Diagnostic concordance in biopsies of deceptive prostatic carcinoma.

OBJECTIVE To explore the diagnostic concordancy of surgery pathologists in deceptive prostatic carcinoma biopsies (DPCs) purposely selected as of difficult interpretation. SETTING Ten experienced pathologists belonging to five tertiary care centers in Mexico City participated. MATERIAL AND METHODS A total of 25 slides (16 DPCs and 9 benign biopsies) were marked and circulated among the participants. They were asked to classify each slide in one of five categories ranging from carcinoma to benign hyperplasia using the criteria of Epstein. The 16 DPCs had been confirmed by immunohistochemistry studies and by review of two expert uropathologists from the MD Anderson Medical Center in Houston TX. A weighted kappa was use to evaluate the concordancy of the pathologists. RESULTS The concordancy with the experts opinion ranged from substantial in one pathologist (KW = 0.77) to bad in another (KW = -0.07) and the mean kappa was 0.32. In two of the three institutions with more than one pathologist, the intrainstitucional concordancy improved (KW of 0.57-0.64). The pathologists were unable to reach a consensus (7 or more agreeing) in 12 of the 25 slides. Overall, 12% of the diagnosis were considered uncertain, 29% were incorrect (45% of DPCs and 19% of benign) and only 59% were correct (55% of DPCs and 67% of benign). CONCLUSION Efforts should be made to improve the diagnostic accuracy of pathologists in DPCs.