Abstract
PURPOSE: Earlier detection of colorectal cancer (CRC) can improve survival rates. A simple, effective blood test may help improve screening participation. The multi-target blood protein (MTBP; ColoSTAT(®)) test and algorithm uses concentrations of five protein biomarkers of CRC and patient's sex and age to generate a CRC likelihood score. We compared the performance of the MTBP test in detecting CRC to colonoscopy, the 'gold standard'. METHODS: This cross-sectional, multicenter study enrolled participants into two cohorts: participants recently colonoscopically diagnosed with CRC and progressing to surgery or neoadjuvant treatment (Cohort 1), and participants with no CRC history who were scheduled for colonoscopy (Cohort 2). Due to COVID-19 pandemic-related recruitment delays, Cohort 1 was supplemented with bio-banked blood samples (BBS) from patients with clinically confirmed CRC. Performance goals for sensitivity and specificity of the MTBP test compared to colonoscopy were ≥ 73% (lower 95% confidence limit [LCL] > 60%) and ≥ 90% (LCL > 80%), respectively. RESULTS: Cohort 1 included 29 patients, Cohort 2 enrolled 768 patients and 192 BBS were included. Median age when providing samples was 64 years (range, 40-85 years); 53.4% were female. Definitive MTBP test results were obtained from 657 samples. 112 and 389 samples met the criteria for inclusion in the primary sensitivity and specificity analyses, respectively. The sensitivity of ColoSTAT(®) for detection of all-stage CRC compared to colonoscopy was 81.3% (95%CL 73.0%-87.4%) and the specificity, 91.0% (95%CL 87.7%-93.5%). CONCLUSIONS: The MTBP test met pre-specified primary performance endpoints and warrants further evaluation in clinical populations at elevated risk of CRC.