Abstract
Timely diagnosis and intervention in colorectal cancer are critical to improving patient outcomes and limiting disease progression. Screening of average-risk individuals is essential for detecting tumors at an earlier, more treatable stage. However, adherence to current screening programs remains suboptimal. Liquid biopsies represent a promising alternative to stool-based tests and may play a key role in optimizing colorectal cancer detection and diagnostic pathways. In this study, 957 patients were recruited across various clinical sites in the United States: 48 with colorectal cancer, 157 with advanced precancerous lesions (APL), 331 with nonadvanced lesions, and 421 with a negative colonoscopy diagnosis. Blood was obtained from patients either prior to scheduled colonoscopy or before surgical resection and any anticancer therapies. Streck plasma samples were analyzed by the Dxcover Liquid Biopsy Platform and classified using machine learning algorithms. When colorectal cancer was classified against all other groups, the ROC curve generated an AUC value of 0.95, and test sensitivity and specificity were 90% and 89%, respectively. The diagnostic model accurately predicted 75% of stage I (3/4), 100% of stage II (15/15), 93% of stage III (14/15), and 100% of stage IV (6/6) colorectal cancers. For the advanced colorectal neoplasia model, 29% of APL were detected. A simple blood test with high sensitivity for early-stage colorectal cancer could significantly enhance patient outcomes. With continued development, this liquid biopsy has the potential to make a substantial impact on the early detection of colorectal cancer. PREVENTION RELEVANCE: Timely diagnosis and intervention in colorectal cancer are critical to improving patient outcomes. A simple blood test with high sensitivity for early-stage colorectal cancer could significantly enhance patient outcomes. With continued development, this liquid biopsy has the potential to make a substantial impact on the early detection of colorectal cancer.