Abstract 3895: Prognostic significance of blood-based multi-cancer detection in cell-free DNA: 4-year outcomes analysis

Abstract

Abstract Background: The third substudy of the case-control Circulating Cell-free Genome Atlas (CCGA; NCT02889978) study reported an observed 99.5% specificity and 89% cancer signal origin prediction accuracy in true positives using a refined version of a blood-based multi-cancer early detection (MCED) test. A shared cancer signal has been detected across >50 cancer types. Most cancer types do not have recommended screening and are diagnosed once symptoms present. Previous analysis of participant outcomes evaluated the prognostic significance of cancer signal detection by an earlier version of the MCED test with 3-year longitudinal follow-up data. We present a longer follow-up time with an updated statistical model to robustly account for variation in cancer type and stage (ie, a cancer spectrum) when estimating survival hazards. Methods: The MCED test utilizes targeted methylation sequencing of plasma cell-free DNA (cfDNA) and machine learning classifiers to estimate a cancer signal from tumor-derived cfDNA. We compared observed overall survival of participants with known cancer who had 4-year longitudinal follow-up data to expected survival from Surveillance, Epidemiology, and End Results (SEER) data matched for age, sex, cancer type, and stage to account for changes in the cancer spectrum between “Cancer Signal Detected” (CSD) and “No Cancer Signal Detected” (NCSD) groups using a refined MCED test version. Hazard ratios (HR) and Kaplan-Meier curves were calculated for these groups. A one sample test was used to estimate the HR and confidence interval against a reference survival distribution. A methylation-based estimate of ctDNA abundance was used to quantify the cfDNA tumor methylated fraction (TMeF). Results: Follow-up data were available for 2471/2513 (98%) evaluable participants with known cancer (median 46.9 mos [IQR: 22.7 - 56.0]). Among 782 (31%) participants with known cancer that died during follow-up, the MCED test detected a cancer signal in 668 (85%). Accounting for the effect of cancer type and stage, the NCSD group had better survival than the CSD group (HR 0.585 [CI: 0.474-0.712]). Both the NCSD group and the CSD group had better survival than expected SEER-based survival (p<0.05, HR 0.432 [0.356-0.521] and p<0.05, HR 0.737 [0.684-0.791], respectively). TMeF increased with clinical stage (p<0.05) and was lower for NCSD cancers relative to CSD cancers for all stages (p<0.05). Conclusions: In the third substudy of CCGA, extended follow-up confirmed the prognostic value of CSD compared to NCSD by the refined MCED test for participants with cancer. Citation Format: Charles Swanton, Matthew Margolis, Earl Hubbell, Kathryn N. Kurtzman, Stephannie Shih, Oliver Venn, Michael Seiden. Prognostic significance of blood-based multi-cancer detection in cell-free DNA: 4-year outcomes analysis [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 3895.