SAN FRANCISCO Remote review of digital immunohistochemistry slides by pathologists seemed to be as accurate as in-person evaluations by light microscopy for diagnosis of dysplastic nevi, but was less successful for diagnosing early malignant melanoma.
Three pathologists viewed 5560 cases of dysplastic nevi with varying atypia and early melanomas on hematoxylin and eosin- and immunohistochemical-stained slides via telepathology. A remote site hosted the digital slides, and the consulting pathologists evaluated them via a Java-enabled Web browser. The pathologists also evaluated a glass set of slides for the cases.
Preliminary results showed that there was a high concordance rate between digital and light microscopic diagnoses for dysplastic nevi, but a few diagnostic discrepancies were seen between telepathology and microscopy when early malignant melanomas were evaluated, Dr. Jill Buckthal-McCuin of the University of Pittsburgh reported in a poster that was presented at the annual meeting of the American Society of Dermatopathology.
Evaluations by telepathology took more than twice as long to perform as microscopy evaluations, mainly because focusing was slow during telepathology. If the time spent on shipping the glass slides for microscopic evaluation were included, however, telepathology was faster, Dr. Buckthal-McCuin added.
An individual evaluator's level of experience in dermatopathology and experience with telepathology may have been a factor in diagnostic accuracy, she suggested.
A dermatopathology fellow made the correct diagnosis in 21 (36%) of 59 cases evaluated with digital telepathology. A staff dermatologist with 1 year of experience was correct in 19 (45%) of 42 diagnoses via telepathology, and said he was not comfortable evaluating 18 other cases, including 8 cases of suspected melanoma.
An attending dermatopathologist with more than 5 years' experience was correct in 23 (42%) of 55 diagnoses via telepathology.
Each of the evaluators misgraded some dysplastic nevi by one degree, which would not affect decisions regarding treatment.
Overall, the diagnoses that were made via telepathology would have resulted in correct treatment in 90% of the cases evaluated by the fellow, 76% of cases that the junior attending dermatopathologist agreed to evaluate, and 67% of the cases reviewed by the senior attending dermatopathologist.
Increasing demands are being placed on pathologists, and in some settings a trained pathologist in a specific subspecialty is not available on site, Dr. Buckthal-McCuin noted.
Digital telepathology might contribute to the more efficient use of pathologists by allowing real-time review of remote cases, delayed image review, subspecialty reviews, and second opinions in a timely fashion, but more study is needed to confirm the utility of this diagnostic method, she said.
The three pathologists in the study reported not feeling comfortable when the slide was moved at the remote site, because an area of interest often was omitted or out of focus. They agreed that the robotic microscope provided excellent detail and would be useful in nonprimary diagnosis.
The fellow complained that the inability to evaluate the immunohistochemical and hematoxylin and eosin slides together hampered the remote diagnosis by telepathology.
Dr. Buckthal-McCuin and her associates plan to study telepathology evaluation of hematoxylin and eosin and immunohistochemical slides together, and said that they expect the degree of concordance with glass-based evaluation to be similar to the findings in this preliminary study.