Prostate-specific antigen associates with extensive lymph node invasion in high-risk prostate cancer

Objective: To evaluate clinical predictors of lymph node invasion (LNI) in patients with high-risk prostate cancer undergoing radical prostatectomy (RP) with extended pelvic lymph node dissection (ePLND). Methods: A contemporary cohort of 116 patients, who underwent ePLND during RP, was retrospectively evaluated. Patients were classified into 3 groups including cases without LNI (group 1), with 1 to 3 positive nodes (group 2; limited LNI), and with more than 3 positive nodes (group 3; extensive LNI). The multinomial logistic regression model (multivariate analysis) evaluated the risk of LNI. Results: Overall, 30 patients (25.9%) had LNI, which was limited in 17 cases (14.7%) and extensive in 13 subjects (11.2%). Median prostate-specific antigen (PSA) was higher in cases with limited (11.4 ng/mL) or extensive (23.5 ng/mL) LNI than cases without (7.3 ng/mL) and the difference was significant (p <.0001). Median proportion of biopsy-positive cores was higher in limited (0.64) or extensive (0.54) LNI than cases without (0.34) and the difference was significant (p < .0001). The distribution of other factors did not show any significant difference among the groups. On multivariate analysis, only higher values of PSA significantly affected the odds of extensive LNI when compared to cases without (odds ratio, 1.054; p = .005); PSA showed a fair discrimination power (area under the curve 0.792). Conclusion: PSA was the only independent predictor of extensive LNI and could be an important preoperative factor for stratifying high-risk patients