Abstract 1265: Comparison of mortality- vs. stage-based endpoints in randomized trials of cancer screening: A systematic review with implications for multi-cancer screening trials

Abstract Background: Randomized controlled trials (RCTs) assessing the endpoint of late-stage cancer incidence, rather than the gold-standard of cancer-specific mortality, are proposed for evaluation of multi-cancer early detection tests. Methods: We conducted a systematic review to assess whether stage-based endpoints can provide a valid alternative for mortality-based endpoints (registration: PROSPERO CRD42023411132). We searched PubMed to identify RCTs of cancer screening interventions published in English before January 20, 2023. Inclusion required that RCTs report, by study arm, numbers of cancer-specific deaths and cancer cases distinguishable between stages I-II vs. III-IV. Data extraction was done by 2 separate reviewers and included the numbers of cancer cases, deaths, and participants by study arm. For each trial, we calculated the percentage reduction in the cumulative (a) late-stage cancer incidence and (b) cancer-specific mortality between the intervention and control arms. We compared these two endpoints based on Pearson correlation coefficients, fixed-effects meta-analysis, and binary statistical significance. Results: We analyzed 41 trials ranging in size from 2,500 to 409,000 participants. Correlation between the reduction in cancer-specific mortality vs. stage III-IV cancer was strong for lung (Pearson’s ρ=0.92, n=12), breast (ρ=0.70, n=6), and ‘other’ cancers (ρ=0.69, n=8) but weak for colorectal (ρ=0.39, n=11) and prostate cancers (ρ<0, n=4). The overall correlation by fixed-effects meta-analysis across cancer types was 0.71 (95%CI=0.47-0.86) (I2=0.55). In larger trials, there was a smaller deviation between the observed reduction in cancer-specific mortality and that predicted, by cancer type, based on the trial’s reduction in stage III-IV cancer (p=0.02). Of 12 trials demonstrating a statistically significant reduction in stage III-IV cancer incidence (p<0.05), 5 also showed reduced cancer-specific mortality (42%). Conclusions: In screening RCTs, reductions in late-stage cancer vs. cancer-specific mortality are correlated, but with important differences by cancer type. Screening interventions that reduce late-stage cancer incidence do not necessarily reduce cancer-specific mortality. Citation Format: Xiaoshuang Feng, Hana Zahed, Justina Onwuka, Mattias Johansson, Ruth Etzioni, Matthew E. Callister, Hilary A. Robbins. Comparison of mortality- vs. stage-based endpoints in randomized trials of cancer screening: A systematic review with implications for multi-cancer screening trials [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1265.