learning and assessment in pathology. J Pathol 2004; 204: 613–618.
25 Chen YK, Hsue SS, Lin DC, Wang WC, Chen JY, Lin CC, Lin LM: An application of virtual microscopy in the teaching of an oral and maxillofacial pathology laboratory course. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008; 105: 342–347.
26 Diaz-Perez JA, Raju S, Echeverri JH: Evalua-tion of a teaching strategy based on integra-tion of clinical subjects, virtual autopsy, pa-thology museum, and digital microscopy for medical students. J Pathol Inform 2014; 5: 25.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
Saco/Bombi/Garcia/Ramírez/Ordi
Pathobiology 2016;83:79–88DOI: 10.1159/000442391
88
27 Gatumu MK, MacMillan FM, Langton PD, Headley PM, Harris JR: Evaluation of usage of virtual microscopy for the study of histology in the medical, dental, and veterinary under-graduate programs of a UK University. Anat Sci Educ 2014; 7: 389–398.
28 Helle L, Nivala M, Kronqvist P, Gegenfurtner A, Bjork P, Saljo R: Traditional microscopy instruction versus process-oriented virtual microscopy instruction: a naturalistic experi-ment with control group. Diagn Pathol 2011; 6(suppl 1):S8.
29 McCready ZR, Jham BC: Dental students’ perceptions of the use of digital microscopy as part of an oral pathology curriculum. J Dent Educ 2013; 77: 1624–1628.
30 Weaker FJ, Herbert DC: Transition of a dental histology course from light to virtual micros-copy. J Dent Educ 2009; 73: 1213–1221.
31 Krippendorf BB, Lough J: Complete and rap-id switch from light microscopy to virtual mi-croscopy for teaching medical histology. Anat Rec B New Anat 2005; 285: 19–25.
32 Kumar RK, Freeman B, Velan GM, De Per-mentier PJ: Integrating histology and histopa-thology teaching in practical classes using vir-tual slides. Anat Rec B New Anat 2006; 289: 128–133.
33 Durosaro O, Lachman N, Pawlina W: Use of knowledge-sharing web-based portal in gross and microscopic anatomy. Ann Acad Med Singapore 2008; 37: 998–1001.
34 Goldberg HR, Dintzis R: The positive impact of team-based virtual microscopy on student learning in physiology and histology. Adv Physiol Educ 2007; 31: 261–265.
35 Triola MM, Holloway WJ: Enhanced virtual microscopy for collaborative education. BMC Med Educ 2011; 11: 4.
36 Bruch LA, De Young BR, Kreiter CD, Haugen TH, Leaven TC, Dee FR: Competency assess-ment of residents in surgical pathology using virtual microscopy. Hum Pathol 2009; 40: 1122–1128.
37 Gongora JH, Barcelo HA: Telepathology and continuous education: important tools for pa-thologists of developing countries. Diagn Pathol 2008; 3(suppl 1):S24.
38 Scoville SA, Buskirk TD: Traditional and vir-tual microscopy compared experimentally in a classroom setting. Clin Anat 2007; 20: 565–570.
39 Sivamalai S, Murthy SV, Gupta TS, Woolley T: Teaching pathology via online digital mi-croscopy: positive learning outcomes for ru-rally based medical students. Aust J Rural Health 2011; 19: 45–51.
40 Inuwa IM, Taranikanti V, Al-Rawahy M, Habbal O: Anatomy practical examinations: how does student performance on computer-ized evaluation compare with the traditional format? Anat Sci Educ 2012; 5: 27–32.
41 Mione S, Valcke M, Cornelissen M: Evalua-tion of virtual microscopy in medical histol-ogy teaching. Anat Sci Educ 2013; 6: 307–315.
42 Tian Y, Xiao W, Li C, Liu Y, Qin M, Wu Y, Xiao L, Li H: Virtual microscopy system at Chinese medical university: an assisted teaching plat-form for promoting active learning and prob-lem-solving skills. BMC Med Educ 2014; 14: 74.
43 Brick KE, Sluzevich JC, Cappel MA, DiCaudo DJ, Comfere NI, Wieland CN: Comparison of virtual microscopy and glass slide microscopy among dermatology residents during a simu-lated in-training examination. J Cutan Pathol 2013; 40: 807–811.
44 Brick KE, Comfere NI, Broeren MD, Gibson LE, Wieland CN: The application of virtual microscopy in a dermatopathology educa-
tional setting: assessment of attitudes among dermatopathologists. Int J Dermatol 2014; 53: 224–227.
45 Koch LH, Lampros JN, Delong LK, Chen SC, Woosley JT, Hood AF: Randomized compar-ison of virtual microscopy and traditional glass microscopy in diagnostic accuracy among dermatology and pathology residents. Hum Pathol 2009; 40: 662–667.
46 Krupinski EA, Tillack AA, Richter L, Hender-son JT, Bhattacharyya AK, Scott KM, Graham AR, Descour MR, Davis JR, Weinstein RS: Eye-movement study and human performance us-ing telepathology virtual slides: implications for medical education and differences with ex-perience. Hum Pathol 2006; 37: 1543–1556.
47 Stewart J III, Bevans-Wilkins K, Bhattacharya A, Ye C, Miyazaki K, Kurtycz DF: Virtual mi-croscopy: an educator’s tool for the enhance-ment of cytotechnology students’ locator skills. Diagn Cytopathol 2008; 36: 363–368.
48 Donnelly AD, Mukherjee MS, Lyden ER, Ra-dio SJ: Virtual microscopy in cytotechnology education: application of knowledge from vir-tual to glass. Cytojournal 2012; 9: 12.
49 Heffner S: Streamlining tumor board reviews. Adv Lab 2008; 17: 20.
50 Spinosa J: Scripp’s tumor board finds value in digital imaging of slides. Dark Rep 2009; 12: 10–15.
51 Conran R, Fontelo P, Liu F, Fontelo M, White E: Slide2Go: a virtual slide collection for pa-thology education. AMIA Annu Symp Proc 2007, p 918.
52 Khalbuss WE, Pantanowitz L, Parwani AV: Digital imaging in cytopathology. Patholog Res Int 2011; 2011: 264683.
53 Rosai J: Digital images of case reports and oth-er articles. Int J Surg Pathol 2007; 15: 5.
Dow
nloa
ded
by:
Uni
vers
itat d
e B
arce
lona
16
1.11
6.10
0.92
- 4
/25/
2016
11:
51:1
8 A
M
Tesis Doctoral. Adela Saco Álvarez
[95]
Estudio número 5
“Virtual Microscopy in the Undergraduate Teaching of
Pathology”
Oriol Ordi, Josep Antoni Bombí, Antonio Martínez, Josep Ramírez, Llúcia
Alòs, Adela Saco, Teresa Ribalta, Pedro L. Fernández, Elias Campo, Jaume
Ordi
Journal of Pathology Informatics 2015 Jan 29; 6:1
Factor de impacto (2016): 0
Microscopía virtual en el diagnóstico rutinario y la docencia
[96]
J Pathol Inform Editor-in-Chief: Anil V. Parwani , Liron Pantanowitz, Pittsburgh, PA, USA Pittsburgh, PA, USA
OPEN ACCESS HTML format
For entire Editorial Board visit : www.jpathinformatics.org/editorialboard.asp
Research Article
Virtual microscopy in the undergraduate teaching of pathology
Oriol Ordi1, Josep Antoni Bombí1,2, Antonio Martínez1,2, Josep Ramírez1,2, Llúcia Alòs1,2, Adela Saco1, Teresa Ribalta1,2, Pedro L. Fernández1,2, Elias Campo1,2, Jaume Ordi1,2,3
1Department of Pathology, School of Medicine, University of Barcelona, 2Department of Pathology, Hospital Clínic, 3 ISGlobal, Barcelona Center for International Health Research (CRESIB), Barcelona, Spain
E-mail: *Prof. Dr. Jaume Ordi - [email protected] *Corresponding author
Received: 28 August 2014 Accepted: 24 November 2014 Published: 29 January 15
INTRODUCTION
In the last 20 years, web‑based resources developed to supplement or replace the traditional methodologies have expanded dramatically. These resources have shown clear benefits, as classes can be delivered to many students simultaneously, and this has helped medical schools to train in a more cost‑effective way.[1]
Histology and pathology play an essential role in education in undergraduate courses in medicine. The practical knowledge of these disciplines has classically been delivered using glass slides and conventional microscopes (CM), as web‑based resources were limited to static images, which were very different from real practice. Virtual microscopy (VM), also referred to as whole slide imaging, has recently started to change the
Copyright: © 2015 Ordi O. This is an open‑access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
This article may be cited as:Ordi O, Bombí JA, Martínez A, Ramírez J, Alòs L, Saco A, et al. Virtual microscopy in the undergraduate teaching of pathology. J Pathol Inform 2015;6:1.
Available FREE in open access from: http://www.jpathinformatics.org/text.asp?2015/6/1/1/150246
Abstract
Background: Little evidence is available concerning the impact of virtual microscopy (VM) in the undergraduate teaching of pathology. We aimed: (1) to determine the impact in student scores when moving from conventional microscopy (CM) to VM; (2) to assess the students’ impressions and changes in study habits regarding the impact of this tool. Methods: We evaluated two groups taking the discipline of pathology in the same course, one using CM and the other VM. The same set of slides used in the CM classes was digitized in a VENTANA iScan HT (Roche Diagnostics, Sant Cugat, Spain) at ×20 and observed by the students using the Virtuoso viewer (Roche Diagnostics). We evaluated the skill level reached by the students with an online test. A voluntary survey was undertaken by the VM group to assess the students’ impressions regarding the resource. The day and time of any accession to the viewer were registered. Results: There were no differences between the two groups in their marks in the online test (mean marks for the CM and the VM groups: 9.87 ± 0.34 and 9.86 ± 0.53, respectively; P = 0.880). 86.6% of the students found the software friendly, easy‑to‑use and effective. 71.6% of the students considered navigation easier with VM than with CM. The most appreciated feature of VM was the possibility to access the images anywhere and at any time (93.3%). 57.5% of the accesses were made on holidays and 41.9% later than 6:00 pm. Conclusions: Virtual microscopy can effectively replace the traditional methods of learning pathology, providing mobility and convenience to medical students.Key words: Pathology, undergraduate teaching, whole slide imaging
Access this article onlineWebsite: www.jpathinformatics.org
DOI: 10.4103/2153-3539.150246
Quick Response Code:
[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]
J Pathol Inform 2015, 1:1 http://www.jpathinformatics.org/content/6/1/1
way in which these disciplines can be delivered online by providing the ability to scan entire glass slides at diagnostic resolution. A digital slide of the tissue section is created and with the use of specific software can be viewed and magnified in real‑time across the web very much like using a CM. Currently, several commercially available systems can digitize glass slides containing tissue sections and produce virtual slides of excellent quality. The rapid progress of this technology and its many potential benefits will probably result in a progressive shift from conventional to VM in routine diagnostic pathology.[2‑4]
Several studies have documented the success of VM in graduate education in medical,[5‑9] dental,[10,11] and veterinary schools.[12‑14] However, although VM has been used for many years in the US, reports on the experience with this technology in the undergraduate teaching of pathology are still limited and there is little evidence about its true impact on students’ knowledge and study habits.
In this study, we report the experience in the transition of a General Pathology course in the medical school of the University of Barcelona from practical teaching based on CM to a new format in which this tool has been totally eliminated, and all microscopy work is conducted on the computer using VM. We specifically aimed: (1) To assess the students’ impressions regarding the impact of VM on their learning and to objectively assess the success of implementing this new technology in their curriculum; (2) to determine whether moving from glass to virtual slides has an impact on student scores in practical exams of pathology; and (3) to evaluate the changes in study habits associated with the introduction of this tool in undergraduate teaching.
METHODS
The study was conducted in the Department of Pathology of the School of Medicine in the University of Barcelona, Spain. This university implemented the open‑source course management system modular object‑oriented dynamic learning environment (Moodle) in 2004, and all the supporting information and most of the activities of each discipline are available to the students in this platform.
The action was conducted in General Pathology, which is delivered in the 3rd year as a 4‑month course and has 6 European Credit Transfer System credits.[15] Two different groups take the discipline each year, the first group from September to January and the second from February to May. All the students in both groups had previously had a whole year of experience using CM in the course of histology, but none had had any previous experience with VM. During the 2013–2014 course, one group studied anatomical pathology using CM whereas the second group used VM.
Characteristics of the Conventional Microscopy CourseThis group took the discipline from September 2013 to January 2014 using CM, following the same rules established in the previous 5 years. The group was divided into six smaller groups each composed of 15 students. Two practical classes were scheduled during the course, each of which included 16 histological slides stained with hematoxylin and eosin, being representative of basic pathological lesions. Eighteen sets of slides including consecutive sections of the 16 cases were available for the use of the students in each practical class. Practical classes of 2 h in duration were delivered in the microscopy room of the medical school and were conducted by a professor of pathology. The professor briefly showed the most relevant features of each particular slide with a microscope connected to a video camera and several screens. Thereafter, each student had a set of the slides for his/her use and had 90 min to observe the slides on his/her own single‑headed microscope with the support of the faculty member, who solved all the questions and problems brought up by the students.
Characteristics of the Virtual Microscopy CourseThis group took the discipline from February to May 2014 using only VM. The practical course included the same 32 cases used in the previously described group. The group was divided into six smaller groups each composed of 15–16 students. A single practical class was scheduled at the beginning of the course (February). The class was delivered in a room equipped with a computer with internet connection and a 52‑inch screen and was conducted by a faculty member, being of 1/2 h in duration. The professor briefly showed how to accede to the website, the general characteristics of the navigator and how to retrieve the supporting information. After this initial session, all the students were allowed to access the virtual slides any day and at any time from any computer connected to Internet. The students were given the opportunity to contact their tutors for any problem or doubt encountered when observing the slides on their computers.
Virtual Slides, Navigation and Supporting FilesAll the cases were digitized in a VENTANA iScan HT (Roche‑Ventana Medical Systems, Tucson, AZ, USA) at a magnification of ×20. The system creates high‑resolution digital images of tissue sections. All files were stored on a server hosted at the Spanish Division of Roche Diagnostics. The students access the virtual slides through a hyperlink on the Moodle platform, using their own computers as virtual microscopes. The images are viewed in the Virtuoso viewer (Roche‑Ventana Medical Systems, Tucson, AZ, USA), which works as a web browser and simulates a CM [Figure 1]. Virtuoso is designed to organize the images into different cases and the cases into groups. No specific
[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]
J Pathol Inform 2015, 1:1 http://www.jpathinformatics.org/content/6/1/1
software installation is required to visualize the virtual slides.
Two supporting pdf files were posted on the main page of the discipline in the Moodle platform. One included general information to guide student accession to the viewer and the username and password necessary to accede to the website. The second pdf file included educational text discussions for each particular case.
Online Evaluation of the Skill LevelAn online quiz was performed to evaluate the skill level reached by the students in the evaluation of the microscopic lesions. A question bank containing 200 multiple choice questions was created in the Moodle platform.[16] All questions were based on static microphotographs prepared by the faculty by selection of specific areas from the same 32 glass slides included in both practical classes and had 5 possible answers. All the questions only had one correct answer qualified with a mark of +1. Each wrong answer was qualified with a mark of –0.25. For the examination, 40 questions were randomly selected from the pool. The test was available on the Moodle platform during a 24 h period. Both the order of the questions, as well as that of the answers was automatically distributed randomly for each student, thus, questions and answers were presented in a different order to every student. There was a 20 min time limit to complete the exam.
Audit of Student Access to the Virtual Slide ViewerFor the VM course, the viewer registered any access to the virtual slides by any of the students. This registration was performed anonymously, as all the students logged in using the same login and password. The day and time of any single accession, as well as the time spent by the student on each slide, were registered, and a pdf file was created with all the information.
Voluntary Student SurveyAt the conclusion of the VM course, a voluntary survey was undertaken by the students to assess the students’
impressions regarding the impact of VM on their learning and the success of implementing this new technology in their curriculum. The survey was designed using the free website https://www.surveymonkey.com/(SurveyMonkey®, Menlo Park, CA, USA) and was posted as an online hyperlink on the Moodle platform, which remained open for a whole week after the online exam was completed. Questions were related to the quality and easiness‑of‑use of the software and navigation, VM versus CM, supporting information, introduction of the activity by the professor and an online quiz [Table 1]. The student survey was designed as a Likert‑scale questionnaire with a five‑point scale with the following options: Strongly agree, agree, undecided, disagree, and strongly disagree. Ethical clearance was granted with participation in evaluations being entirely voluntary and completely anonymous.
Data AnalysisStatistical analysis was performed using the SPSS (version 18.0; SPSS, Inc., Chicago, IL). The results are presented as absolute numbers and percentages or mean and standard deviation. The analysis was mostly descriptive and included Chi‑square tests.
RESULTS
Characteristics of the GroupsThe CM course had 88 students, 67.0% (59/88) females and 33.0% (29/88) males, with a mean age of 20.6 ± 1.4. The VM course had 93 students, 68.8% (64/93) females and 31.2% (29/93) males, with a mean age of 20.8 ± 1.3. No differences were observed between the two groups.
Characteristics of the Virtual SlidesThe size of the files ranged from 149,321 to 1,851,049 Kb (mean 751,562.7 ± 413,330.2 Kb). The total weight of the 32 files was 24,050,005 Kb. The scanned images can be viewed up to a magnification of ×400 and are always in focus, with optimized contrast and adjusted illumination. At high magnifications it is easy for the student to maintain orientation with respect to the entire section, because the system indicates the position of the slide on a thumbnail showing a small representation of the section [Figure 1].
Online Evaluation of the Skill LevelThe mean mark in the online test in the CM course was 9.87 ± 0.34 (range: 8.3–10), with all the students passing the exam. Seventy‑six out of 88 (86.4%) answered all the questions correctly. The mean mark in the VM course was 9.86 ± 0.53 (range: 6.7–10) with 91/93 students (97.8%) passing the online test. Eighty‑five out of 93 (91.4%) answered all the questions correctly. No differences were observed between the two groups (P = 0.880, Student’s t‑test).
Figure 1: Screen shot of the virtual microscope display. A thumbnail showing a small representation of the whole section makes easier for the student to maintain orientation with respect to the entire section
[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]
J Pathol Inform 2015, 1:1 http://www.jpathinformatics.org/content/6/1/1
Audit of Student Accesses to the ViewerThe number of visits to the VM from the opening of the website to the day of the exam is shown in Figure 2; about 80.3% of the accesses (862/1073) were done in the week prior to the examination; 57.5% of the visits were made on holidays and 42.5% on working days. The times of access during the day are shown in Figure 3. The earliest access was at 8:33 am and the latest at 00:55 am, with 58.1% of the visits being observed between 8:00 am and 6:00 pm and 41.9% later than 6:00 pm. The mean length of the students’ accesses was 4 min 2 s. Thus, the overall time spent by the students on studying the 32 slides included in the course (including the 30‑min time of the initial presentation of the website) was 2 h 50 min.
Student SurveySixty‑one out of 93 (65.6%) students participated in the survey. Table 1 shows the questions included in the survey and the students’ answers, and Figure 4 shows the mean students’ ratings for the main items of the survey concerning the use of VM; 86.6% of the students found the software friendly, easy‑to‑use and effective for the purposes of the course. The most appreciated feature of VM was the possibility to access the images anywhere and at any time (93.3%), and 71.6% of the students thought that navigation with the virtual was easier than with glass slides. Although a significant percentage of the students were neutral to both methods, most (50.8%) preferred VM to CM.
Consequences on the Workload for the Faculty StaffThe workload in terms of time in classroom teaching for the faculty staff was reduced from 20 h to 2.5. None of the students required faculty assistance. No questions to the tutors were registered in relation to problems or
doubts encountered when observing the slides on their computers.
CONCLUSIONS
The results of this study confirm that VM can effectively replace CM to teach pathology in undergraduate courses in medical schools and show that the microscopic skills acquired with VM are comparable to those acquired with CM, the classical tool for teaching pathology. The overall feedback from the students was highly positive. Students complemented the ease of use of the software. Students felt they worked faster with VM, and over 70% thought that the navigation with the VM was easier than with the CM. The most appreciated feature of VM was the possibility to access the images anywhere and at any
Table 1: Student’s responses to the survey regarding the use of virtual and the conventional microscope
Questions included in the student’s survey Strongly agree
Agree Neutral Disagree Strongly disagree
The software is friendly, easy‑to‑use and effective for the purposes of the course 43.3 43.3 10.0 3.4 0The access to the virtual slides is quick 21.7 53.3 13.4 8.3 3.3I liked the possibility to access the images anywhere and at any time 65.0 28.3 5.0 1.7 0The quality of the image of virtual slides is adequate 26.7 38.3 21.7 8.3 5.0VM allows time saving 31.2 44.3 21.3 1.6 1.6The identification of cells and structures with VM is easy 6.7 41.7 38.3 10.0 3.3Navigation with the VM viewer is easier than with glass slides 33.3 38.3 26.7 1.7 0I had problems with the navigation 3.3 13.3 15.0 31.7 36.7The presentation of the virtual viewer and the slides by the professor is useful 6.7 26.7 33.3 16.7 16.6The presentation of the virtual viewer by the professor is unnecessary 21.7 6.7 35.0 18.3 18.3The supporting material (pdf document) is useful and adequate 73.8 21.3 4.9 0 0The test based on image captures is a good way to evaluate the knowledge acquired with VM
14.7 57.4 18.1 4.9 4.9
I prefer virtual to CM 21.3 29.5 34.4 11.5 3.3I prefer conventional to VM 1.6 13.1 42.6 23.0 19.7
The figures indicate percentages, VM: Virtual microscopy, CM: Conventional microscopy
Figure 2: Number of accesses to the virtual slides from the opening of the website to the day of the exam. The X axis shows the day of accession in relation to the exam (day 0). The Y axis shows the absolute number of accessions. Red marks identify accessions on weekends/holidays (Saturday, Sunday or other holidays). Gray marks indicate work days
[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]
J Pathol Inform 2015, 1:1 http://www.jpathinformatics.org/content/6/1/1
time. This finding was highlighted by data obtained from the audit of accesses to the navigator, showing that over half of the accesses were made on holidays, and over one‑third were made after working hours. Finally, the introduction of VM resulted in a significant reduction of the workload for the students and for the faculty staff of pathology in terms of classroom teaching, although there was a significant increase in the time related to the preparation of the material. Interestingly, the transition from conventional to VM was not a gradual process, but a sudden change, showing that positive results may be immediate.
Curriculum reform in medical schools worldwide has focused on a reduction in contact hours to decompress crowded programs, an emphasis on independent learning, and on the development of interpersonal skills and problem‑solving abilities.[11] Achieving this objective has inevitably meant that time has been reallocated from traditional areas to new educational activities deemed to be more important. In some medical schools, this has led to curricula that offer diminished opportunities for students to learn the basic medical sciences.[1,5,11,17] In this context of standardized curricula and the growing number of medical students, new strategies have been employed to improve the student experience of learning pathology. The introduction of VM, an adequate alternative to the traditional methods of teaching pathology, can help the students to achieve a satisfactory knowledge of these basic disciplines in these newly reformed curricula. Our results showed no differences in the exams between the VM and the CM groups, plus the overall favorable feelings of the students about VM are in keeping with the adequacy of this method. For the medical staff, using this tool not only results in a reduction of the work load but also allows obtaining more information about how students learn pathology, how instructive the laboratory classes are, and which slides are of real didactic value for the students.
Virtual microscopy imitates the use of a traditional microscope and glass slide. One of the main advantages of
this tool for students is that the slides are always in focus, with optimized contrast and adjusted illumination. Indeed, over 70% of the students thought that the navigation with the VM viewer was easier than with glass slides.[2‑4]
The anonymous survey showed that the students found VM useful. Our results are in keeping with previous reports showing that the students’ experience with VM is very favorable.[11] This provides clear evidence of the learning benefits derived from using this tool. VM allows students to independently explore the entire histological slide, as well as control the content and its rhythm of delivery. As observed in previous studies,[5,6,11,18,19] this interactive technology makes microscopic laboratory studies in pathology more efficient and teaching resources more portable and independent of class schedules. As shown in our study, according to the students, the most appreciated feature of VM was the possibility to accede to the images anywhere and at any time. Indeed, data obtained from the audit of accesses to the navigator showed that over half of the accesses were made on holidays and over one‑third were made after working hours. The adoption of electronic course materials, along with almost universal use of personal and laptop computers by the medical school students facilitates the introduction of VM.[11]
Although the initial equipment and software cost for creating VM is high, this new technology has the potential to revolutionize the way individuals teach and learn from microscopic images. With VM, the most representative slides with the best quality material can reassuringly be included in teaching sets. Not only can such materials be easily added to the virtual sets, but compared with glass slides these digital slides will not fade, break, or disappear. Scanned slides for dedicated teaching should be de‑identified prior to making them available for general users. One of the main advantages of VM is the portability (time and location), and ease of maintenance. Finally, this tool may allow reducing or even eliminating the expensive laboratories of microscopy.[2‑5]
Figure 3: Time of accesses to the virtual slides during the day. The X axis shows the time of the day and the Y axis shows the absolute number of accessions
Figure 4: Mean students’ ratings for the main items of the survey regarding the use of virtual microscopy
[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]
J Pathol Inform 2015, 1:1 http://www.jpathinformatics.org/content/6/1/1
The main strength of our study is that it allows adequate comparison of two very similar groups from the same course working with the same material and that it provides objective data on how students learn pathology. A possible limitation of our study is the use of an examination system restricted to standard static images. However, this allowed comparison of the results with the CM group, as all the image captures were made from the same cases. Examinations that apply virtual technology require more sophisticated management software.[4] Finally, the very good results obtained in the examinations should be considered as the consequence of the extremely high marks required in Spain to accede to the medical schools.
In conclusion, evidence showing that the microscopic skills achieved by students with VM are comparable to these acquired with CM indicates that this technology can effectively replace the traditional methods of learning pathology. One of its main advantages is that it provides mobility and convenience to medical students.
REFERENCES
1. Williams G, Lau A. Reform of undergraduate medical teaching in the United Kingdom: A triumph of evangelism over common sense. BMJ 2004;329:92‑4.
2. Al‑Janabi S, Huisman A, Van Diest PJ. Digital pathology: Current status and future perspectives. Histopathology 2012;61:1‑9.
3. Pantanowitz L, Valenstein PN, Evans AJ, Kaplan KJ, Pfeifer JD, Wilbur DC, et al. Review of the current state of whole slide imaging in pathology. J Pathol Inform 2011;2:36.
4. Pantanowitz L, Szymas J, Yagi Y, Wilbur D. Whole slide imaging for educational purposes. J Pathol Inform 2012;3:46.
5. Blake CA, Lavoie HA, Millette CF. Teaching medical histology at the University of South Carolina School of Medicine: Transition to virtual slides
and virtual microscopes. Anat Rec B New Anat 2003;275:196‑206.6. Boutonnat J, Paulin C, Faure C, Colle PE, Ronot X, Seigneurin D. A pilot
study in two French medical schools for teaching histology using virtual microscopy. Morphologie 2006;90:21‑5.
7. Foster K. Medical education in the digital age: Digital whole slide imaging as an e‑learning tool. J Pathol Inform 2010;1.
8. Goldberg HR, Dintzis R. The positive impact of team‑based virtual microscopy on student learning in physiology and histology. Adv Physiol Educ 2007;31:261‑5.
9. Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F. Comparison of a virtual microscope laboratory to a regular microscope laboratory for teaching histology. Anat Rec 2001;265:10‑4.
10. Chen YK, Hsue SS, Lin DC, Wang WC, Chen JY, Lin CC, et al. An application of virtual microscopy in the teaching of an oral and maxillofacial pathology laboratory course. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008;105:342‑7.
11. Weaker FJ, Herbert DC. Transition of a dental histology course from light to virtual microscopy. J Dent Educ 2009;73:1213‑21.
12. Dee FR, Meyerholz DK. Teaching medical pathology in the twenty‑first century: Virtual microscopy applications. J Vet Med Educ 2007;34:431‑6.
13. Mills PC, Bradley AP, Woodall PF, Wildermoth M. Teaching histology to first‑year veterinary science students using virtual microscopy and traditional microscopy: A comparison of student responses. J Vet Med Educ 2007;34:177‑82.
14. Neel JA, Grindem CB, Bristol DG. Introduction and evaluation of virtual microscopy in teaching veterinary cytopathology. J Vet Med Educ 2007;34:437‑44.
15. Doménech Martínez E, Armas Ramos H, Castro Conde JR, González Díaz JP, Méndez Pérez A, Ormazábal Ramos C, et al. Study of the introduction of the European Credit Transfer System (ECTS) in pediatrics and modification of the teaching methodology. An Pediatr (Barc) 2006;65:415‑27.
16. Inuwa IM, Taranikanti V, Al‑Rawahy M, Habbal O. Anatomy practical examinations: How does student performance on computerized evaluation compare with the traditional format? Anat Sci Educ 2012;5:27‑32.
17. Bloodgood RA, Ogilvie RW. Trends in histology laboratory teaching in United States medical schools. Anat Rec B New Anat 2006;289:169‑75.
18. Sims MH, Mendis‑Handagama C, Moore RN. Virtual microscopy in a veterinary curriculum. J Vet Med Educ 2007;34:416‑22.
19. Szymas J, Lundin M. Five years of experience teaching pathology to dental students using the WebMicroscope. Diagn Pathol 2011;6 Suppl 1:S13.
[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]
Tesis Doctoral. Adela Saco Álvarez
[103]
V. Discusión
Microscopía virtual en el diagnóstico rutinario y la docencia
[104]
Tesis Doctoral. Adela Saco Álvarez
[105]
El uso de la MV en el diagnóstico primario se está extendiendo
considerablemente en los últimos años, y cada vez más centros disponen de esta
herramienta en los Servicios de Anatomía Patológica. Esto es debido a las numerosas
ventajas que ofrece este sistema, como las herramientas informáticas, la portabilidad
o la facilidad para visualizar la misma preparación por un grupo amplio de personas. A
estas cualidades hay que añadir otras de más reciente aparición, como la
cuantificación automática de células marcadas con tinciones inmunohistoquímicas o el
reconocimiento de patrones histológicos, las cuales sirven de ayuda a la hora de
realizar el diagnóstico y disminuyen la variabilidad entre observadores. A pesar de
estas ventajas existen reticencias a su uso en el diagnóstico primario, siendo la
principal el desconocimiento de si existe suficiente evidencia científica que asegure su
no inferioridad respecto a la MC. La realización de los estudios de validación de la MV
en el diagnóstico primario se ve facilitada por la existencia de guías publicadas por la
American Telemedicine Association, el College of American Pathologists y la Canadian
Association of Pathologist, así como del Libro Blanco de la Sociedad Española de
Patología; donde se recogen una serie de recomendaciones útiles a la hora de realizar
la validación del diagnóstico primario con MV.
Nuestro estudio 1 tiene como principal objetivo evaluar si existe suficiente
bibliografía que ponga de manifiesto una buena concordancia entre los diagnósticos
con MV y MC; para ello hemos llevado a cabo una revisión de los estudios de
validación publicados hasta el momento, valorando si tanto la metodología como las
muestras evaluadas son adecuadas. Los resultados de este estudio pusieron de
manifiesto que, independientemente de la subespecialidad, todos los estudios sobre
validación tienen una muy buena correlación entre los diagnósticos alcanzados con MV
y MC; por lo cual la MV parece ser una herramienta adecuada para el diagnóstico
histológico de rutina, presentando además múltiples ventajas sobre la MC. Sin
embargo, a pesar de la buena evidencia demostrada con la MV en la práctica rutinaria
existen numerosas áreas de la Anatomía Patológica que presentan deficiencias en los
estudios de validación o una total ausencia de los mismos. Entre estas áreas se
encuentran la hematopatología, la patología hepática, ginecológica, ósea, endocrina y
de partes blandas. Algunas de estas áreas son similares entre ellas y presentan
Microscopía virtual en el diagnóstico rutinario y la docencia
[106]
características superponibles a otras que sí tienen estudios de validación, por lo cual
no resultan necesarios estudios adicionales. Sin embargo, otras como la
hematopatología o la patología hepática presentan características propias, por lo que
resultan imprescindibles nuevos estudios de validación antes del uso de la MV para
realizar el diagnóstico primario de forma generalizada en un Servicio de Anatomía
Patológica.
La citología parece ser una excepción, pues la aplicación de la MV en esta área
presenta una mayor controversia por la necesidad usar ejes adicionales a la hora de
escanear para permitir el enfoque a distintos planos, incrementando de forma muy
significativa el tamaño de las imágenes generadas.
Varios estudios ponían de manifiesto que, como ocurre con otras herramientas
nuevas, existe una curva de aprendizaje por lo que el tiempo empleado en el
diagnóstico, y en menor medida la concordancia intra e inter-observador, pueden ser
subóptimos en las fases iniciales del uso de esta tecnología [43–46,48].
Algunos estudios destacan la existencia de otras aplicaciones para la MV además
del diagnóstico primario, como por ejemplo la creación de reconstrucciones 3D de
imágenes en 2D de biopsias, lo cual puede ayudar a mejorar la comprensión de los
patrones de crecimiento y de la disposición de las células en el espacio [21,68]. Otra
área que se encuentra en reciente expansión es la referente al reconocimiento de
patrones histológicos usando análisis de imagen. Esta herramienta puede incrementar
significativamente la reproducibilidad de los diagnósticos entre patólogos en muchas
subespecialidades, consiguiendo un diagnóstico más preciso, con la repercusión que
esto tiene sobre el pronóstico y tratamiento de los pacientes [69–74].
Ante la luz de estos resultados realizamos nuevos estudios de validación del uso
de la MV en el diagnóstico de biopsias ginecológicas (estudio 2) y biopsias pequeñas
hepáticas (estudio 3), pues no existía en la bibliografía actual suficiente evidencia
sobre la no inferioridad del diagnóstico de rutina con MV comparado con MC.
Los resultados del estudio 2 muestran una elevada concordancia entre los
diagnósticos realizados con MC y MV (superior al 94%) en una serie amplia de
muestras ginecológicas de rutina. El valor de Kappa, considerado como la medida del
Tesis Doctoral. Adela Saco Álvarez
[107]
nivel de concordancia entre observadores corregido por el azar, resultó casi perfecto
(0,914). La cantidad de discrepancias observadas en nuestro estudio se encuentra
dentro del rango de la variabilidad intra-observador que se suele observar en patología
[75,76]. El diagnóstico final de consenso coincidió con el diagnóstico realizado con MV
en un 22,2% de las discrepancias mayores y en un 35,3% de las menores. Ninguna de
las discrepancias se consideró como relacionada con una pobre calidad de imagen de
MV o con la imposibilidad de conseguir una magnificación suficiente, sino que
estuvieron asociadas fundamentalmente a interpretaciones diferentes de casos
difíciles o a la presencia de lesiones de pequeño tamaño que no fueron identificadas
en la evaluación. Por tanto, nuestros resultados confirman que la MV puede ser
utilizada de forma segura en el diagnóstico histológico primario de la patología
ginecológica.
Ocho de las nueve discrepancias observadas en el estudio (88,9%) fueron en el
diagnóstico de H-SIL como L-SIL o como negativas o cambios reactivos en el cérvix
uterino (cuatro casos de cada uno). Así mismo, 13 de las 17 discrepancias menores
(76,5%) estuvieron relacionadas con el diagnóstico de L-SIL versus epitelio cervical
normal o reactivo. Como consecuencia, el valor Kappa de las biopsias o escisiones
cervicales de pacientes referidas por citología cérvico-vaginal anormal fue 0,832,
claramente inferior al valor observado en el grupo general de todas las biopsias. Esta
observación es concordante con los resultados de varios estudios que muestran la
existencia de una importante variación intra e inter-observador en la interpretación de
biopsias de cérvix uterino teñidas con H&E, con coeficientes de Kappa situados en
general entre 0,45 y 0,50; indicando un nivel de concordancia solo moderado [75,77–
82]. La estimación de la reproducibilidad en la interpretación de las muestras cérvico-
vaginales en el curso de un estudio de triaje de lesiones ASCUS/LSIL comparando los
diagnósticos del centro de origen con los resultados de la revisión de patólogos
expertos mostró una reproducibilidad moderada (Kappa<0,5) en la interpretación de
biopsias [83]. En este estudio la ausencia de reproducibilidad fue sustancialmente
mayor en biopsias por punción tipo punch que en muestras por escisión, y la
variabilidad fue mayor en la evaluación de lesiones de bajo grado; al igual que fue
observado en nuestro estudio.
Microscopía virtual en el diagnóstico rutinario y la docencia
[108]
En nuestro estudio el patólogo responsable de la evaluación con MV tenía
solamente una mínima experiencia previa en el uso de la herramienta, pero ello no
afectó de forma significativa a la exactitud del diagnóstico, ni siquiera en el periodo
inicial del estudio. No obstante, sí se detectó un incremento claro del nivel de
reproducibilidad durante el periodo de la realización del estudio, así como una
disminución en el número de casos en los que el patólogo requirió el uso de la MC para
confirmar el diagnóstico. El diagnóstico de consenso en los casos discrepantes
coincidió con el de la evaluación con MC en el 82,4% de las biopsias en el periodo
inicial, presentando una tendencia más equilibrada en el segundo periodo,
coincidiendo con el de la evaluación con MV en un 55,5% de los casos. Todas estas
evidencias indican que el incremento en la experiencia con el uso de la MV mejora los
resultados diagnósticos y que los patólogos, tras un periodo de adquisición de
experiencia, son capaces de diagnosticar prácticamente todas las muestras de forma
fiable exclusivamente con el uso de la MV.
El patólogo que trabajó con la MV no refirió dificultades para alcanzar el
diagnóstico con el aumento de 200x con el que se realizó el escaneo, indicando que no
resulta necesaria una mayor resolución en la mayor parte de las biopsias ginecológicas.
Esta estrategia de escaneo es la usada en la mayor parte de los estudios de validación
porque comporta un ahorro en el tiempo de escaneo y en los requerimientos de disco
para el almacenaje de imágenes [43,84–89]. Sin embargo, es probable que se requiera
un incremento en la magnificación del escaneo (400x) para un porcentaje pequeño de
casos para poder alcanzar un diagnóstico de forma segura.
La principal fortaleza de nuestro estudio es que el mayor análisis de validación
realizado hasta la fecha en biopsias ginecológicas, el cual incluye un número
significativo de casos que permiten un estudio estadístico robusto. Hasta el momento
solo una publicación había analizado la correlación entre MV y MC en la evaluación de
secciones congeladas de 52 lesiones ováricas, mostrando que, al igual que se observa
en nuestro estudio, la correlación entre ambos diagnósticos es muy buena [7]. La
segunda fortaleza de nuestro estudio es que los patólogos implicados en el diagnóstico
tuvieron a su disposición la información clínica relevante para el diagnóstico.
Tesis Doctoral. Adela Saco Álvarez
[109]
La principal limitación de nuestro estudio es que no se evaluó la variabilidad
intra-observador de los diagnósticos con MV y MC, lo que se considera como el mejor
diseño a la hora de evaluar la concordancia entre los dos sistemas [3]. Sin embargo, la
buena reproducibilidad inter-observador alcanzada con el estudio sugiere que los
resultados serían muy similares si las evaluaciones fuesen realizadas por el mismo
observador.
En conclusión, los diagnósticos realizados con MV son altamente concordantes
con los emitidos usando MC en las biopsias ginecológicas de rutina, por lo que la MV es
una herramienta que permite realizar diagnósticos primarios, presentando además
múltiples ventajas respecto a la MC.
La patología hepática resultó ser otra de las subespecialidades sin estudios de
validación que permitieran el uso de la MV en el diagnóstico primario. Esta área de la
patología tiene características que la diferencian de las demás subespecialidades;
entre ellas el pequeño tamaño de algunas biopsias y la presencia de muestras
derivadas del trasplante hepático, donde pequeñas variaciones en la histología pueden
cambiar de forma significativa el diagnóstico y el manejo clínico del paciente. Por este
motivo llevamos a cabo el estudio 3, donde evaluamos la concordancia tanto intra
como inter-observador en el diagnóstico de biopsias con aguja hepáticas usando MC y
MV.
Los resultados muestran una elevada concordancia intra-observador entre la
evaluación con MC y con MV de dos patólogos, siendo mayor al 90% con ambos
observadores. El valor de Kappa, considerado como la medida del nivel de
concordancia entre observadores corregido por el azar, fue casi perfecto; siendo de 0.9
para las comparaciones intra-observador de ambos observadores, y de 0.9 y 1 para las
comparaciones inter-observador con MC y MV respectivamente.
El nivel de discrepancias entre los diagnósticos con MV y MC fue menor al 10%, y
solo se identificaron discrepancias menores, pues no representaron ningún impacto en
el manejo del paciente. Además, cabe destacar que ninguna de ellas fue relativa a una
deficiente calidad de imagen o a un aumento insuficiente. Todas las discrepancias
Microscopía virtual en el diagnóstico rutinario y la docencia
[110]
encontradas fueron secundarias al pequeño tamaño del material o a la dificultad
intrínseca del caso.
Existen numerosos estudios que muestran variaciones sustanciales inter e intra-
observador en el diagnóstico de biopsias hepáticas, aunque se use MC en ambas
visualizaciones. Estos estudios se encuentran limitados a patologías concretas como la
esteatohepatitis no alcohólica o la hepatitis crónica de origen viral [90–94]. En uno de
los estudios donde se evaluaba la concordancia intra-observador de 50 biopsias
hepáticas orientadas como esteatohepatitis no alcohólica, Kleiner et al reportaron un
valor de Kappa de 0.61 [91]. Nuestro estudio mostró una tasa de concordancia mayor,
con un valor de Kappa que variaba entre 0.7 y 0.9 en las diferentes comparaciones, a
pesar de que se evaluó un numero mayor de casos y se incluyó esteatohepatis
alcoholica y no alcoholica. Tres de los estudios evaluaban la concordancia intra-
observador en el diagnóstico de la hepatitis crónica de origen viral. La evaluación de
los grados de fibrosis y el estadio en esos estudios mostraron valores de Kappa que
oscilaban entre 0.72 y 1 [92–94], los cuales son comparables con los hallados en
nuestro estudio (entre 0.7 y 0.9). Estas discrepancias pueden ser atribuidas a la
inherente variabilidad intra-observador en el diagnóstico de las biopsias con aguja
hepáticas. Algunos de estos estudios evaluaron distintas características histológicas
por separado, mostrando altas tasas de concordancia en la evaluación de la esteatosis
(K=0,79), necrosis periportal (K=0,74) y fibrosis (K=0,86), y tasas más bajas en la
necrosis lobulillar (K=0,42) [90,91]. En nuestro estudio los niveles de concordancia
entre la evaluación con MV y MC de todos esos hallazgos histológicos fueron aún
mejores. Estos resultados confirmaron que la MV puede ser usada para el diagnóstico
primario de rutina de las biopsias hepáticas.
Una posible limitación de nuestro estudio es el bajo número de casos de cada
una de las patologías hepáticas, comparado con el de los estudios previos [90–94]. Sin
embargo, nuestro estudio está diseñado para evaluar el diagnóstico en la práctica
rutinaria, donde se recibe una gran variedad de patologías distintas, y no para evaluar
una patología concreta.
Tesis Doctoral. Adela Saco Álvarez
[111]
El análisis por separado de la patología de trasplante (n=64) mostró una alta
concordancia intra-observador entre diagnósticos con MV y MC, siendo casi perfecta
(93.7%; κ= 0.9 para un observador y 87.5%; κ= 0.8 para el otro). No hubo diferencias en
el diagnóstico de rechazo.
Cabe destacar, que los patólogos no refirieron ninguna dificultad para alcanzar el
diagnóstico con la magnitud de 400x usada en las preparaciones. Al contrario de lo que
sucede en otras áreas de la patología, donde objetivos menores son suficientes, en la
patología hepática 400x parece la magnitud más aducada debido a las características
propias de esta subespecialidad; como son el pequeño tamaño de las biopsias con
aguja o la necesidad de evaluar cambios histológicos muy sutiles [43,84,85,87,88].
En conclusión, los diagnósticos de biopsias pequeñas hepáticas realizados con
MV tienen una gran concordancia intra e inter-observador respecto a los llevados a
cabo con MC. Nuestros resultados confirman que la MV puede ser usada de forma
segura para el diagnóstico histológico primario de biopsias hepáticas, tanto de hígados
de trasplante como de nativos.
Cabe destacar que las altas tasas de concordancia halladas tanto en patología
hepática (estudio 3) como ginecológica (estudio 2) son similares a las publicadas en
otros estudios de validación del uso de la MV en biopsias de piel [43,95], mama
[45,96], próstata [76,85], vejiga urinaria [97], patología gastrointestinal [86,98] y
pediátrica [89], recogidos en la revisión realizada en el estudio 1. Los patólogos
encargados del diagnóstico con MV también destacaron algunas de sus ventajas, como
la posibilidad de usar herramientas informáticas y de análisis de imagen. Éstas
permiten realizar de forma más objetiva la medición de lesiones y de la profundidad de
infiltración, lo cual es una información muy relevante en algunas áreas anatómicas
como la vulva, cérvix uterino o endometrio, entre otras. La posibilidad de visualización
de varias preparaciones a la vez en una misma pantalla también resulta muy util,
especialmente en las biopsias hepáticas donde se realizan múltiples tinciones
frecuentemente. Además, la MV facilita la realización de consultas diagnósticas, así
como los comités multidiscipliarios.
Microscopía virtual en el diagnóstico rutinario y la docencia
[112]
A pesar de los buenos resultados de la MV en el diangóstico rutinario, hay varias
consideraciones a tener en cuenta que dificultan el uso de este sistema, siendo las
principales el rechazo de los patólogos a abandonar la MC y el elevado coste
económico de la adquisición de equipamiento y del archivo de las imágenes generadas.
También cabe destacar los aspectos legales que envuelven al diagnóstico primario con
MV referentes a la confidencialidad de datos, la calidad de imagen, la calidad de los
monitores, el espacio de almacenamiento y la confianza a la hora de realizar los
diagnósticos [17–22]. Estas cuestiones comienzan a ser reguladas en otros países como
EEUU, donde la Food and Drug Administration (FDA) empieza a establecer la normativa
a seguir a la hora de diagnosticar usando MV. En el ámbito nacional, aún no existe
legislación al respecto, pero cabe esperar que en un futuro cercano ésta exista.
La docencia es otro de los ámbitos donde la MV puede tener un gran impacto
positivo debido a sus numerosas ventajas, destacando la portabilidad o la posibilidad
de visualizar una misma preparación por un gran número de personas. Al igual que
ocurre con el diagnóstico primario, resulta fundamental asegurarse de que la MV es
capaz de sustituir a la MC sin que el aprendizaje de los alumnos de pre y post-grado se
vea afectado de forma desfavorable. Con este objetivo realizamos el estudio 4, el cual
mostro la existencia de numerosas publicaciones científicas con unos buenos
resultados en el uso de la MV aplicada a la docencia de pre y post-grado, evidenciando
la no inferioridad de esta nueva tecnología respecto a la MC. En el caso de la docencia
de pre-grado tanto en facultades de Medicina como de Odontología, Veterinaria o
Parasitología, los resultados muestran que la MV puede sustituir a la MC sin que los
conocimientos de los estudiantes se vean afectados, presentando además numerosas
ventajas sobre esta última [32,49,57,59,99–104]. La opinión de los estudiantes sobre el
uso de la MV es muy positiva; siendo la fácil accesibilidad a las preparaciones la
característica mejor valorada [35,37]. Otras ventajas con buena aceptación por parte
del alumnado fueron la facilidad de uso, el fomento de la cooperación entre
estudiantes y el autoaprendizaje. La realización de los exámenes mediante MV
también destaca por sus ventajas, pues se valora realmente el conocimiento de los
alumnos, el material se encuentra homogeneizado y con una buena calidad de imagen
en todo momento.
Tesis Doctoral. Adela Saco Álvarez
[113]
El resultado final de las evaluaciones en los diferentes estudios o bien no mostró
diferencias entre los estudiantes de pre-grado que se prepararon usando MV y MC, o
bien resultó favorable al uso de MV, pues estos estudiantes parecían reconocer
patrones histológicos con mayor facilidad debido a que no tenían que centrarse en el
aprendizaje del uso del microscopio convencional [105–108]
En el caso de la docencia de post-grado, los estudios también ponen de
manifiesto unos resultados favorables al uso de la MV. La MV permite crear series de
casos interesantes que ayudan a unificar y homogenizar el aprendizaje,
independientemente del centro donde se encuentre el patólogo en formación,
preservando a su vez el material histológico.
Algunos estudios indican que el uso de preparaciones digitales con anotaciones
previamente creadas por los docentes hace que el aprendizaje sea más dirigido a las
características histológicas que llevan al diagnóstico, haciendo el proceso de
aprendizaje más eficiente y mejorando las puntuaciones finales de los exámenes [109].
La MV aplicada al aprendizaje de citopatología también presenta muy buenos
resultados en distintos estudios, solucionando los problemas técnicos que impedían la
homogeneización de las preparaciones. En este ámbito, el uso de anotaciones parece
mejorar aún más los resultados finales de los examenes [110].
Existen otros usos de la MV aplicados a la docencia que incluyen “tumor boards”
y tele-aprendizaje; este último permite tanto a patologos en formación como a
especialistas disponer de series de casos interesantes destinados a la formación
continuada [111,112].
En resumen, el estudio 4 ha puesto de manifiesto que a pesar de que en el
momento actual el uso de la MV en la docencia suele ser complementario a la MC,
especialmente en el caso de la educación de post-grado, existen sufientes evidencias
de que puede sustituir totalmente a la MC sin que el proceso de aprendizaje se vea
afectado negativamente. Además, la MV presenta numerosas ventajas y mejora la
homogeneización del material destinado al estudio.
Microscopía virtual en el diagnóstico rutinario y la docencia
[114]
Los buenos resultados puestos de manifiesto en este estudio y las ventajas de la
MV propiciaron la implementación de esta tecnología en la Universidad de Barcelona,
con el objetivo de realizar las clases de Anatomía Patológica a alumnos de la Facultad
de Medicina. Durante este proceso realizamos el estudio 5 con el fin de documentar la
transición de MC a MV y para validar la aplicación de esta última la docencia de pre-
grado, así como recoger su valoración por parte del alumnado.
Los resultados confirman que la MV puede reemplazar de forma efectiva a la MC
en la docencia de estudiantes de pre-grado, y muestran que las habilidades alcanzadas
con MV son comparables a aquellas alcanzadas con MC. La aceptación de esta nueva
tecnología por parte de los estudiantes fue muy positiva, resaltando el fácil uso del
programa informático. La característica mejor valorada de la MV fue la posibilidad de
acceder a todas las imágenes en cualquier momento y en cualquier ubicación. Esta
ventaja fue remarcada por los datos derivados de la auditoría del acceso al navegador,
mostrando que sobre la mitad de los accesos se llevaron a cabo en vacaciones y un
tercio fuera de horas lectivas.
La introducción de la MV se tradujo en una disminución del trabajo realizado por
los estudiantes y docentes en términos de tiempo empleado en las clases, aunque el
tiempo de preparación del material se vio incrementado de forma significativa.
En nuestro estudio la transición entre MC y MV no se realizó progresivamente,
sino que se llevó a cabo de forma abrupta, demostrando que los resultados positivos
se pueden ver de forma inmediata.
Las recientes reformas en los programas docentes se centran en la reducción de
horas de clase y prácticas presenciales, para dar mayor énfasis al autoaprendizaje, el
desarrollo de habilidades personales y la capacidad para la resolución de problemas
[104]. Con el fin de alcanzar estos objetivos el tiempo empleado en actividades más
tradicionales ha sido reubicado en nuevas actividades más importantes; lo que ha
dado paso a que en algunas facultades de medicina hayan disminuido
considerablemente las oportunidades de los estudiantes de aprender ciencias médicas
básicas [57,104,113,114]. Por esta razón resulta necesario establecer nuevas
estrategias, con el objetivo de mejorar el aprendizaje de Anatomía Patológica. La
Tesis Doctoral. Adela Saco Álvarez
[115]
introducción de la MV es una alternativa adecuada a los sistemas de aprendizaje
tradicionales, pudiendo ayudar a los estudiantes a alcanzar un conocimiento
satisfactorio en estas disciplinas básicas.
Nuestro estudio no mostró diferencias significativas en los resultados de los
exámenes entre el grupo que realizó las practicas con MC y el que lo hizo con MV, con
la ventaja adicional de que los estudiantes se sintieron más cómodos con el uso de la
MV para el aprendizaje. Para los docentes, la MV no solo disminuyó el tiempo
empleado en las clases, sino que también ayudó a conocer cómo es el proceso de
aprendizaje de los alumnos de Anatomía Patológica, cómo de instructivas son las
clases con MC y qué preparaciones resultan realmente didácticas.
La VM imita el funcionamiento de la MC, con la ventaja de que las preparaciones
se encuentran siempre enfocadas y con un óptimo contraste e iluminación. Además,
hasta un 70% de los estudiantes encontraron más sencilla la navegación de las
preparaciones digitales que las convencionales con microscopio óptico.
Las encuestas anónimas pusieron de manifiesto que la mayoría de estudiantes
encontraba útil la MV; este resultado concuerda con el de otros estudios donde la MV
también fue valorada de forma positiva por parte del alumnado [104]. Esto pone de
manifiesto los beneficios de la aplicación de la MV a la docencia, pues permite que los
estudiantes exploren las preparaciones de forma independiente, controlando el
contenido y el ritmo de aprendizaje. Además, el uso generalizado de ordenadores
personales y la experiencia previa de los estudiantes con dispositivos informáticos,
hacen que la adaptación y el manejo de la MV resulte muy sencilla [104].
Los resultados son concordantes con los de otros estudios, recogidos en nuestro
estudio 4, los cuales ponen de manifiesto que esta tecnología hace que las prácticas de
laboratorio de microscopía sean más eficientes y hace posible la portabilidad de los
recursos docentes sin depender de calendarios académicos [30,37,57,104]. Además,
en nuestro estudio la característica mejor valorada de la MV fue la posibilidad de
acceder al sistema desde cualquier lugar y en cualquier momento, lo que quedó
confirmado con los datos de la auditoría del acceso al navegador.
Microscopía virtual en el diagnóstico rutinario y la docencia
[116]
A pesar de que el coste de la implementación de la MV es alto, este nuevo
sistema puede representar un gran avance en la forma de enseñar y aprender
Anatomía Patológica. La MV posibilita que el material más representativo y de mejor
calidad de cada caso pueda ser incluido, eliminando la posibilidad de deterioro o
pérdida de las preparaciones; aunque es necesario que las muestras pasen por un
proceso de anonimato previo, para respetar la confidencialidad de datos de los
pacientes. Otra de las principales ventajas de la MV es la facilidad de mantenimiento,
pues esta herramienta puede reducir e incluso eliminar los gastos destinados al
mantenimiento de los laboratorios de microscopía [2,38,42,57].
La principal fortaleza de nuestro estudio es la comparación de dos grupos muy
similares del mismo curso, que trabajaban con el mismo material; lo cual provee datos
objetivos sobre el efecto de la MV en el proceso de aprendizaje. Una posible limitación
de nuestro estudio es que los exámenes se realizaron usando imágenes estáticas en
vez de preparaciones virtuales que permitan la navegación, pues se requieren
programas informáticos más sofisticados para poder realizarlos de esta manera.
Aunque hay que tener en consideración que en ambos grupos los exámenes se
realizaron con imágenes provenientes de los mismos casos. Por último, cabe destacar
que las altas calificaciones alcanzadas por los estudiantes son consecuencia de las altas
notas necesarias para ingresar en la Facultad de Medicina.
En resumen, la evidencia muestra que las habilidades en microscopía alcanzadas
con MV son comparables con aquellas alcanzadas con MC, lo que indica que esta
tecnología puede reemplazar de forma satisfactoria los métodos de aprendizaje
tradicionales, aportando a su vez mayor portabilidad entre otras ventajas.
La conclusión final de nuestros estudios es que la MV es una herramienta capaz
de sustituir a la MC tanto en la docencia de pre y post-grado, como en el diagnóstico
rutinario de biopsias. Existen suficientes estudios de validación que aseguran la no
inferioridad de la MV respecto a la MC en docencia y en el diagnóstico primario de
gran parte de las áreas de la Anatomía Patológica; siendo necesaria la realización de
nuevos estudios de validación que incluyan algunas subespecialidades que no
disponen de los mismos. La expansión de esta tecnología por los Servicios de Anatomía
Tesis Doctoral. Adela Saco Álvarez
[117]
Patológica y las Facultades hará que cada vez más centros puedan disponer de sus
múltiples ventajas.
Microscopía virtual en el diagnóstico rutinario y la docencia
[118]
Tesis Doctoral. Adela Saco Álvarez
[119]
VI. Conclusiones
Microscopía virtual en el diagnóstico rutinario y la docencia
[120]
Tesis Doctoral. Adela Saco Álvarez
[121]
1. Existe una muy buena correlación entre los diagnósticos realizados con MC y
MV en gran parte de las subespecialidades de la Anatomía Patológica, por lo
que se puede considerar que la implementación de esta tecnología para el
diagnóstico primario es segura (estudio 1)
2. Existe una ausencia total de estudios de validación en algunas áreas de la
Anatomía Patológica como en la patología ginecológica, la patología hepática,
la hematopatología, la patología ósea, endocrina y de partes blandas, razón por
la cual es necesario realizar estudios de validación que incluyan muestras de
estas subespecialidades (estudio 1)
3. La MV puede ser utilizada de forma segura en el diagnóstico histológico
primario de la patología ginecológica, puesto que la concordancia inter-
observador entre los diagnósticos obtenidos con la MV y la MC es muy alta
(estudio 2)
4. Existe una curva de aprendizaje en el uso de la MV, pues la experiencia se
traduce en una mejoría de los resultados y permite a los patólogos diagnosticar
con seguridad prácticamente todas las muestras usando exclusivamente la MV
(estudio 2)
5. El escaneo a 200x aumentos es adecuado para realizar un diagnóstico de la
mayoría de las biopsias ginecológicas de forma fiable (estudio 2)
6. La MV puede ser usada de forma segura para el diagnóstico histológico
primario de biopsias hepáticas con aguja, tanto en hígado nativo como en
hígado de trasplante, mostrando unos altos índices de concordancia tanto intra
como inter-observador (estudio 3)
7. El escaneo a 400x aumentos ha demostrado ser adecuado para realizar el
diagnóstico de las biopsias hepáticas con aguja de forma satisfactoria (estudio
3)
Microscopía virtual en el diagnóstico rutinario y la docencia
[122]
8. Existe suficiente evidencia de que la MV es una herramienta adecuada para la
docencia de pre y post-grado tanto en los estudios de Medicina, como de
Veterinaria u Odontología (estudio 4)
9. Las habilidades en microscopía alcanzadas por los estudiantes de pregrado de
Medicina usando la MV son comparables con las alcanzadas con la MC.
Además, este cambio puede realizarse de forma rápida (estudio 5)
10. Los alumnos valoran de forma positiva la implementación de la MV en la
docencia. Las principales ventajas reconocidas por ellos son el fácil uso del
programa informático y la posibilidad de realizar las prácticas a cualquier hora y
desde cualquier lugar (estudio 5)
Tesis Doctoral. Adela Saco Álvarez
[123]
VII. Bibliografía
Microscopía virtual en el diagnóstico rutinario y la docencia
[124]
Tesis Doctoral. Adela Saco Álvarez
[125]
[1] Weinstein RS. Prospects for telepathology. Hum Pathol 1986;17:433–4.
[2] Al-Janabi S, Huisman AAA, Van Diest PJ. Digital pathology: Current status and
future perspectives. Histopathology 2012;61:1–9. doi:10.1111/j.1365-
2559.2011.03814.x.
[3] Pantanowitz L, Sinard JH, Henricks WH, Fatheree LA, Carter AB, Contis L, et al.
Validating whole slide imaging for diagnostic purposes in pathology: guideline
from the College of American Pathologists Pathology and Laboratory Quality
Center. Arch Pathol Lab Med 2013;137:1710–22. doi:10.5858/arpa.2013-0093-
CP.
[4] Cross SS, Dennis T, Start RD. Telepathology: current status and future prospects
in diagnostic histopathology. Histopathology 2002;41:91–109.
[5] Evans AJ, Kiehl T-R, Croul S. Frequently asked questions concerning the use of
whole-slide imaging telepathology for neuropathology frozen sections. Semin
Diagn Pathol 2010;27:160–6.
[6] Evans AJ, Chetty R, Clarke BA, Croul S, Ghazarian DM, Kiehl T-R, et al. Primary
frozen section diagnosis by robotic microscopy and virtual slide telepathology:
the University Health Network experience. Semin Diagn Pathol 2009;26:165–76.
[7] Fallon MA, Wilbur DC, Prasad M. Ovarian frozen section diagnosis: use of whole-
slide imaging shows excellent correlation between virtual slide and original
interpretations in a large series of cases. Arch Pathol Lab Med 2010;134:1020–3.
doi:10.1043/2009-0320-OA.1.
[8] Gould P V, Saikali S. A comparison of digitized frozen section and smear
preparations for intraoperative neurotelepathology. Anal Cell Pathol (Amst)
Microscopía virtual en el diagnóstico rutinario y la docencia
[126]
2012;35:85–91. doi:10.3233/ACP-2011-0026.
[9] Kaplan KJ, Burgess JR, Sandberg GD, Myers CP, Bigott TR, Greenspan RB. Use of
robotic telepathology for frozen-section diagnosis: a retrospective trial of a
telepathology system for intraoperative consultation. Mod Pathol an Off J
United States Can Acad Pathol Inc 2002;15:1197–204.
doi:10.1097/01.MP.0000033928.11585.42.
[10] Pantanowitz L, Dickinson K, Evans AJ, Hassell LA, Henricks WH, Lennerz JK, et al.
American Telemedicine Association clinical guidelines for telepathology. J Pathol
Inform 2014;5:39. doi:10.4103/2153-3539.143329.
[11] Slodkowska J, Pankowski J, Siemiatkowska K, Chyczewski L. Use of the virtual
slide and the dynamic real-time telepathology systems for a consultation and
the frozen section intra-operative diagnosis in thoracic/pulmonary pathology.
Folia Histochem Cytobiol 2009;47:679–84. doi:10.2478/v10042-010-0009-z.
[12] Wilbur DC. Digital pathology: get on board-the train is leaving the station.
Cancer Cytopathol 2014;122:791–5. doi:10.1002/cncy.21479.
[13] Ayad E. Virtual telepathology in Egypt, applications of WSI in Cairo University.
Diagn Pathol 2011;6 Suppl 1:S1. doi:10.1186/1746-1596-6-S1-S1.
[14] Romero Lauro G, Cable W, Lesniak A, Tseytlin E, McHugh J, Parwani A, et al.
Digital pathology consultations-a new era in digital imaging, challenges and
practical applications. J Digit Imaging 2013;26:668–77. doi:10.1007/s10278-013-
9572-0.
[15] Wienert S, Beil M, Saeger K, Hufnagl P, Schrader T. Integration and acceleration
of virtual microscopy as the key to successful implementation into the routine
Tesis Doctoral. Adela Saco Álvarez
[127]
diagnostic process. Diagn Pathol 2009;4:3. doi:10.1186/1746-1596-4-3.
[16] Wilbur DC, Madi K, Colvin RB, Duncan LM, Faquin WC, Ferry JA, et al. Whole-
slide imaging digital pathology as a platform for teleconsultation: a pilot study
using paired subspecialist correlations. Arch Pathol Lab Med 2009;133:1949–53.
doi:10.1043/1543-2165-133.12.1949.
[17] Bernard C, Chandrakanth SA, Cornell IS, Dalton J, Evans A, Garcia BM, et al.
Guidelines from the Canadian Association of Pathologists for establishing a
telepathology service for anatomic pathology using whole-slide imaging. J Pathol
Inform 2014;5:15. doi:10.4103/2153-3539.129455.
[18] Hedvat C V. Digital microscopy: past, present, and future. Arch Pathol Lab Med
2010;134:1666–70. doi:10.1043/2009-0579-RAR1.1.
[19] Ho J, Ahlers SM, Stratman C, Aridor O, Pantanowitz L, Fine JL, et al. Can digital
pathology result in cost savings? A financial projection for digital pathology
implementation at a large integrated health care organization. J Pathol Inform
2014;5:33. doi:10.4103/2153-3539.139714.
[20] Isaacs M, Lennerz JK, Yates S, Clermont W, Rossi J, Pfeifer JD. Implementation of
whole slide imaging in surgical pathology: A value added approach. J Pathol
Inform 2011;2:39. doi:10.4103/2153-3539.84232.
[21] Pantanowitz L. Digital images and the future of digital pathology. J Pathol Inform
2010;1. doi:10.4103/2153-3539.68332.
[22] Thorstenson S, Molin J, Lundstrom C. Implementation of large-scale routine
diagnostics using whole slide imaging in Sweden: Digital pathology experiences
2006-2013. J Pathol Inform 2014;5:14. doi:10.4103/2153-3539.129452.
Microscopía virtual en el diagnóstico rutinario y la docencia
[128]
[23] Hartman DJ, Parwani A V, Cable B, Cucoranu IC, McHugh JS, Kolowitz BJ, et al.
Pocket pathologist: A mobile application for rapid diagnostic surgical pathology
consultation. J Pathol Inform 2014;5:10. doi:10.4103/2153-3539.129443.
[24] Roy S, Pantanowitz L, Amin M, Seethala RR, Ishtiaque A, Yousem SA, et al.
Smartphone adapters for digital photomicrography. J Pathol Inform 2014;5:24.
doi:10.4103/2153-3539.137728.
[25] Speiser JJ, Hughes I, Mehta V, Wojcik EM, Hutchens KA. Mobile
teledermatopathology: using a tablet PC as a novel and cost-efficient method to
remotely diagnose dermatopathology cases. Am J Dermatopathol 2014;36:54–7.
doi:10.1097/DAD.0b013e3182863186.
[26] Gavrielides MA, Conway C, O’Flaherty N, Gallas BD, Hewitt SM. Observer
performance in the use of digital and optical microscopy for the interpretation
of tissue-based biomarkers. Anal Cell Pathol (Amst) 2014;2014:157308.
doi:10.1155/2014/157308.
[27] Nassar A, Cohen C, Agersborg SS, Zhou W, Lynch KA, Barker EA, et al. A multisite
performance study comparing the reading of immunohistochemical slides on a
computer monitor with conventional manual microscopy for estrogen and
progesterone receptor analysis. Am J Clin Pathol 2011;135:461–7.
doi:10.1309/AJCP4VFKA5FCMZNA.
[28] Micsik T, Kiszler G, Szabo D, Krecsak L, Hegedus C, Tibor K, et al. Computer Aided
Semi-Automated Evaluation of HER2 Immunodetection--A Robust Solution for
Supporting the Accuracy of Anti HER2 Therapy. Pathol Oncol Res 2015;21:1005–
11. doi:10.1007/s12253-015-9927-6.
Tesis Doctoral. Adela Saco Álvarez
[129]
[29] Krenacs T, Zsakovics I, Diczhazi C, Ficsor L, Varga VS, Molnar B. The potential of
digital microscopy in breast pathology. Pathol Oncol Res 2009;15:55–8.
doi:10.1007/s12253-008-9087-z.
[30] Boutonnat J, Paulin C, Faure C, Colle PE, Ronot X, Seigneurin D. A pilot study in
two French medical schools for teaching histology using virtual microscopy.
Morphologie 2006;90:21–5.
[31] Braun MW, Kearns KD. Improved learning efficiency and increased student
collaboration through use of virtual microscopy in the teaching of human
pathology. Anat Sci Educ 2008;1:240–6. doi:10.1002/ase.53.
[32] Linder E, Lundin M, Thors C, Lebbad M, Winiecka-Krusnell J, Helin H, et al. Web-
based virtual microscopy for parasitology: A novel tool for education and quality
assurance. PLoS Negl Trop Dis 2008;2:e315. doi:10.1371/journal.pntd.0000315.
[33] Paulsen FP, Eichhorn M, Brauer L, Bräuer L. Virtual microscopy-The future of
teaching histology in the medical curriculum? Ann Anat 2010;192:378–82.
doi:10.1016/j.aanat.2010.09.008.
[34] Foster K. Medical education in the digital age: Digital whole slide imaging as an
e-learning tool. J Pathol Inform 2010;1:38–40. doi:10.4103/2153-3539.68331.
[35] Merk M, Knuechel R, Perez-Bouza A. Web-based virtual microscopy at the RWTH
Aachen University: Didactic concept, methods and analysis of acceptance by the
students. Ann Anat 2010;192:383–7. doi:10.1016/j.aanat.2010.01.008.
[36] Collier L, Dunham S, Braun MW, O’Loughlin VD. Optical versus virtual: Teaching
assistant perceptions of the use of virtual microscopy in an undergraduate
human anatomy course. Anat Sci Educ 2012;5:10–9. doi:10.1002/ase.262.
Microscopía virtual en el diagnóstico rutinario y la docencia
[130]
[37] Szymas J, Lundin M. Five years of experience teaching pathology to dental
students using the WebMicroscope. Diagn Pathol 2011;6:S13.
doi:10.1186/1746-1596-6-S1-S13.
[38] Pantanowitz L, Szymas J, Yagi Y, Wilbur D. Whole slide imaging for educational
purposes. J Pathol Inform 2012;3:46. doi:10.4103/2153-3539.104908.
[39] Husmann PR, O’Loughlin VD, Braun MW. Quantitative and qualitative changes in
teaching histology by means of virtual microscopy in an introductory course in
human anatomy. Anat Sci Educ 2009;2:218–26. doi:10.1002/ase.105.
[40] Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F. Comparison of a virtual
microscope laboratory to a regular microscope laboratory for teaching
histology. Anat Rec 2001;265:10–4.
[41] Cornish TC, Swapp RE, Kaplan KJ. Whole-slide imaging: routine pathologic
diagnosis. Adv Anat Pathol 2012;19:152–9.
doi:10.1097/PAP.0b013e318253459e.
[42] Pantanowitz L, Valenstein PN, Evans AJ, Kaplan KJ, Pfeifer JD, Wilbur DC, et al.
Review of the current state of whole slide imaging in pathology. J Pathol Inform
2011;2:36. doi:10.4103/2153-3539.83746.
[43] Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJA, Ten Kate FJW, et al.
Whole slide images for primary diagnostics in dermatopathology: a feasibility
study. J Clin Pathol 2012;65:152–8. doi:10.1136/jclinpath-2011-200277.
[44] Randell R, Ruddle RA, Mello-Thoms C, Thomas RG, Quirke P, Treanor D. Virtual
reality microscope versus conventional microscope regarding time to diagnosis:
an experimental study. Histopathology 2013;62:351–8. doi:10.1111/j.1365-
Tesis Doctoral. Adela Saco Álvarez
[131]
2559.2012.04323.x.
[45] Krishnamurthy S, Mathews K, McClure S, Murray M, Gilcrease M, Albarracin C,
et al. Multi-institutional comparison of whole slide digital imaging and optical
microscopy for interpretation of hematoxylin-eosin-stained breast tissue
sections. Arch Pathol Lab Med 2013;137:1733–9. doi:10.5858/arpa.2012-0437-
OA.
[46] Houghton JP, Ervine AJ, Kenny SL, Kelly PJ, Napier SS, McCluggage WG, et al.
Concordance between digital pathology and light microscopy in general surgical
pathology: a pilot study of 100 cases. J Clin Pathol 2014;67:1052–5.
doi:10.1136/jclinpath-2014-202491.
[47] Sanders DSA, Grabsch H, Harrison R, Bateman A, Going J, Goldin R, et al.
Comparing virtual with conventional microscopy for the consensus diagnosis of
Barrett’s neoplasia in the AspECT Barrett’s chemoprevention trial pathology
audit. Histopathology 2012;61:795–800. doi:10.1111/j.1365-2559.2012.04288.x.
[48] Randell R, Ruddle RA, Thomas RG, Mello-Thoms C, Treanor D. Diagnosis of major
cancer resection specimens with virtual slides: impact of a novel digital
pathology workstation. Hum Pathol 2014;45:2101–6.
doi:10.1016/j.humpath.2014.06.017.
[49] Fonseca F-P, Santos-Silva A-R, Lopes M-A, Almeida O-P de, Vargas P-A.
Transition from glass to digital slide microscopy in the teaching of oral pathology
in a Brazilian dental school. Med Oral Patol Oral Cir Bucal 2015;20:e17-22.
[50] Kumar RK, Velan GM, Korell SO, Kandara M, Dee FR, Wakefield D. Virtual
microscopy for learning and assessment in pathology. J Pathol 2004;204:613–8.
Microscopía virtual en el diagnóstico rutinario y la docencia
[132]
doi:10.1002/path.1658.
[51] Ho J, Parwani A V, Jukic DM, Yagi Y, Anthony L, Gilbertson JR. Use of whole slide
imaging in surgical pathology quality assurance: design and pilot validation
studies. Hum Pathol 2006;37:322–31.
[52] Garcia Rojo M, Felix Conde A, Ordi J, Ruiz Martin J, Corominas JM, Alvarez
Alegret R, et al. Libro blanco de la Anatomía Patológica en España. 2015.
[53] Winokur TS, McClellan S, Siegal GP, Redden D, Gore P, Lazenby A, et al. A
prospective trial of telepathology for intraoperative consultation (frozen
sections). Hum Pathol 2000;31:781–5. doi:10.1053/hupa.2000.8452.
[54] Neel JA, Grindem CB, Bristol DG. Introduction and evaluation of virtual
microscopy in teaching veterinary cytopathology. J Vet Med Educ 2007;34:437–
44. doi:10.3138/jvme.34.4.437.
[55] Mills PC, Bradley AP, Woodall PF, Wildermoth M. Teaching histology to first-year
veterinary science students using virtual microscopy and traditional microscopy:
a comparison of student responses. J Vet Med Educ 2007;34:177–82.
[56] Dee FR, Meyerholz DK. Teaching medical pathology in the twenty-first century:
virtual microscopy applications. J Vet Med Educ 2007;34:431–6.
doi:10.3138/jvme.34.4.431.
[57] Blake C a, Lavoie HA, Millette CF. Teaching medical histology at the University of
South Carolina School of Medicine: Transition to virtual slides and virtual
microscopes. Anat Rec B New Anat 2003;275:196–206. doi:10.1002/ar.b.10037.
[58] Goldberg HR, Dintzis R. The positive impact of team-based virtual microscopy on
student learning in physiology and histology. AJP Adv Physiol Educ 2007;31:261–
Tesis Doctoral. Adela Saco Álvarez
[133]
5. doi:10.1152/advan.00125.2006.
[59] Chen Y-K, Hsue S-S, Lin D-C, Wang W-C, Chen J-Y, Lin C-C, et al. An application of
virtual microscopy in the teaching of an oral and maxillofacial pathology
laboratory course. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
2008;105:342–7. doi:10.1016/j.tripleo.2007.03.020.
[60] Wellnitz U, Fritz P, Voudouri V, Linder A, Toomes H, Schmid J, et al. The validity
of telepathological frozen section diagnosis with ISDN-mediated remote
microscopy. Virchows Arch 2000;437:52–7.
[61] Oberholzer M, Fischer HR, Christen H, Gerber S, Bruhlmann M, Mihatsch MJ, et
al. Telepathology: frozen section diagnosis at a distance. Virchows Arch
1995;426:3–9.
[62] Della Mea V, Cataldi P, Pertoldi B, Beltrami CA. Combining dynamic and static
robotic telepathology: a report on 184 consecutive cases of frozen sections,
histology and cytology. Anal Cell Pathol 2000;20:33–9.
[63] Dawson PJ, Johnson JG, Edgemon LJ, Brand CR, Hall E, Van Buskirk GF.
Outpatient frozen sections by telepathology in a Veterans Administration
medical center. Hum Pathol 2000;31:786–8. doi:10.1053/hupa.2000.8451.
[64] Becker RLJ, Specht CS, Jones R, Rueda-Pedraza ME, O’Leary TJ. Use of remote
video microscopy (telepathology) as an adjunct to neurosurgical frozen section
consultation. Hum Pathol 1993;24:909–11.
[65] Adachi H, Inoue J, Nozu T, Aoki H, Ito H. Frozen-section services by
telepathology: experience of 100 cases in the San-in District, Japan. Pathol Int
1996;46:436–41.
Microscopía virtual en el diagnóstico rutinario y la docencia
[134]
[66] Nakayama I, Matsumura T, Kamataki A, Uzuki M, Saito K, Hobbs J, et al.
Development of a teledermatopathology consultation system using virtual
slides. Diagn Pathol 2012;7:177. doi:10.1186/1746-1596-7-177.
[67] Jones NC, Nazarian RM, Duncan LM, Kamionek M, Lauwers GY, Tambouret RH,
et al. Interinstitutional whole slide imaging teleconsultation service
development: assessment using internal training and clinical consultation cases.
Arch Pathol Lab Med 2015;139:627–35. doi:10.5858/arpa.2014-0133-OA.
[68] Song Y, Treanor D, Bulpitt AJ, Magee DR. 3D reconstruction of multiple stained
histology images. J Pathol Inform 2013;4:S7. doi:10.4103/2153-3539.109864.
[69] Helin HH, Lundin M, Lundin J, Martikainen P, Tammela T, Helin HH, et al. Web-
based virtual microscopy in teaching and standardizing Gleason grading. Hum
Pathol 2005;36:381–6. doi:10.1016/j.humpath.2005.01.020.
[70] Caie PD, Turnbull AK, Farrington SM, Oniscu A, Harrison DJ. Quantification of
tumour budding, lymphatic vessel density and invasion through image analysis
in colorectal cancer. J Transl Med 2014;12:156. doi:10.1186/1479-5876-12-156.
[71] Neil DAH, Roberts ISD, Bellamy COC, Wigmore SJ, Neuberger JM. Improved
access to histopathology using a digital system could increase the organ donor
pool and improve allocation. Transpl Int 2014;27:759–64. doi:10.1111/tri.12320.
[72] Neltner JH, Abner EL, Schmitt FA, Denison SK, Anderson S, Patel E, et al. Digital
pathology and image analysis for robust high-throughput quantitative
assessment of Alzheimer disease neuropathologic changes. J Neuropathol Exp
Neurol 2012;71:1075–85. doi:10.1097/NEN.0b013e3182768de4.
[73] Riber-Hansen R, Vainer B, Steiniche T. Digital image analysis: a review of
Tesis Doctoral. Adela Saco Álvarez
[135]
reproducibility, stability and basic requirements for optimal results. APMIS
2012;120:276–89. doi:10.1111/j.1600-0463.2011.02854.x.
[74] Webster JD, Dunstan RW. Whole-slide imaging and automated image analysis:
considerations and opportunities in the practice of pathology. Vet Pathol
2014;51:211–23. doi:10.1177/0300985813503570.
[75] Nelson D, Ziv A, Bandali KS. Going glass to digital: virtual microscopy as a
simulation-based revolution in pathology and laboratory science. J Clin Pathol
2012;65:877–81. doi:10.1136/jclinpath-2012-200665.
[76] Rodriguez-Urrego PA, Cronin AM, Al-Ahmadie HA, Gopalan A, Tickoo SK, Reuter
VE, et al. Interobserver and intraobserver reproducibility in digital and routine
microscopic assessment of prostate needle biopsies. Hum Pathol 2011;42:68–
74. doi:10.1016/j.humpath.2010.07.001.
[77] Robertson AJ, Anderson JM, Beck JS, Burnett RA, Howatson SR, Lee FD, et al.
Observer variability in histopathological reporting of cervical biopsy specimens. J
Clin Pathol 1989;42:231–8.
[78] McCluggage WG, Walsh MY, Thornton CM, Hamilton PW, Date A, Caughley LM,
et al. Inter- and intra-observer variation in the histopathological reporting of
cervical squamous intraepithelial lesions using a modified Bethesda grading
system. Br J Obstet Gynaecol 1998;105:206–10.
[79] McCluggage WG, Bharucha H, Caughley LM, Date A, Hamilton PW, Thornton
CM, et al. Interobserver variation in the reporting of cervical colposcopic biopsy
specimens: comparison of grading systems. J Clin Pathol 1996;49:833–5.
[80] de Vet HC, Koudstaal J, Kwee WS, Willebrand D, Arends JW. Efforts to improve
Microscopía virtual en el diagnóstico rutinario y la docencia
[136]
interobserver agreement in histopathological grading. J Clin Epidemiol
1995;48:869–73.
[81] Creagh T, Bridger JE, Kupek E, Fish DE, Martin-Bates E, Wilkins MJ. Pathologist
variation in reporting cervical borderline epithelial abnormalities and cervical
intraepithelial neoplasia. J Clin Pathol 1995;48:59–60.
[82] Bergeron C, Ordi J, Schmidt D, Trunk MJ, Keller T, Ridder R. Conjunctive
p16INK4a testing significantly increases accuracy in diagnosing high-grade
cervical intraepithelial neoplasia. Am J Clin Pathol 2010;133:395–406.
doi:10.1309/AJCPXSVCDZ3D5MZM.
[83] Stoler MH, Schiffman M. Interobserver reproducibility of cervical cytologic and
histologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study.
JAMA 2001;285:1500–5.
[84] Bauer TW, Schoenfield L, Slaw RJ, Yerian L, Sun Z, Henricks WH. Validation of
whole slide imaging for primary diagnosis in surgical pathology. Arch Pathol Lab
Med 2013;137:518–24. doi:10.5858/arpa.2011-0678-OA.
[85] Camparo P, Egevad L, Algaba F, Berney DM, Boccon-Gibod L, Comperat E, et al.
Utility of whole slide imaging and virtual microscopy in prostate pathology.
APMIS 2012;120:298–304. doi:10.1111/j.1600-0463.2011.02872.x.
[86] Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJA, ten Kate FJW, et al.
Whole slide images for primary diagnostics of gastrointestinal tract pathology: a
feasibility study. Hum Pathol 2012;43:702–7.
doi:10.1016/j.humpath.2011.06.017.
[87] Gilbertson JR, Ho J, Anthony L, Jukic DM, Yagi Y, Parwani A V. Primary histologic
Tesis Doctoral. Adela Saco Álvarez
[137]
diagnosis using automated whole slide imaging: a validation study. BMC Clin
Pathol 2006;6:4. doi:10.1186/1472-6890-6-4.
[88] Bauer TW, Slaw RJ. Validating whole-slide imaging for consultation diagnoses in
surgical pathology. Arch Pathol Lab Med 2014;138:1459–65.
doi:10.5858/arpa.2013-0541-OA.
[89] Al-Janabi S, Huisman A, Nikkels PGJ, ten Kate FJW, van Diest PJ. Whole slide
images for primary diagnostics of paediatric pathology specimens: a feasibility
study. J Clin Pathol 2013;66:218–23. doi:10.1136/jclinpath-2012-201104.
[90] Rousselet M-C, Michalak S, Dupre F, Croue A, Bedossa P, Saint-Andre J-P, et al.
Sources of variability in histological scoring of chronic viral hepatitis. Hepatology
2005;41:257–64. doi:10.1002/hep.20535.
[91] Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al.
Design and validation of a histological scoring system for nonalcoholic fatty liver
disease. Hepatology 2005;41:1313–21. doi:10.1002/hep.20701.
[92] Regev A, Berho M, Jeffers LJ, Milikowski C, Molina EG, Pyrsopoulos NT, et al.
Sampling error and intraobserver variation in liver biopsy in patients with
chronic HCV infection. Am J Gastroenterol 2002;97:2614–8. doi:10.1111/j.1572-
0241.2002.06038.x.
[93] Robert M, Sofair AN, Thomas A, Bell B, Bialek S, Corless C, et al. A comparison of
hepatopathologists’ and community pathologists’ review of liver biopsy
specimens from patients with hepatitis C. Clin Gastroenterol Hepatol
2009;7:335–8. doi:10.1016/j.cgh.2008.11.029.
[94] Skripenova S, Trainer TD, Krawitt EL, Blaszyk H. Variability of grade and stage in
Microscopía virtual en el diagnóstico rutinario y la docencia
[138]
simultaneous paired liver biopsies in patients with hepatitis C. J Clin Pathol
2007;60:321–4. doi:10.1136/jcp.2005.036020.
[95] Brick KE, Sluzevich JC, Cappel MA, DiCaudo DJ, Comfere NI, Wieland CN.
Comparison of virtual microscopy and glass slide microscopy among
dermatology residents during a simulated in-training examination. J Cutan
Pathol 2013;40:807–11. doi:10.1111/cup.12189.
[96] Reyes C, Ikpatt OF, Nadji M, Cote RJ. Intra-observer reproducibility of whole
slide imaging for the primary diagnosis of breast needle biopsies. J Pathol Inform
2014;5:5. doi:10.4103/2153-3539.127814.
[97] Comperat E, Egevad L, Lopez-Beltran A, Camparo P, Algaba F, Amin M, et al. An
interobserver reproducibility study on invasiveness of bladder cancer using
virtual microscopy and heatmaps. Histopathology 2013;63:756–66.
doi:10.1111/his.12214.
[98] Singson RP, Natarajan S, Greenson JK, Marchevsky AM. Virtual microscopy and
the Internet as telepathology consultation tools. A study of gastrointestinal
biopsy specimens. Am J Clin Pathol 1999;111:792–5.
[99] Farah CS, Maybury TS. The e-evolution of microscopy in dental education. J Dent
Educ 2009;73:942–9.
[100] Diaz-Perez JA, Raju S, Echeverri JH. Evaluation of a teaching strategy based on
integration of clinical subjects, virtual autopsy, pathology museum, and digital
microscopy for medical students. J Pathol Inform 2014;5:25. doi:10.4103/2153-
3539.137729.
[101] Gatumu MK, MacMillan FM, Langton PD, Headley PM, Harris JR. Evaluation of
Tesis Doctoral. Adela Saco Álvarez
[139]
usage of virtual microscopy for the study of histology in the medical, dental, and
veterinary undergraduate programs of a UK University. Anat Sci Educ
2014;7:389–98.
[102] Helle L, Nivala M, Kronqvist P, Gegenfurtner A, Bjork P, Saljo R. Traditional
microscopy instruction versus process-oriented virtual microscopy instruction: a
naturalistic experiment with control group. Diagn Pathol 2011;6 Suppl 1:S8.
doi:10.1186/1746-1596-6-S1-S8.
[103] McCready ZR, Jham BC. Dental students’ perceptions of the use of digital
microscopy as part of an oral pathology curriculum. J Dent Educ 2013;77:1624–
8.
[104] Weaker FJ, Herbert DC. Transition of a Dental Histology Course from Light to
Virtual Microscopy. J Dent Educ 2009;73:1213–21.
[105] Anyanwu GE, Agu AU, Anyaehie UB. Enhancing learning objectives by use of
simple virtual microscopic slides in cellular physiology and histology: impact and
attitudes. Adv Physiol Educ 2012;36:158–63. doi:10.1152/advan.00008.2012.
[106] Krippendorf BB, Lough J. Complete and rapid switch from light microscopy to
virtual microscopy for teaching medical histology. Anat Rec - Part B New Anat
2005;285:19–25. doi:10.1002/ar.b.20066.
[107] Scoville SA, Buskirk TD. Traditional and virtual microscopy compared
experimentally in a classroom setting. Clin Anat 2007;20:565–70.
doi:10.1002/ca.20440.
[108] Sivamalai S, Murthy SV, Gupta T Sen, Woolley T. Teaching pathology via online
digital microscopy: Positive learning outcomes for rurally based medical
Microscopía virtual en el diagnóstico rutinario y la docencia
[140]
students. Aust J Rural Health 2011;19:45–51. doi:10.1111/j.1440-
1584.2010.01176.x.
[109] Marsch AF, Espiritu B, Groth J, Hutchens KA. The effectiveness of annotated (vs.
non-annotated) digital pathology slides as a teaching tool during dermatology
and pathology residencies. J Cutan Pathol 2014;41:513–8.
doi:10.1111/cup.12328.
[110] Stewart J 3rd, Bevans-Wilkins K, Bhattacharya A, Ye C, Miyazaki K, Kurtycz DFI.
Virtual microscopy: an educator’s tool for the enhancement of cytotechnology
students’ locator skills. Diagn Cytopathol 2008;36:363–8. doi:10.1002/dc.20821.
[111] Hamilton PW, Wang Y, McCullough SJ. Virtual microscopy and digital pathology
in training and education. Apmis 2012;120:305–15. doi:10.1111/j.1600-
0463.2011.02869.x.
[112] Rosai J. Digital images of case reports and other articles. Int J Surg Pathol
2007;15:5. doi:10.1177/1084713806296004.
[113] Williams G, Lau A. Reform of undergraduate medical teaching in the United
Kingdom: a triumph of evangelism over common sense. BMJ 2004;329:92–4.
doi:10.1136/bmj.329.7457.92.
[114] Bloodgood RA, Ogilvie RW. Trends in histology laboratory teaching in United
States medical schools. Anat Rec B New Anat 2006;289:169–75.
doi:10.1002/ar.b.20111.
Tesis Doctoral. Adela Saco Álvarez
[141]
Microscopía virtual en el diagnóstico rutinario y la docencia
[142]
Top Related