Prostate cancer and radiation therapy--the message conveyed by serum prostate-specific antigen

Prostate-specific antigen (PSA) is a powerful pretreatment prognosticator and a sensitive post-treatment outcome measure for clinically localized prostate cancer treated with radiation therapy. Today, the pretreatment serum PSA level appears to supersede both grade and T-stage as a determinant of outcome. This study was undertaken to attempt a reconciliation between the old (pre-PSA) and the new (PSA) data-in particular to address the question of why stage and grade apparently play so little role in this PSA era.

We analyzed the outcome of two cohorts of men with T1-T4, NO, or NX, MO prostate cancer, one group (648 patients) treated and followed in the pre-PSA era (1966-1988), another group (707 patients) treated and followed in the PSA era (1987-1993)--who received definitive radiation as their only initial treatment. The patterns of relapse and prognostic factors for these groups were compared and contrasted using univariate and multivariate techniques.

At a median follow-up of 6.5 years, the relapse patterns in the pre-PSA series were: local in 109 (17%), nodal in 17 (3%), and distant metastatic in 186 (29%). Actuarial local and metastatic rates at 5 years were 13 and 26%, respectively. Local recurrence was only weakly predictable, Gleason grade being the only significant, albeit weak, covariate. Metastatic failure, however, was highly significantly and meaningfully correlated with Gleason grade and T-stage. Because metastasis was the most common adverse end point in this series, overall freedom from progression also correlated with grade and stage. At a median follow-up of 31 months, the patterns of failure in the PSA series were: local in 77 (11%), nodal in 3 (< 1%), and distant metastatic in 24 (3%). Actuarial local and metastatic rates at 5 years were 30 and 6%, respectively. Local recurrence was highly and meaningfully correlated with pretreatment PSA level, which was the only significant determinant of this end point. Metastatic failure was highly correlated with Gleason grade and T-stage, with PSA playing a much lesser, though significant role. The inversion of failure patterns (local vs. distant) between the two series was striking. The high incidence of local failure in the PSA series was almost entirely related to positive prostatic biopsies pursuant to the investigation of the postradiation rising PSA profile. Of the 77 local recurrences, 69 (90%) were identified in this way. Among 99 men with rising PSA values who underwent investigation (CT scans, bone scans, and biopsies), disease was found in 86, and the patterns of disease in these 86 were: local only in 62 (72%), local and metastatic in 7, and metastatic in 17 (20%). The most common event in the PSA series was the rising PSA profile, and this, too, strongly correlated with the pretreatment PSA level.

Based on our earlier finding that the major source of pretreatment serum PSA in patients with clinically localized disease is the primary tumour itself and on the findings in the present report, we conclude that the new major message conveyed by serum PSA relates to the primary tumor and its likely outcome. Gleason grade and T-stage remain major determinants of metastatic relapse. The total and permanent eradication of prostate cancer from the prostate with conventional doses of external beam radiation therapy is harder to achieve than generally appreciated.