Abstract 784: Systematic review and meta-analysis of the accuracy of tumor origin detection in blood-based multi-cancer early detection (MCED) in the general population

Abstract Background: Among the recently developed blood-based multi-cancer early detection (MCED) tests, determining the location of tumors is pivotal to guiding appropriate treatment. We aim to systematically review and statistically examine the accuracy of tumor of origin prediction among blood-based MCED tests. Methods: Original articles were searched from Pubmed, Cochrane, and Embase that featured blood-based screening tests, multiple cancer types, and asymptomatic human subjects. We excluded studies with small samples (n<30), non-screening tests, and non-blood-based tests. For cell-free DNA (cfDNA) based assays, measurements of diagnostic accuracy were pooled for meta-analysis. Findings: Of 1,074 records identified and screened, 4 cohort studies were analyzed using cfDNA-based diagnostic tests. The accuracy of tissue-of-origin (TOO) prediction for 3,762 cancer samples across cancer types was 0.79 (95% CI 0.64 - 0.91). Among six cancer types, colorectal cancers had the highest accuracy, and liver & bile duct cancers had the lowest, although the difference was statistically insignificant (0.89 (95% CI 0.79-0.97) vs. 0.68 (95% CI 0.40-0.90)). Additionally, cases were most frequently misclassified as colorectal cancer (Table 1). Lastly, two studies reported increased prediction accuracies when two sites of origin were assigned rather than a single site. The information for localizing TOO was derived from methylation patterns of cfDNA in two studies, fragmentation profiles of cfDNA in another study, and protein markers in the other research. Interpretation: Our results demonstrate that the primary site of cancers was accurately discerned in 79% of cases by MCED tests. However, the performance varies across cancer types. Further research on performance according to cancer stages and in combination with other molecular profiling is warranted. Table 1. Events/Total Accuracy of prediction (95% CI) misclassification type freqeuncy Total 3,129/3,762 0.79 (0.64 - 0.91) Colorectal 638/723 0.89 (0.79 - 0.97) stomach & esophagus 26/553 breast 8/553 Breast 443/538 0.86 (0.64 - 0.99) colorectal 20/349 pancreas & GB 4/349 lung 4/349 Ovarian Breast 132/159 0.85 (0.62 - 0.99) colorectal 39/261 uterus 13/147 Pancreatic & GB 289/338 0.82 (0.69 - 0.93) colorectal 12/300 stomach & esophagus 10/300 Stomach & Esophagus 225/311 0.75 (0.48 - 0.95) colorectal 39/261 lung 10/261 Lung 530/656 0.72 (0.45 - 0.92) colorectal 29/560 head and neck 8/560 Liver & Bile duct 95/144 0.68 (0.40 - 0.90) colorectal 13/220 pancreatic & GB 8/220 Abbreviations: CI = confidence interval; GB = gallbladder Citation Format: Youjin Oh, Joo Hee Park, Liam Il-Young Chung, Richard Duan, Trie Arni Djunadi, Sung Mi Yoon, Zunairah Shah, Chan Mi Jung, Ilene Hong, Leeseul Kim, Horyun Choi, Young Kwang Chae. Systematic review and meta-analysis of the accuracy of tumor origin detection in blood-based multi-cancer early detection (MCED) in the general population [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 784.