OBJECTIVE: To evaluate ultrasensitive (US) measurements of prostate-specific antigen (PSA) level in men with prostate cancer, and to correlate the findings with currently accepted values for PSA recurrence, as PSA is widely accepted as a surrogate marker for disease recurrence after treatment for prostate cancer, and although USPSA assays can detect minute quantities of PSA, the significance of this is unclear. PATIENTS AND METHODS: In all, 225 patients had a radical prostatectomy (RP) and were followed with USPSA measurements. PSA recurrence was defined as two or more consecutive increasing PSA values after RP of > or = 0.200 ng/mL. This was deemed clinically significant if it was associated with adjuvant treatment with a repeated nadir PSA level, or a PSA level that continued to increase on watchful waiting. USPSA values were compared between patients who recurred and those who did not, to determine any association with PSA recurrence and clinical outcomes. RESULTS: There was a PSA recurrence in 21 patients; all had clinical evidence of recurrence. Although the difference in mean USPSA levels was statistically significant for those patients who did and did not recur, the overlap in values invalidated any clinical utility. However, an undetectable level achieved during the follow-up appeared to confer a favourable prognosis. CONCLUSION: After RP patients might have PSA levels detectable by USPSA assays, i.e. <0.1 ng/mL. The amount of 'background noise' produced within this range precludes the ability to use this test as a clinical indicator of disease recurrence. However, undetectable levels appear to confer a favourable prognosis.