Objective: To examine the usefulness of clinical stage, tumour differentiation and prostate-specific antigen (PSA) level, alone and in combination, to predict regional nodal metastases in individual patients with localized prostate cancer.
Patients and methods: The usefulness of digital rectal examination (DRE), biopsy Gleason sum and PSA, alone and in combination, to predict nodal metastases in an individual patient was examined. The study included 689 patients who had laparoscopic or open pelvic lymph node dissection for clinical stage T1-3 prostate cancer. The Kruskal-Wallis test, Mantel-Haenszel test, chi-squared test and logistic regression were used for continuous, ordinal, categorical, and multivariate analysis, respectively.
Results: Of the 689 patients who underwent radical prostatectomy, 52 (8%) had nodal metastases. Although clinical stage, DRE, pre-operative PSA level and biopsy Gleason sum were significantly related in the univariate analysis, only pre-operative PSA level and biopsy Gleason sum were significant predictors of lymph node status in a multivariate analysis. However, based on a receiver operating characteristic curve, a model with satisfactory sensitivity and specificity could not be obtained.
Conclusion: Current estimations of primary prostate cancer biology using pre-operative PSA level, clinical stage and biopsy Gleason sum are not sufficiently sensitive to predict nodal metastases, and pelvic lymphadenectomy remains the definitive method of detection.