Metastasis-free survival as a surrogate for OS in men receiving salvage radiation therapy for recurrence after prostatectomy

October 2022 Clinical Practice Andrea Enguita

Metastasis-free survival (MFS) is a validated surrogate endpoint for overall survival (OS) in men treated for localised prostate cancer. Using the data of NRG/RTOG 9601, MFS appeared to be a strong surrogate for OS for recurrent prostate cancer as well, while prostate-specific antigen-based endpoints are weak surrogate endpoints and should not be used in this setting.


“Surrogate markers for overall survival are important tools in clinical studies, although their validation is sometimes difficult and a surrogate marker (e.g. DFS or progression-free survival) does not always translate into an overall survival benefit. This study shows that metastasis-free survival is a surrogate marker for overall survival in prostate cancer according to the Prentice criteria.”

The Intermediate Clinical Endpoints in Cancer of the Prostate (ICECaP) working group identified metastasis-free survival (MFS) as a valid surrogate for overall  survival (OS) in men treated for localised prostate cancer when assessed at both the individual patient level and trial level. However, surrogate endpoints are disease state-specific and may not function equally well across disease states. Therefore, it remains unknown if this holds true in biochemically recurrent disease after radical prostatectomy. Using the data from the phase III NRG/RTOG 9601 trial, this study aimed to determine the performance of multiple intermediate clinical endpoints (ICEs) as possible surrogate endpoints for OS in men receiving salvage radiation therapy for recurrent prostate cancer after prostatectomy.

Study design

This secondary analysis used data from patients enrolled on NRG/RTOG 9601. This phase III trial randomly assigned 760 men with cancer recurrence after prostatectomy to salvage radiation plus two years of antiandrogen therapy with bicalutamide or placebo. The primary endpoint was OS. ICEs assessed for surrogacy were biochemical failure (BF) as defined in NRG/RTOG 9601 (BF1, prostate-specific antigen (PSA) nadir +0.3-0.5 ng/mL or initiation of salvage hormone therapy), as defined in NRG/RTOG 0534, (BF2,  PSA nadir +2 ng/mL), and as defined by the American Urological Association (PSA of 0.2 ng/mL followed by a confirmatory value of ≥0.2 ng/ml), distant metastasis (DM), and MFS.

Potential surrogacy for OS of the ICEs was assessed with the Prentice criteria and a two-stage meta-analytic approach. The Prentice method consists of four criteria that must be satisfied to establish surrogacy: 1) the treatment should have a significant effect on the true endpoint, 2) there must be a significant treatment effect on the surrogate endpoint, 3) a significant association between the surrogate and the true endpoint is required, and 4) the impact of the treatment effect on the true endpoint should be fully captured by the surrogate. The meta-analytic approach requires two conditions: 1) that the ICE correlates with OS and 2) that the treatment effect on the ICE and OS are correlated. Condition one was assessed at the patient level with Kendall’s t and condition two by repeatedly randomly dividing the entire trial cohort into ten pseudo-trial centres and calculating the average R2 between treatment hazard ratios for ICE and OS.


In total, 421 men experienced BF1 (169 vs. 252 in the antiandrogen and placebo arms, respectively), whereas 234 men met the BF2 definition (90 vs. 144), and 483 men experienced BF3 (178 vs. 305). In total, 156 men developed distant metastases (63 vs. 93), and there were 239 deaths from any cause (108 vs. 131).

Assessment with the Prentice criteria: The addition of antiandrogen therapy to stereotactic radiotherapy (SRT) improved OS vs. SRT alone (HR[95%CI]: 0.77[0.59-0.99]; p= 0.04),  thus satisfying criterion one. BF1 (HR[95%CI]: 0.48[0.40-0.59]; p< 0.001], BF2 (HR[95%CI]: 0.48[0.37-0.62]; p< 0.001), BF3 (HR[95%CI]: 0.29[0.24-0.35]; p< 0.001), DM (HR[95%CI]: 0.63[0.46-0.87]; p= 0.005) and MFS (HR[95%CI]: 0.73[0.58-0.92]; p= 0.006) were all improved by the addition of antiandrogen therapy to SRT, thus satisfying the second criterion. To assess the third criterion, the association between each ICE and OS was determined. BF1, BF2, DM, and MFS had a statistically significant association with OS at 3, 5, and 7 years post-treatment, whereas BF3 did not. BF1, BF2, DM, and MFS also met criterion 4.  

Assessment with two-stage meta-analytic approach: For condition one, there was strong correlation between MFS and OS (t= 0.86), moderate correlation between DM and OS (t= 0.66), and weaker correlation between BF1 (t= 0.25), BF2 (t= 0.40), and BF3 (t= 0.14) and OS. Similarly for condition two, the treatment effect of antiandrogen therapy on MFS and OS were correlated (average R2 across 500 permutations = 0.67), but this was not true for DM and OS (R2= 0.18), BF1 and OS (R2= 0.09), BF2 and OS(R2= 0.12), nor BF3 and OS (R2= 0.13).


In the setting of a retrospective analysis of a completed phase III randomised trial (NRG/RTOG 9601), MFS appeared to be a strong surrogate for OS for recurrent prostate cancer in men receiving post-prostatectomy SRT, consistent with data in the primary treatment setting. Although PSA-based endpoints are prognostic, they are weakly correlated  with OS at the patient level, and the treatment effect on PSA-based end points are weakly correlated to the treatment effect on OS from antiandrogen therapy. Using MFS as a surrogate endpoint for OS may have the potential to expedite future clinical trial completion in studies assessing SRT for recurrent prostate cancer after prostatectomy.


Jackson WC, Tang M, Schipper MJ, et al. Biochemical Failure Is Not a Surrogate End Point for Overall Survival in Recurrent Prostate Cancer: Analysis of NRG Oncology/RTOG 9601. J Clin Oncol. 2022;40(27):3172-9.