Integrating APCS Score With Dual Biomarker Testing ( FIT + mSEPT9 ) for Precision Risk Stratification and Predictive Modeling in High‐Risk Populations for Colorectal Cancer

ABSTRACT Objective To evaluate the diagnostic performance of integrating the Asia‐Pacific Colorectal Screening (APCS) score with dual biomarker testing comprising the fecal immunochemical test (FIT) and blood‐based methylated Septin9 (mSEPT9) for risk stratification and detection of advanced colorectal neoplasia (ACN) in a high‐risk population, using colonoscopy with pathological examination as the reference standard. Methods A retrospective analysis included 115 patients who underwent FIT, plasma mSEPT9 testing, and colonoscopy between January 2024 and December 2025. FIT was analyzed quantitatively for ROC analysis, while the 20 μg/g threshold was used for binary operating characteristics. Logistic models included ordinal APCS tier, ln(FIT + 1), and mSEPT9 status. The triple‐model threshold was selected by the Youden index, and 1000 bootstrap resamples were used for internal validation. Results Among 115 participants, 45 had ACN, including 20 CRCs and 25 advanced adenomas. Quantitative FIT yielded an AUC of 0.836 (95% CI, 0.762–0.910). Apparent AUCs were 0.873 for APCS + FIT and 0.912 for APCS + FIT + mSEPT9 (ΔAUC = 0.039; p = 0.047). At a cutoff of 0.318, the triple model showed 91.1% sensitivity, 68.6% specificity, 65.1% PPV, and 92.3% NPV. The optimism‐corrected AUC was 0.886 for the triple model and 0.852 for APCS + FIT; corrected ΔAUC was 0.034 (95% CI, −0.007 to 0.078). Conclusion The integrated model showed high apparent discrimination and sensitivity for ACN, but the incremental benefit over APCS + FIT was uncertain after internal validation. Prospective external validation is required before clinical implementation.