P118 External validation of model predicting early-onset surgery in Crohn’s Disease patients

Yao, J.(1);Hu, B.(2);Wang, H.(3);Zhi, M.(1);

(1)The Sixth Affiliated Hospital of Sun Yat-Sen University, Department of Gastroenterology- Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Disease, Guangzhou, China;(2)The Sixth Affiliated Hospital of Sun Yat-Sen University, Department of Colorectal Surgery- Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Disease, Guangzhou, China;(3)The Sixth Affiliated Hospital of Sun Yat-Sen University, Department of Gastrointestinal Surgery- Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Disease, Guangzhou, China

Background

Screening Crohn’s disease (CD) patients with high risk of early-onset surgery is crucial in launching therapeutic strategies. We have already identified disease behavior, smoking, body mass index, C-reactive protein level at diagnosis, previous perianal or intestinal surgery, maximum bowel wall thickness, use of biologics, and exclusive enteral nutrition as independent significant factors associated with 1-year surgery surgeries risk and further established a prognostic model (Fig.1,World J Gastroenterol. 2020;26(5):524-534). We aimed to validate this model using external cohort.

Methods

 This retrospective study was conducted from Jan, 1, 2017, to Dec, 31, 2019 in three tertiary referral centers including Sixth Affiliated Hospital of Sun Yat-Sen University, Second Affiliated Hospital of Zhejiang University, and Second Affiliated Hospital of Military Medical University. Data of patients with a confirmed diagnosis of CD were collected through hospital electronic system. The published model was validated with calibration using the Hosmer-Lemeshow goodness-of-fit test, and discrimination was assessed using areas under the curve (AUC).

Results

A total of 756 patients were enrolled in our study with 101 (13.4%) excluded for the sake of incomplete data and loss of follow-up. Of the enrolled patients, 74.8% were male (n = 490) at the mean age of 28.4 ± 11.0 years, with the mean follow-up period of 21.8 ± 8.1 months. An ideal predictive ability of this model was confirmed by receiver operating characteristic curves and AUC as high as 94.5%. Besides, acceptable sensitivity of 69.5% and excellent specificity of 97.0% supported further clinical promotion and application of this model.

Conclusion

This model owns ideal ability to predict 1-year surgery risk in CD patients, which definitely help clinical decision-making and acid therapeutic strategies launching.