Department of Oncology and Radiotherapy, Val de Grace Hospital, Paris, France.Department of Pathology, Val de Grace Hospital, Paris, France.
This study aimed at determining whether virtual microscopy improves the accuracy in the pathological examination of prostate needle biopsies regarding maximum tumor length, percentage of positive cores, and Gleason grading. We assessed a series of 816 prostate needle biopsy cores in 68 consecutive patients with prostate adenocarcinoma. Biopsy specimens were reviewed using conventional examination. Then, slides were converted to whole slide imaging (Olympus BX51). Tumor was measured, and Gleason score was assigned using the OlyVia software. Optically evaluated pathological features were compared with digital findings to determine whether one of these two methods for the assignment of a preoperative Gleason score is appropriate for predicting the definitive Gleason score of radical prostatectomy. When comparing optical and digital measurements, maximum tumor length in biopsy cores and percent prostate needle biopsy with cancer showed no significant difference. The mean variation in the measurement of tumor length was 2.65mm per biopsy. Among 240 biopsy cores involved with cancer, the concordance rate for Gleason score assignment was 75.8% (κ=0.49, good agreement). When considering the higher Gleason score assignment as the score for the entire case (ISUP 2005), the concordance rate was 69.1% (κ=0.46, good agreement). When comparing the biopsy scores with the definitive score of radical prostatectomy, the concordance rate was significantly increased from 54.4% for conventional examination (κ=0.23, marginal agreement) to 66.2% for virtual slide examination (κ=0.42, good agreement). Virtual microscopy does not compromise, but might improve, the accuracy of grading in prostate needle biopsies. This requires further assessment.
Chargari C, Comperat E, Magné N, Védrine L, Houlgatte A, Egevad L, Camparo P. Are you the author?
Reference: Pathol Res Pract. 2011 May 13. Epub ahead of print.