Multivariate Analysis of Factors Affecting Biochemical Recurrence After Radical Prostatectomy

Abstract

Objective: This study aims to evaluate the effects on biochemical recurrence (BCR) of the following proposed prognostic factors after radical prostatectomy (RP): patients' clinical T stage, Gleason grade group (GG) of RP specimen, technique of operation used (open RP vs. robot-assisted laparoscopic RP), presence of positive surgical margin (PSM), length of PSM, GG at PSM, extraprostatic extension (EPE) at PSM, and presence of detectable PSA at 4-6 weeks after RP. It also aims to identify which among the aforementioned variables are independent predictors of risk for BCR.

Patients and Methods: This is a retrospective study. Included in the study were patients who underwent RP (Open and Robot-assisted Laparoscopic technique) at two tertiary hospital branches of an academic medical center from April 2009 to December 2015 with histopathology reports read by a single urologic pathologist and with complete follow- up for at least one year. Excluded were those who underwent RP but without complete follow- up. Using Pearson chi-square and z-test with level of significance set at 0.05, the clinicopathologic variables including: patients clinical stage, GG of RP specimen, length of PSM, GG at positive margins, presence of EPE at positive margins, and presence of detectable PSA after the surgery were assessed in order to know which among these factors were predictive of BCR. Multinomial regression analysis was also used to identify which among the variables were independent predictors of risk for BCR.

Results: A total of 165 patients underwent RP from April 2009 to December 2015, among which 72 patients were eligible for inclusion in the final analysis. Clinical T2 stage was found to be a predictor of BCR with odds ratio of 13.000 (95%CI: 3.705 - 45.620; p < 0.001) as compared to stage T1. GG of final histopathology report of prostatectomy specimen was found to be a predictor of BCR, as those with grade groups 4 and 5 had significantly increased risk of BCR with odds ratio of 70.778 (95%CI: 8.207 - 610.426; p < 0.001) as compared to those with grade groups 1 to 3. Patients with positive margins had increased risk of BCR, with odds ratio of 13.458 (95%CI: 13.472 - 52.171; p < 0.001) compared to those with negative margins. GG at the PSM was found to be a predictor of BCR, with a grade grouping of 4 or 5 at the positive margin predicting BCR with odds ratio of 20.625 (95%CI: 2.241 - 189.847; p = 0.008) as compared to grade grouping of 1 or 2 at the margin. Detectable PSA after RP was found to be a predictor of BCR, with odds ratio of 115.000 (95%CI: 19.457 - 679.712; p < 0.001) as compared to undetectable PSA after RP. Technique of RP (p = 0.177), measured length of PSM (p = 0.713), and EPE at PSM (p = 0.146) were not found to predict BCR. Furthermore, clinical T stage (p = 0.007) and detectable PSA after RP (p < 0.001) were found to be independent predictors of BCR among the risk factors examined.

Conclusion: Of the independent variables examined, clinical T stage, GG of RP specimen, presence of PSM, GG at positive margins, and detectable PSA were found to be significant predictors of BCR. Technique of RP, measured length of PSM, and EPE at PSM were not found to predict BCR. Furthermore, multivariate analysis showed that only clinical T stage and detectable PSA after RP were independent predictors of BCR. Attentive assessment of these predictors in the preoperative period should aid the urologist in clinical decision-making and in advising patients regarding their prognosis.