[861] The Predictive Power of Prostate Cancer Extent Measurements in a Large Watchful Waiting Cohort

R Rajab, G Fisher, L Ambroisine, CS Foster, H Moller, M Kattan, V Reuter, P Scardino, J Cuzick, DM Berney. Barts and The London School of Medicine, London, United Kingdom; University of Liverpool; Memorial Sloan Kettering Hospital, New York; Kings College, London, United Kingdom; The Cleveland Clinic, Cleveland, United Kingdom

Background: Decisions on radical or conservative treatment for prostate cancer are difficult. Cancer extent in prostate biopsies can be measured by a number of different methods. We wished to examine these methods in a large cohort of conservatively treated prostate cancers with considerable follow up information available, to determine the optimum method for predicting disease outcome.
Design: Patients diagnosed with gland confined prostate cancer between 1990 and 1996 and managed conservatively were identified from U.K cancer registries by the Transatlantic Prostate Group. The clinical end point was death from prostate cancer. Diagnostic specimens were reviewed and measurements of disease burden were made: fraction of positive chips in TURP specimens, fraction of positive cores in needle biopsies and the cumulative percentage of cancer in needle cores. The percentage of cancer in needle cores was measured. These assessments of disease burden were correlated with prostate cancer related deaths on univariate and multivariate analysis with Gleason score.
Results: 1656 patients had histological specimens available for review. Mean follow up was 117 months. Trans-urethral specimens accounted for 911 cases and 745 were needle core biopsies. On univariate analysis, the fraction of positive chips was the strongest indicator of outcome (X²= 145.83). Biopsy cores showed a significant but lower association on univariate analysis: fraction of positive cores (X²=26.24) and fractional length of tumour involvement (X²=22.12). In multivariate analysis when compared with Gleason score, the extent of disease was still significant for the TURP specimens and also for the fraction of positive cores, but not for the fractional percentage of disease on biopsy.
Conclusions: Measurements of disease burden are helpful in predicting outcome in conservatively treated prostate cancers and could influence decision making for radical therapy. With long follow-up data available, volume of disease on prostatic chips was considerably more informative than that from biopsy material. However the fraction of positive biopsies was still a significant indicator of outcome on multivariate analysis and had better predictive ability than measurements of the percentage of core involvement.
Category: Genitourinary (including renal tumors)

Tuesday, March 10, 2009 2:15 PM

Platform Session: Section A, Tuesday Afternoon