DOES 12-MONTH PFS PREDICT BETTER FUTURE SURVIVAL? A CONDITIONAL OVERALL SURVIVAL ANALYSIS IN RELAPSED/REFRACTORY DLBCL
Author(s)
Sneha Rai, MSc Statistics1, Jatin Gupta, MBA, MPharm1, Mohd Kashif Siddiqui, MBA, MPH, PharmD2.
1EBM Health Consultants, Delhi, India, 2EBM Health, London, United Kingdom.
1EBM Health Consultants, Delhi, India, 2EBM Health, London, United Kingdom.
OBJECTIVES: In oncology, survival evidence is often presented from baseline, but clinical decisions are rarely made only at baseline. Patients with relapsed/refractory diffuse large B-cell lymphoma (R/R DLBCL) who remain progression-free at 12 months may have a different subsequent prognosis from the overall treated population at baseline. This study aimed to estimate conditional overall survival (OS) among patients with R/R DLBCL who remained progression-free at 12 months.
METHODS: Progression-free survival (PFS) and OS data were available from randomised controlled trials (RCTs) assessing pharmacological treatments for R/R DLBCL. Conditional OS was estimated among patients who remained progression-free at 12 months, defined as the probability of surviving for an additional period after achieving the 12-month PFS landmark. Conditional OS estimates were generated at clinically relevant future time horizons, including 24, 36, and 60 months after the landmark. Analyses were conducted overall and, where feasible, by treatment arm, treatment class, and comparator type.
RESULTS: A total of 6 RCTs that included rituximab-based chemotherapy as the control arm, were included in the analysis. Among patients who remained progression-free at 12 months, conditional OS at 24, 36, and 60 months was 90% (95%CI: 86.9, 93.2), 87% (95%CI: 82.7, 91.4), and 84.5% (95% CI: 78.3, 91.1), respectively. Across studies, the probability of subsequent 60-month conditional OS among 12-month progression-free survivors ranged from 7.8% to 95.6%, indicating substantial between-trial heterogeneity and uncertainty in later time horizons.
CONCLUSIONS: Among patients with R/R DLBCL who remained progression-free at 12 months, conditional OS provided a clinically interpretable estimate of subsequent prognosis. This approach complements conventional baseline survival analyses by evaluating whether early disease control is associated with durable survival. Findings should be interpreted in the context of landmark selection, reconstructed data, cross-trial heterogeneity, and limited numbers at risk at later timepoints.
METHODS: Progression-free survival (PFS) and OS data were available from randomised controlled trials (RCTs) assessing pharmacological treatments for R/R DLBCL. Conditional OS was estimated among patients who remained progression-free at 12 months, defined as the probability of surviving for an additional period after achieving the 12-month PFS landmark. Conditional OS estimates were generated at clinically relevant future time horizons, including 24, 36, and 60 months after the landmark. Analyses were conducted overall and, where feasible, by treatment arm, treatment class, and comparator type.
RESULTS: A total of 6 RCTs that included rituximab-based chemotherapy as the control arm, were included in the analysis. Among patients who remained progression-free at 12 months, conditional OS at 24, 36, and 60 months was 90% (95%CI: 86.9, 93.2), 87% (95%CI: 82.7, 91.4), and 84.5% (95% CI: 78.3, 91.1), respectively. Across studies, the probability of subsequent 60-month conditional OS among 12-month progression-free survivors ranged from 7.8% to 95.6%, indicating substantial between-trial heterogeneity and uncertainty in later time horizons.
CONCLUSIONS: Among patients with R/R DLBCL who remained progression-free at 12 months, conditional OS provided a clinically interpretable estimate of subsequent prognosis. This approach complements conventional baseline survival analyses by evaluating whether early disease control is associated with durable survival. Findings should be interpreted in the context of landmark selection, reconstructed data, cross-trial heterogeneity, and limited numbers at risk at later timepoints.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
CO26
Topic
Clinical Outcomes, Methodological & Statistical Research, Study Approaches
Topic Subcategory
Relating Intermediate to Long-term Outcomes
Disease
Oncology