BCRFS and TTBCR Are Inconsistent OS Surrogates in Prostate Cancer

"Given that our findings demonstrate the prognostic association between BCR and OS and the relative importance of noncancer-related deaths and cancer-specific deaths in nonmetastatic [advanced prostate cancer], BCR is clearly not an ideal end point for biomarker development or validation in general," according to the study authors.

Although biochemical recurrence (BCR) appeared to be prognostic for overall survival (OS) in patients with localized prostate cancer, BCR-free survival (BCRFS) and time to BCR (TTBCR) did not consistently meet all surrogacy criteria for OS, according to findings from a meta-analysis published in Journal of Clinical Oncology.

With respect to studies assessing radiotherapy dose escalation, investigators highlighted a 29% reduction in the hazard of BCR when patients received dose-escalated radiation (HR, 0.71; 95% CI, 0.63-0.79); notably, dose escalation did not significantly correlate with OS (HR, 0.98; 95% CI, 0.87-1.11). Additionally, adding short-term androgen deprivation therapy (ADT) to radiation significantly reduced the hazard of BCR (HR, 0.53; 95% CI, 0.48-0.59), and death (HR, 0.91; 95% CI, 0.84-0.99). Prolonged ADT also correlated with a significantly lower hazard of BCR (HR, 0.54; 95% CI, 0.48-0.61) and death (HR, 0.86; 95% CI, 0.78-0.94).

In trials assessing low- and high-dose radiotherapy, BCR events at 48 months correlated with a significantly higher risk of death (HR, 2.46; 95% CI, 2.08-2.92) Investigators also reported that BCR events at 48 months correlated with significantly worse OS in trials assessing radiotherapy alone vs radiotherapy plus ADT (HR, 1.51; 95% CI, 1.35-1.70) and those evaluating long-term vs short-term ADT plus radiotherapy (HR, 2.31; 95% CI, 2.04-2.61). BCR events also had similar correlations with distant metastasis-free survival (DMFS) and prostate cancer–specific mortality (PCSM) across all 3 groups of trials.

After adjusting for BCR by 48 months following random assignment, investigators noted a small, nonsignificant effect of combining short-term ADT with radiotherapy on OS (HR, 0.96; 95% CI, 0.87-1.06). Some evidence suggested an OS benefit when prolonging ADT duration with radiotherapy when adjusting for BCR by 36 months (HR, 0.94; 95% CI, 0.85-1.05), although there was no significant effect observed at 48 months (HR, 1.00; 95% CI, 0.90-1.12).

“Overall, these results strongly suggest that BCR-based end points should not be the primary end point of any randomized trial in localized [advanced prostate cancer],” the study authors wrote. “Given that our findings demonstrate the prognostic association between BCR and OS and the relative importance of noncancer-related deaths and cancer-specific deaths in nonmetastatic [advanced prostate cancer], BCR is clearly not an ideal end point for biomarker development or validation in general.”

Investigators of this meta-analysis evaluated patient data from 11 trials assessing radiotherapy dose escalation, ADT use, and ADT prolongation to determine the surrogacy of BCR-based end points for OS via different surrogacy analysis methods. Evaluating BCR’s surrogacy involved applying the Prentice criteria and the two-stage meta-analytic approach.

To meet the Prentice criteria, treatment needed to have a significant impact on the true end point and the intermediate clinical end point. Additionally, there needed to be a significant relationship between the intermediate clinical end point and the true end point, and the impact of treatment on the true end point had to be mediated by the effect of treatment on the intermediate clinical end point.

The study’s primary end point was the extent to which TTBCR was a surrogate intermediate clinical end point for OS in patients receiving radiotherapy with or without ADT for localized advanced prostate cancer. Investigators also assessed BCRFS.

Investigators of the meta-analysis assessed data from 10,741 patients, including 3639 enrolled on trials evaluating radiotherapy dose escalation, 3930 on trials investigating ADT use, and 3772 on trials assessing prolongation of ADT. The median age in the overall population was 70 years (interquartile range [IQR], 65-74). Additionally, 73% of patients had cT1 or cT2 disease, 85% had a Gleason score of 7 or less, and the median baseline prostate-specific antigen level was 11.1 ng/mL (IQR, 7.1-18.0).

According to findings from the meta-analytic approach, the Kendall’s tau for OS ranged from 0.589 (95% CI, 0.587-0.591) to 0.693 (95% CI, 0.690-0.694) with respect to BCRFS, and from 0.233 (95% CI, 0.227-0.235) to 0.414 (95% CI, 0.408-0.415) with respect to TTBCR.

Reference

Roy S, Romero T, Michalski JM, et al. Biochemical recurrence surrogacy for clinical outcomes after radiotherapy for adenocarcinoma of the prostate. J Clin Oncol. Published online August 28, 2023. doi:10.1200/JCO.23.00617