Performance of APCS Risk Score and Fecal Immunochemical Test in Predicting Colorectal Cancer: A Retrospective Diagnostic Accuracy Study.

Abstract

Objective: To evaluate the diagnostic performance of APCS and FIT for predicting colorectal cancer (CRC), and to develop an enhanced predictive model incorporating clinical and laboratory parameters.

Methods: A retrospective study of 662 patients undergoing colonoscopy at Rattanaburi Hospital, Thailand (January 2022–April 2025). Clinical data, quantitative FIT, APCS scores, and histopathology were analyzed

Results: Mean age was 61.3 years; 42.45% were male. Adenomas comprised 93.96% of lesions, and adenocarcinomas (CA) 2.87%. The subgroup with APCS ≥4 and positive FIT had the highest CA detection rate (6.29%). For predicting CA, FIT alone had an AUROC of 0.56 (sensitivity 68.4%, specificity 45.3%), while APCS ≥4 alone had an AUROC of 0.59 (sensitivity 68.4%, specificity 50.7%). Combining FIT and APCS ≥4 increased specificity to 79.0% (sensitivity 47.4%, AUROC 0.64). Multivariable regression identified FIT positivity (OR=7.66), APCS ≥4 (OR=4.54), diabetes mellitus (OR=8.84), and hematocrit (OR=0.70) as independent CA predictors (p<0.01 for all except APCS p=0.01). An enhanced model combining these variables significantly improved performance (AUROC 0.86).

Conclusion: Standalone APCS or FIT scores offer limited predictive accuracy (AUROC <0.65). However, incorporating diabetes and hematocrit into an enhanced model significantly improves diagnostic performance (AUROC 0.86), showing high potential for risk-based screening.