Pathologic reassessment of prostate cancer surgical specimens before molecular retrospective studies.

PURPOSE The retrospective use of prostate cancer tissue is crucial to design tumor marker prognostic studies. We hypothesize that comparison between recent and more historical cases could introduce biases due to stage and grade migration upon time. DESIGN We reviewed 544 margin-free specimens from patients treated for clinically localized prostate cancer by radical prostatectomy between 2000 and 2005. One hundred and ninty-two patients that underwent biochemical recurrence after surgery were matched with 352 patients without progression, according to pretreatment prostate specific antigen, Gleason score, pathologic stage, and follow-up duration (at least 36 months). RESULTS The reassessment led to histopathologic reclassification of standard prognostic variables in 15% of cases, including modifications in the Gleason score (n = 63), pathologic stage (n = 12), and margin status (n = 30). Most discrepancies with the initial reports are explained either by differences in the scoring system upon time, or by the exam of additional tissues sections. The impact of reclassification led to increase adverse prognostic factors more frequently in the group of patients with progression (Chi(2), P < 0.0001). CONCLUSION Careful reassessment of prostate cancer samples should be mandatory before molecular prognostic studies to ensure a more uniform pathologic evaluation, and might be reported in the "recommendations for tumor markers prognostic studies" (REMARK).