Abstract 7202: Real-world data and clinical experience from over 100,000 multi-cancer early detection tests

Abstract Background: A multi-cancer early detection (MCED) test, which has been commercially available since April 19, 2021, detects a cancer signal and predicts cancer signal origin (CSO) by analyzing methylation patterns in cell-free DNA. Here, we report real-world clinical experience with >100, 000 MCED tests ordered and processed through October 14, 2023. Methods: A Quality Assurance program was set up to characterize use of MCED testing in a real-world setting and included voluntarily reported outcome data from healthcare providers to assess cancer signal detection rate (CSDR; cancer signal detected results [true + false positives] divided by all tests) and CSO accuracy. This real-world data analysis included 111, 080 tests returned to individuals ≥22 years (56% male, median age 58 y [IQR 50-67]). All data were de-identified following the HIPAA-compliant, independent expert determination method. Only the first test administered to each individual was used in this analysis. CSDR was compared to an independent model of CSDR based on estimated cancer incidence from SEER. When available, the proportion of reported diagnosed cancers is broken down by asymptomatic patients (ie, without clinical suspicion of cancer) and stage. CSO prediction accuracy was calculated for reported diagnosed cancers with a known cancer type. Results: The overall real-world CSDR was 0.91% (1, 011/111, 080) and increased with age (p<1e-10), consistent with the independent CSDR model prediction and previously published data. Among asymptomatic patients with a positive MCED test and a reported cancer status, 49.4% (128/259) received a cancer diagnosis with a median time to diagnosis of 43 days (IQR: 21-94). Excluding recurrences and cases with multiple diagnoses, cancer stage was available for 69 asymptomatic patients, with 59% [41/69] diagnosed with stages I-III. CSO prediction accuracy across the 19 different CSO categories in this dataset was 89% (201/227) and represented 30 distinct AJCC-defined cancer types ultimately diagnosed. Many cancer diagnoses (60.8% [138/227]) were types without recommended screening guidelines, including 18 cases of head and neck, 15 pancreas and gall bladder, and 9 stomach and esophagus cancers. Conclusions: Real-world CSDR was consistent with previously reported large-scale clinical studies and an independent CSDR model. Although outcomes data is limited by incomplete reporting, a known issue in clinical surveillance data, these data provide valuable insight into real world use of MCED tests. The MCED test reliably detected a cancer signal across a wide range of cancer types, including unscreened cancers. High CSO prediction accuracy guided diagnostic workups, resulting in typically rapid cancer diagnoses. These real-world data demonstrate the successful detection of both conventionally screened and unscreened cancers using the MCED test. Citation Format: Marc Matrana, Vershalee Shukla, Dallas Kingsbury, Martin Poliak, Jordan Lipton, Louis B. Malinow, Oliver Venn, John Beausang, Geoff Stanley, Earl Hubbell, Kathryn N. Kurtzman, Jeffrey M. Venstrom, Rita Shaknovich, Candace Westgate. Real-world data and clinical experience from over 100, 000 multi-cancer early detection tests [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7202.