Correlation between tumor size change and outcome in a rare cancer immunotherapy basket trial

Background: RECIST criteria for progressive disease, partial response, and complete response, reflecting þ20%, −30%, and −100% tumor size changes, respectively, are critical outcome variables in oncology clinical trials. Herein, we evaluated post-immunotherapy tumor size change correlation with outcomes. Methods: We used a unique clinical trial data resource, a multicenter basket trial in patients with rare solid tumors treated with nivolumab (anti-PD-1) and ipilimumab (anti-CTLA-4) between 2017 and 2023 (National Cancer Institute/Southwest Oncology Group-sponsored DART trial [NCT02834013]) (open at 1083 sites at its peak). Outcome associations were evaluated by survival analysis techniques including Martingale residuals. Results: In 638 evaluable patients, we found strong linear relationships between percent change in tumor measurement up to a 40%-50% increase and progression-free (PFS) and overall survival (OS) (both Cox regression P<.001; landmark analyses based on day 65). Pearson R correlation between survival estimates and tumor change category were −0.94, −0.89, and −0.89 (PFS) and −0.84, −0.90, and −0.90 (OS) for median, 6-month (PFS), and 1-year (OS) and for 1-year (PFS) and 2-year (OS) estimates. Conclusions: Percent change in tumor measurement per RECISTv1.1 (the sum of longest dimensions of target lesions) has a linear association with PFS and OS up to a 40% to 50% increase in tumor measurement in this cohort of patients with rare cancers who received combination immune checkpoint blockade. Quantitative first scan tumor measurement changes include important information to evaluate the potential efficacy of a therapy beyond the proportion of patients who achieve an objective response.