EVALUATING INDIVIDUAL-LEVEL ASSOCIATION BETWEEN RADIOGRAPHIC PROGRESSION-FREE SURVIVAL (PFS) AND OVERALL SURVIVAL (OS) IN METASTATIC-HORMONE SENSITIVE PROSTATE CANCER (MHSPC): AN INDIRECT ILLNESS-DEATH MODELING APPROACH
Author(s)
Paul Serafini, BA, Mir Sohail Fazeli, MD, PhD, Mir-Masoud Pourrahmat, MSc, Murat Kurt, BS, MS, PhD;
Evidinno Outcomes Research Inc., Vancouver, BC, Canada
Evidinno Outcomes Research Inc., Vancouver, BC, Canada
OBJECTIVES: Therapeutic advances in mHSPC prolonged survival, delaying the availability of mature OS data in randomized controlled trials (RCTs) and prompting evaluation of rPFS as a surrogate for OS. This study evaluated individual-level association between rPFS and OS using published survival data from mHSPC trials.
METHODS: Reconstructed rPFS and OS data from RCTs identified through a published systematic literature review (Shore et al. 2025) were pooled separately across all arms and studies, and analyzed in an illness-death model previously validated for estimation individual-level association between rPFS and OS in mHSPC (Serafini et al. 2025) versus a published benchmark (Halabi et al. 2024). Proposed framework elicited pre-progression death (PPD) probability and hazard rate of an exponentially distributed post-progression survival (PPS) to simulate paired rPFS-OS data from the modeled long-term rPFS and OS curves. Correlation between rPFS and OS was measured by Pearson’s r, Spearman’s rho, and Kendall’s tau; sensitivity of which were assessed with respect to model choice for rPFS projections, changes in evidence base according to availability of rPFS and OS data from the RCTs, and pre-specified variations in PPD probability (±10%), hazard rate of PPS (±10%) and time horizon (±60 months)
RESULTS: Evidence base consisted of 24 RCTs published between 1995-2023. Of these seven (n=6,555) reported rPFS data, 23 (n=18,631) reported OS data, and six (n=6,493) reported both rPFS and OS data. Among RCTs reporting both endpoints, predicted Pearson’s r, Spearman’s rho, Kendall’s tau were 0.998, 0.977, and 0.921, respectively with negligibly narrow 95% CIs. In sensitivity analyses, ranges for r, rho, and tau were (0.985-0.998), (0.919-0.980), and (0.786-0.922), respectively.
CONCLUSIONS: Consistent with Halabi et al. (2024), results demonstrate a strong individual-level association between rPFS and OS in mHSPC, supporting rPFS as a surrogate for OS while reflecting the impact of changes in evidence base and mHSPC treatment landscape on this relationship.
METHODS: Reconstructed rPFS and OS data from RCTs identified through a published systematic literature review (Shore et al. 2025) were pooled separately across all arms and studies, and analyzed in an illness-death model previously validated for estimation individual-level association between rPFS and OS in mHSPC (Serafini et al. 2025) versus a published benchmark (Halabi et al. 2024). Proposed framework elicited pre-progression death (PPD) probability and hazard rate of an exponentially distributed post-progression survival (PPS) to simulate paired rPFS-OS data from the modeled long-term rPFS and OS curves. Correlation between rPFS and OS was measured by Pearson’s r, Spearman’s rho, and Kendall’s tau; sensitivity of which were assessed with respect to model choice for rPFS projections, changes in evidence base according to availability of rPFS and OS data from the RCTs, and pre-specified variations in PPD probability (±10%), hazard rate of PPS (±10%) and time horizon (±60 months)
RESULTS: Evidence base consisted of 24 RCTs published between 1995-2023. Of these seven (n=6,555) reported rPFS data, 23 (n=18,631) reported OS data, and six (n=6,493) reported both rPFS and OS data. Among RCTs reporting both endpoints, predicted Pearson’s r, Spearman’s rho, Kendall’s tau were 0.998, 0.977, and 0.921, respectively with negligibly narrow 95% CIs. In sensitivity analyses, ranges for r, rho, and tau were (0.985-0.998), (0.919-0.980), and (0.786-0.922), respectively.
CONCLUSIONS: Consistent with Halabi et al. (2024), results demonstrate a strong individual-level association between rPFS and OS in mHSPC, supporting rPFS as a surrogate for OS while reflecting the impact of changes in evidence base and mHSPC treatment landscape on this relationship.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
CO28
Topic
Clinical Outcomes
Topic Subcategory
Relating Intermediate to Long-term Outcomes
Disease
SDC: Oncology