Patient-centered trial design in phase III advanced cancer trials: Control arm and endpoint selection, 2015–2024.

Abstract

715 Background: Control arm selection affects patient experience and evidence supporting oncology practice. Contemporary patterns in comparator and endpoint selection across advanced cancer trials remain incompletely described. We evaluated Phase III advanced/metastatic cancer trials registered from 2015–2024. Methods: Using the ClinicalTrials.gov API, we identified Phase III interventional advanced/metastatic cancer trials initiated from 2015–2024 with completed, terminated, or active-not-recruiting status. Control arms were classified as active comparator, placebo-containing, active+BSC/placebo, observation, BSC only, or investigator’s choice. Primary endpoints were categorized as OS, PFS/DFS, ORR, PRO, or other. Temporal trends were assessed using Spearman correlation. Multivariable logistic regression evaluated factors associated with placebo-containing controls, adjusting for year, sponsor, endpoint, and tumor type. Results: Among 935 trials, active comparator arms were most common (58.0%), followed by placebo-containing controls (22.7%), active+BSC/placebo (14.9%), observation (2.6%), BSC only (1.1%), and investigator’s choice (0.9%). Active comparator use increased from 46.8% in 2015 to 66.7% in 2023, while placebo-containing designs declined from 32.3% to 14.3%. Each additional initiation year was associated with lower odds of placebo-containing control use (OR 0.91, 95% CI 0.85–0.97; p=0.005). Industry-sponsored trials were more likely to use placebo-containing controls (OR 1.85, 95% CI 1.28–2.67; p=0.001), as were trials with PFS/DFS primary endpoints (OR 1.85, 95% CI 1.32–2.59; p<0.001). Compared with lung cancer trials, placebo-containing controls were more common in prostate (OR 2.42; p=0.008) and pancreatic cancer trials (OR 2.65; p=0.025). PROs were primary endpoints in only 9 trials (1.0%), with no increase over time. OS-primary designs were most common in hepatic, pancreatic, and gastric cancers. Conclusions: Active comparator use increased and placebo-containing controls declined in Phase III advanced cancer trials from 2015–2024, consistent with an evolving therapeutic landscape. However, placebo-containing controls remained independently associated with industry sponsorship and PFS/DFS-based primary endpoints, while endpoint selection varied markedly by tumor type. PROs remained rare primary endpoints throughout. These findings suggest that comparator and endpoint choices affecting evidentiary strength and patient-centeredness cluster within specific trial contexts, warranting attention in future oncology trial design guidance. Selected tumor types. Tumor Type N Placebo (%) OS Primary (%) Active (%) Lung 206 19.4 37.4 66.5% Breast 136 25.7 10.3 51.5% Prostate 59 33.9 28.8 49.2% Pancreatic 34 29.4 73.5 47.1% Gastric 58 27.6 63.8 43.1% Colorectal 89 12.4 22.5 64.0%