Positive margin length and Gleason grade of tumor focus at the margin predict for biochemical failure after radical prostatectomy in patients with pT2 prostate cancer.
103 Background: While three prospective trials have demonstrated benefit from adjuvant radiation (XRT) after radical prostatectomy (RP) in patients with positive surgical margins (PSM), its use varies amongst physicians. Many rely on clinical acumen to determine the optimal strategy for application of XRT post RP. We aim to determine if the length of PSM and highest Gleason grade (GG) of tumor at the PSM (hGGPSM) can be used to identify patients at greatest risk of biochemical failure (BCF) post RP. Methods: A retrospective review of all RP patients at The University of Texas MD Anderson Cancer Center from 2002 to 2010 was performed. After a single pathologist review, patients with organ confined disease (pT2), pathologic N0/Nx and a PSM were included. BCF was defined as 2 sequential PSA values of ≥0.2 or any detectable PSA prompting XRT. Patients receiving adjuvant XRT or with <12 months follow-up were excluded. Results: 205 patients met the inclusion criteria. Median PSA was 5.3 ng/mL (0.5-33) and median follow-up was 64 months (13-130). The majority were low clinical stage (cT1c: 65%), low (11%)/intermediate (82%) grade and had a single site of a PSM (90%). BCF occurred in 47 patients for a 5 yr BCF free survival (BCFFS) of 69%. PSM length was significantly associated with BCFFS (≤1mm vs >1, p=0.02). When accounting for hGGPSM, Gl 3 tumors were less likely to experience BF (5 yr BCFFS-96%) regardless of PSM length, while BCFFS for Gl >3 tumors were significantly lower dependent upon length of PSM ( ≤1mm vs >1mm, p=0.03). On multivariable analysis length of PSM (p=0.05) and hGGPSM (p=0.007) remained independent predictors of BCF (Table). Conclusions: Length of PSM and hGGPSM are independent predictors of BCF. These should be considered when evaluating patients for adjuvant XRT and in risk stratifying patients in prospective clinical trials. [Table: see text]