The Relationship Between Progression-Free Survival and Overall Survival in Relapsed/Refractory Multiple Myeloma: A Meta-Regression of Clinical Trial Data
Author(s)
Svenja Petersohn, PhD1, Sonja Kroep, PhD1, Cameron Williams, PhD1, Molly Purser, PhD2, Gbenga Kazeem, PhD3, Simon McNamara, PhD3;
1OPEN Health HEOR & Market Access, Rotterdam, Netherlands, 2GSK, Upper Providence, PA, USA, 3GSK, Stevenage, United Kingdom
1OPEN Health HEOR & Market Access, Rotterdam, Netherlands, 2GSK, Upper Providence, PA, USA, 3GSK, Stevenage, United Kingdom
Presentation Documents
OBJECTIVES: Survival outcomes have improved in patients with relapsed/refractory multiple myeloma (RRMM) with the introduction of novel therapies, resulting in increased time to obtain mature overall survival (OS) data. Therefore, there is a need for identification of surrogate endpoints that support estimation of OS. This meta-regression investigated progression-free survival (PFS; hazard ratios [HRs] and medians) as a surrogate for OS based on published trials in second-line or later (2L+) RRMM.
METHODS: A systematic literature review identified efficacy and safety data from phase 2/3 RRMM clinical trials (2008 to February 2024). Studies with ≥50 patients/arm that reported relevant outcomes for 2L+ patients aged ≥18 years, who progressed during/after their most recent line, were eligible. Feasibility assessments evaluated heterogeneity between studies. Spearman’s correlation coefficient estimated strength of associations; weighted least squares (WLS) regression analyses quantified relationships, and surrogate threshold effects (STE) were generated. Sensitivity analyses were also performed.
RESULTS: Thirty studies (N=111 to 792) were included in the base-case analysis, with 28 assessed for HR and 16 assessed for absolute median. Twenty-seven trials were phase 3, 21 were in the 2L+ setting and 9 in 3L+, and 5 allowed cross-over from comparator to treatment arm. Significant positive correlations were found between PFS and OS (Spearman’s coefficient: 0.52 [95% CI 0.13-0.83] for HRs; 0.88 [95% CI 0.72-0.94] for absolute medians). WLS models suggested moderate predictive power for HRs (slope 0.53; p<0.001; R2=0.42) and absolute medians (slope 1.8; p<0.001; R2=0.84). STE predicted a PFS HR of 0.749 or lower would result in a significant OS treatment effect. Sensitivity analyses were consistent with the base models.
CONCLUSIONS: These results support PFS as a surrogate marker for OS in 2L+ RRMM trials, indicating that substantial PFS benefits are expected to result in OS benefits with further data maturity. Funding: GSK
METHODS: A systematic literature review identified efficacy and safety data from phase 2/3 RRMM clinical trials (2008 to February 2024). Studies with ≥50 patients/arm that reported relevant outcomes for 2L+ patients aged ≥18 years, who progressed during/after their most recent line, were eligible. Feasibility assessments evaluated heterogeneity between studies. Spearman’s correlation coefficient estimated strength of associations; weighted least squares (WLS) regression analyses quantified relationships, and surrogate threshold effects (STE) were generated. Sensitivity analyses were also performed.
RESULTS: Thirty studies (N=111 to 792) were included in the base-case analysis, with 28 assessed for HR and 16 assessed for absolute median. Twenty-seven trials were phase 3, 21 were in the 2L+ setting and 9 in 3L+, and 5 allowed cross-over from comparator to treatment arm. Significant positive correlations were found between PFS and OS (Spearman’s coefficient: 0.52 [95% CI 0.13-0.83] for HRs; 0.88 [95% CI 0.72-0.94] for absolute medians). WLS models suggested moderate predictive power for HRs (slope 0.53; p<0.001; R2=0.42) and absolute medians (slope 1.8; p<0.001; R2=0.84). STE predicted a PFS HR of 0.749 or lower would result in a significant OS treatment effect. Sensitivity analyses were consistent with the base models.
CONCLUSIONS: These results support PFS as a surrogate marker for OS in 2L+ RRMM trials, indicating that substantial PFS benefits are expected to result in OS benefits with further data maturity. Funding: GSK
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
CO30
Topic
Clinical Outcomes
Topic Subcategory
Relating Intermediate to Long-term Outcomes
Disease
SDC: Oncology