Enhancing Cancer Prediction Among People With Elevated CEA: A Novel Approach Using CEA Levels and Clinical Risk Scoring.

Abstract

Carcinoembryonic antigen (CEA) is a blood biomarker with significance in cancer monitoring and prognosis prediction. Despite not being a primary diagnostic tumor marker, it is frequently measured in clinical practice and health check-ups in primary care services, often prompting extensive investigations due to its non-specific nature. This study investigated the integration of CEA levels with a clinical risk scoring system to enhance cancer prediction in individuals with elevated CEA. We retrospectively reviewed records, focusing on people who participated in general health check-up programs. Participants with elevated CEA levels (≥5 ng/mL) and no previous cancer diagnosis were selected. A multivariable logistic regression model was employed to establish the risk scoring system, which was then evaluated using Receiver Operating Characteristic (ROC) analysis. Among 17 859 participants, 310 had elevated CEA levels, with 7.40% subsequently diagnosed with cancer, primarily colorectal cancer. Combining the clinical risk scoring system with CEA levels significantly improved cancer prediction, with an Area under the ROC curve of 0.918, compared to 0.825 for CEA levels alone. CEA, alongside clinical risk prediction, shows promise in predicting cancer among individuals with high CEA levels. This integrated approach can optimize early cancer detection strategies and improve patient outcomes.