Abstract 1039: A novel serum lipidomics signature validated in a prospective multi-center trial enables accurate and low-cost CRC screening

Abstract

Abstract Background: Blood-based colorectal cancer (CRC) testing is expected to increase compliance to recommended screening guidelines. However, existing methods based on cfDNA typically suffer from low performance in detecting early-stage cancer and pre-malignant advanced adenomas (AA); and from high-cost, making routine screening financially impractical. Attempts to identify other blood biomarkers suffered from low reproducibility across studies, due to confounding factors related to utilizing samples from diagnosed and potentially treated patients, and artifacts related to differential handling of cases and control samples. Method: Serum samples were obtained from the Israeli Multi-OMICS screening study (IMOSS-500K), in which ~500,000 serum samples were collected as part of standard clinical routine by Maccabi Health between July 2021 and January 2023. De-identified electronic health records were used to identify 152 primary CRC patient, diagnosed within 1 year after serum collection; and another 270 age- and gender-matched controls. The cohort was divided to train and test sub-cohorts, comprising of patients diagnosed within 2-12 and 0-2 months after sample collection, respectively. A 2nd test cohort was obtained from Rambam Hospital, included 97 samples from treatment-native CRC patients, and 24 age- and gender-matched controls. A 3rd test cohort, obtained from IMOSS-500K, included 139 patients that were diagnosed with AA within 2 months from serum collection; and 139 age- and gender-matched controls. All samples were analyzed with a proprietary high-throughput LC-MS based multi-OMICS platform detecting 10,000’s of metabolites, lipids, and protein biomarker ions per sample. A diagnosis score was defined via a logistic regression model in the train sub-cohort and the diagnostic performance for was evaluated in the 3 test cohorts. Results: We identified a novel signature for CRC based on the concentration of 4 structurally related serum lipids, whose structure was not previously reported in literature. Evaluating the performance of the score in the test sub-cohort, showed a remarkable AUC of 0.96 for the detection of CRC, with a high sensitivity of 89% at 90% specificity; with 90% sensitivity achieved already at stage I. The high diagnostic performance for detecting CRC was reproducible in the 2nd test cohort, with an AUC of 0.97, and sensitivity of 94% at 90% specificity. Analyzing the 3nd test cohort, our signature achieved a sensitivity of 37% at 90% specificity for the detection of AA. Conclusion: We developed a novel low-cost blood-lipidomics based method for CRC detection, suitable for routine screening. Comprehensive validation through a multi-center prospective study showed similar performance to that published for the stool-cfDNA test ColoGuard, and higher performance than other published blood-cfDNA methods in the detection of early-stage CRC and AA. Citation Format: Boris Sarvin, Avi Shoshan, Eldad Kepten, Mark Vernik, Ilya Podolsky, Carmel Shor, Shira Shaham-Niv, Tomer Shlomi. A novel serum lipidomics signature validated in a prospective multi-center trial enables accurate and low-cost CRC screening [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 1039.