Prostate-Specific Antigen Levels Posttreatment May Not Predict Overall Survival

THURSDAY, Aug. 31, 2023 (HealthDay News) -- Treatments that reduce the risk for biochemical recurrence (BCR) based on rising prostate-specific antigen levels after radiotherapy do not necessarily improve a patient’s long-term overall survival (OS), according to a study published online Oct. 28 in the Journal of Clinical Oncology.

Soumyajit Roy, M.B.B.S., from Rush University in Chicago, and colleagues conducted a meta-analysis of 11 randomized controlled trials that used treatment intensification strategies with definitive radiotherapy known to reduce BCR to evaluate its surrogacy. For the meta-analysis, individual patient data were collected for each trial for an overall sample of 10,741 patients. The Prentice criteria and the two-stage meta-analytic approach were used to evaluate the surrogacy of BCR-free survival (other-cause mortality as an event) and time to BCR (other-cause mortality as a competing risk).

The researchers found that dose escalation, addition of short-term androgen deprivation therapy (ADT), and prolongation of ADT duration significantly improved BCR (hazard ratios [HRs], 0.71, 0.53, and 0.54, respectively). OS was significantly improved by adding short-term ADT (HR, 0.91) and prolonging ADT (HR, 0.86). However, dose escalation did not improve OS (HR, 0.98; 95 percent confidence interval [CI], 0.87 to 1.11). In all three groups, BCR at 48 months was associated with inferior OS (HRs, 2.46, 1.51, and 2.31, respectively). But when adjusting for BCR at 48 months, there was no significant treatment effect on OS (HR, 1.10 [95 percent CI, 0.96 to 1.27]; HR, 0.96 [95 percent CI, 0.87 to 1.06]; and HR, 1.00 [95 percent CI, 0.90 to 1.12], respectively). For BCR-free survival and OS, the patient-level correlation ranged from 0.59 to 0.69 and time to BCR and OS ranged from 0.23 to 0.41. The correlation of treatment effect on BCR-free survival and time to BCR had R² levels on OS of 0.563 and 0.160, respectively.

“There was a poor to modest correlation between these BCR-based surrogate end points and OS, suggesting that BCR-based end points are not appropriate surrogate end points,” write the authors.

Abstract/Full Text (subscription or payment may be required)