Yuan Ze,1 Hui-Ming Tu,2 Yuan-Yuan Zhao,1 Lin Zhang3,4 

1Tumor Research and Therapy Center, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan, 250021, People’s Republic of China; 2Department of Gastroenterology, Affiliated Hospital of Jiangnan University, Wuxi, 214122, People’s Republic of China; 3Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, Hefei, 230026, People’s Republic of China; 4School of Population Medicine and Public Health, Peking Union Medical College/Chinese Academy of Medical Sciences, Beijing, 100053, People’s Republic of China

Correspondence: Hui-Ming Tu, Department of Gastroenterology, Affiliated Hospital of Jiangnan University, No. 1000, Hefeng Road, Binhu District, Wuxi, 214122, People’s Republic of China, Tel +86-13861753621, Email tuhuiming1891@163.com Yuan-Yuan Zhao, Tumor Research and Therapy Center, Shandong Provincial Hospital Affiliated to Shandong First Medical University, No. 324, Jingwu Weiqi Road, Huaiyin District, Jinan, 250021, People’s Republic of China, Tel +86-13589097201, Email zhaoyuanyuan21@outlook.com

Purpose: This study aimed to develop and validate a nomogram for predicting positive colonoscopy results using the data from non-invasive screening strategies.
Methods: The volunteers participated in primary colorectal cancer (CRC) screenings using Asia-Pacific colorectal screening (APCS) scoring, faecal immunochemical testing (FIT) and stool deoxyribonucleic acid (sDNA) testing and underwent a colonoscopy. The positive colonoscopy results included CRC, advanced adenoma (AA), high-grade intraepithelial neoplasia (HGIN), and low-grade intraepithelial neoplasia (LGIN). The enrolled participants were randomly selected for training and validation sets in a 7:3 ratio. A model for predicting positive colonoscopy results was virtualized by the nomogram using logistic regression analysis.
Results: Among the 179 enrolled participants, 125 were assigned to training set, while 54 were assigned to validation set. After multivariable logistic regression was done, APCS score, FIT result, and sDNA result were all identified as the predictors for positive colonoscopy results. A model that incorporated the above independent predictors was developed and presented as a nomogram. The C-index of the nomogram in the validation set was 0.768 (95% CI, 0.644– 0.891). The calibration curve demonstrated a good agreement between prediction and observation. The decision curve analysis (DCA) curve showed that the model achieved a net benefit across all threshold probabilities. The AUC of the prediction model for predicting positive colonoscopy results was much higher than that of the FIT + sDNA test scheme.
Conclusion: The nomogram for predicting positive colonoscopy results was successfully developed based on 3 non-invasive screening tools (APCS scoring, FIT and sDNA test).

Keywords: nomogram, colorectal cancer, primary screening, faecal immunochemical testing, stool deoxyribonucleic acid