Prospective study of a multi-modal blood-based test for colorectal cancer screening.

Abstract

70 Background: Colorectal cancer (CRC) screening has proven to be a useful tool for the detection and prevention of CRC. However, adherence is low despite the multiple screening options available. A blood-based CRC screening test with optimized sensitivity and specificity may improve screening adherence. We aimed to prospectively evaluate the performance of a multimodal blood-based colorectal neoplasia screening test in individuals presenting for colonoscopy. Methods: This prospective, observational study was developed in four hospitals in Spain and enrolled individuals (45 – 84 years of age) undergoing colonoscopy, including diagnostic and screening colonoscopies, between March 2020 and September 2021. Eligible individuals consented to use of medical records for research and to provide a blood sample prior to colonoscopy procedure. Individuals were followed for one year. Whole blood was collected and shipped ambient to a central laboratory for analysis (Shield, Guardant Health, Redwood City, CA, USA). The blood-based colorectal neoplasia test is a 500kb next-generation sequencing based panel and bioinformatic platform that incorporates cell-free DNA (cfDNA) methylation-based partitioning to identify cancer related genomic alterations and epigenomic modifications (methylation and modifications in chromatin state). Results are integrated to yield a binary “positive” or “negative” result. Here we present findings correlating the blood-based test results with colonoscopy findings and findings from available 1-year follow-up data. Results: 556 out of 598 eligible individuals had an evaluable colonoscopy and a blood-based test result that passed quality control standards. Of the 556 individuals, 52% were female; median age was 55 years (21% age 45 – 49, 74% age 50 – 74, 4% age 75+). Reasons for colonoscopy were symptoms (49%), average risk screening (33%), positive stool-based test (6%), positive family history (11%) or other reasons (1%). Colorectal adenocarcinoma prevalence rate was 1.4%. Sensitivity for CRC detection was 100% (8/8; Stage I, 1; Stage II, 3; Stage III, 2; Stage IV, 2). Sensitivity for advanced adenoma (AA) detection was 22%. Specificity for advanced neoplasia (CRC + AA) was 91%. Negative predictive value (NPV) for CRC detection was 100% and positive predictive value (PPV) was 13%. 1-year follow-up data was available for 504 individuals, of whom 9 patients developed cancer (none CRC), of whom 1 had an initial positive Shield result. Conclusions: In this prospective study of individuals eligible for colonoscopy, sensitivity and specificity of the blood-based test showed a performance consistent with the available stool-based non-invasive screening options. This, combined with a more acceptable mode of testing suggests that this blood-based test may be a viable CRC screening option.