The Asia-Pacific Colorectal Screening (APCS) score and a modified version that incorporates body mass index (BMI) has been widely recognised. This study aimed to compare the external validity of the APCS and the modified APCS to predict advanced colorectal neoplasia (ACN) in an asymptomatic Chinese population.Methods
A prospective, multi-centre colonoscopy study was conducted in 7 endoscopic units across North Shanghai, China from 2016 to 2017. Average-risk subjects were recruited during their attendance for screening colonoscopy. Information on the predictors, including age, gender, family history of colorectal cancer and smoking status were collected by self-reported questionnaires. Body weight and height were measured by trained professionals. All screening colonoscopy procedures were performed by senior colonoscopists. The histology of all resected lesions was reported using the same diagnostic criteria by pathologists who were blinded to the study design. Cumulative scores were derived for each subject and categorised as average, moderate and high risk. The risk of ACN was assessed for each risk tier and the excess risk in moderate and high-risk tiers was compared with subjects with average risk. The C-statistics was employed to evaluate the discriminatory capability of both APCS scores.Results
Among a total of 2813 enrolled subjects (mean age 58.7±8.4 years), 679 cases (24.1%) were found to have colorectal neoplasia, of which 142 (5.0%) had ACN and 29 patients (1.0%) had invasive cancers (table 1). In both APCS scoring systems, the risk of ACN was significantly different in low, moderate and high-risk tier (p<0.001, table 2). When compared with average risk subjects, those in the high-risk tier had a 3.06 (95% confidence interval (CI) 1.22, 7.68) and 4.07 (95% CI 1.48, 11.20) fold increased risk of ACN for the APCS and modified APCS system, respectively. The C-statistics of APCS and modified APCS were 0.645 (95% CI, 0.598, 0.683), and 0.648 (95% CI, 0.640, 0.696), respectively.Conclusions
Both the APCS and modified APCS scores were similarly efficient to identify those subjects with high risk of can. Their feasibility and acceptability in different population groups should be further explored.