DETECTION OF PROSTATE-CANCER RELAPSE WITH PROSTATE-SPECIFIC - ANTIGENMONITORING AT LEVELS OF 0.001 TO 0.1 MU-G. L/

Purpose: The development of prostate specific antigen (PSA) assays wit h detection limits of approximately 0.001 mu g./l. is technically feas ible. We examined if serum PSA changes of 0.001 to 0.1 mu g./l. for up to 3 years after radical prostatectomy have any clinical value. Mater ials and Methods: We studied 148 patients with a postoperative PSA of less than 0.1 mu g./l. by a conventional PSA assay. At least 3 serial serum samples were collected per patient along with detailed clinicopa thological features. Serial serum samples were analyzed for PSA with t he ultrasensitive method. Associations between increase in serum PSA a nd clinicopathological features were analyzed with the unconditional l ogistic regression model. Results: After establishing a set of interpr etative criteria, we divided the patients into 51 with biochemical rel apse, 93 who were free of relapse and 4 with equivocal status. Between the groups with and without relapse there was no difference in year o f surgery, age at operation or length of followup. Compared to patient s without relapse, those with biochemical relapse were likely to have positive surgical margins (p < 0.01), larger tumor volumes (p < 0.01), greater preoperative PSA (p = 0.03) and disease extending outside the prostate (p = 0.02). The relative risks for biochemical relapse estim ated by a univariate logistic regression model were 3.1 (95% confidenc e interval 1.39 to 6.82, p < 0.01) for positive surgical margin, 3.4 ( 95% confidence interval 1.46 to 8.13, p < 0.01) for tumor volume, 2.3 (95% confidence interval 1.08 to 5.02, p = 0.03) for high preoperative PSA and 2.7 (95% confidence interval 1.12 to 6.26, p = 0.03) for extr aprostatic tumor extension. At multivariate analysis with the same mod el the associations between positive surgical margins and biochemical relapse (relative risk 2.95, p = 0.04) and tumor volume (relative risk 3.36, p = 0.03) remained significant. These associations were still o bserved when we analyzed a subset of patients classified as having bio chemical relapse based on PSA changes of 0.001 to 0.08 mu g./l. Conclu sions: Increases in postoperative serum PSA at levels of 0.001 to 0.1 mu g./l. after radical prostatectomy are associated with clinicopathol ogical features of poor prognosis. Monitoring postoperative cases with a highly sensitive PSA assay (detection limit 0.001 mu g./l.) could o ffer a simple and effective means of detecting clinically important bi ochemical relapse early after radical prostatectomy. These patients ma y be suitable for early intervention when effective treatments for rel apse become available.