Prognostic factors of persistently detectable PSA after radical prostatectomy.

OBJECTIVES To determine predictive factors of detectable prostate-specific antigen (PSA) in patients submitted to radical prostatectomy (RP) and to define the prognostic role of this event. METHODS A total of 318 patients who underwent RP between 2002 and 2007 were selected from our prospective database. Selection criteria were: no neo-adjuvant therapy; surgical specimens analyzed and reviewed according to a standardized protocol by two pathologists; clinical stage T1,T2 or T3 N0; pathological stage T2-3/N0-1. RESULTS Median age was 65. 22 years. All patients had a PSA greater than 20 ng/mL (6.9%). Fifty-six patients had poorly differentiated prostate cancer at biopsy (17.6%) and 77 after pathological examination. Cancer stage was cT2/3 in 128 (40.2%) patients, pT3 in 79 (24.8%) patients and pN1 in 20 patients (6.2%). Surgical margins were positive in 89 cases (28%). Thirty-three of the 318 patients had detectable PSA (10.3%) after RP. Multivariate analysis confirmed PSA (odds ratio 3.07; P = 0.0008), pT3a/b stage (odds ratio 2.72; P = 0.0466) and nodal metastasis (odds ratio 5.68; P = 0.0060) as independent predictors of detectable PSA after RP. Detectable PSA had a great impact on prognosis. Twenty-four of these 33 patients experienced a PSA progression and needed a second treatment. In a multivariate model, detectable PSA functioned as an independent predictor of PSA progression (hazard ratio 4.54; P = 0.0000). CONCLUSIONS In our experience, a detectable PSA after RP can be predicted by preoperative PSA, pathological stage and nodal status. Moreover, it represents a significant risk factor of PSA progression. The strong imbalance towards risk factors of systemic disease supports the use of hormonal therapy in case of progression.