Virtual Microscopy in Melanoma Synoptic Reporting; a Validation and a Comparative Study
A Al Habeeb, A Evans, D Ghazarian. University Health Network, Toronto, ON, Canada
Background: The current melanoma reporting relies heavily on proper synoptic reporting for more accurate, complete and reproducible reports. Aims: 1. To evaluate the current virtual microscopy (VM) in filling synoptic reports for melanoma cases. 2. To assess any possible impact of 20x scans vs. 40x scans on the accuracy.
Design: 10 consecutive cases of melanoma were collected. A total of 12 slides (10 H&E and 2 IHC slides) were scanned at both 20x and 40x using aperio scanscope CS. The cases were evaluated separately by 2 dermatopathologists. The results were evaluated.
Results: The melanoma cases were 1 in-situ, 2 nodular, 3 superficial spreading, one lentigo maligna melanoma, and 3 undetermined. Clark's level ranged from 1 to level 4. The melanoma thickness ranged from 0.0-4.3 mm (median 1.62mm). 6 cases were incompletely excised. There was one ulcerated case. None showed perineural or lymphovascular invasion or satellitosis. The mitotic index was reported in 7 cases ranging from 7-23/ mm ² with a median of 15.6. Pathologist A recorded the type of melanoma as originally reported in 9/10 cases. There was a discrepancy in 1 case were the type was recorded as superficial spreading instead of lentigo maligna melanoma. There was 100% concordance rate in Clark's level. The median thickness was 1.63 mm with no significant differences between 20x vs. 40x. All cases were reported as non-ulcerated missing one ulcerated melanoma. The margin status concordance rate was 100%. The mitotic index median was 5.4 (20x) vs. 6.0 (40x). Pathologist B reported melanoma types as originally reported in 9/10 cases with one case called lentigo maligna melanoma instead of superficial spreading melanoma. Clark's level estimation concurred with the original reports in all cases. The median thickness was 1.7 mm. One ulcerated melanoma was missed. The margin status concordance rate was 100%. The mitotic index median was 14 (x20) vs. 6.2 (x40).
Conclusions: The current system of virtual microscopy produces excellent quality images that make melanoma synoptic reporting feasible at both 20x and 40x scans. Although, in most cases 20x scans are sufficient, 40x scans can be resorted to in difficult cases for a far superior image quality. The current user interface (mouse) is not convenient specially when dealing with high volume. Also automated image analysis could play a major role in solving mitotic index evaluation, which inherently carries a high intra- and inter-observer variability.
Wednesday, March 24, 2010 1:00 PM
Poster Session VI # 78, Wednesday Afternoon