Deep-Learning, Radiomics and Clinic Based Fusion Models for Predicting Response to Infliximab in Crohn's Disease Patients: A Multicentre, Retrospective Study
Weimin Cai,Xiao Wu,Kun Guo,Yongxian Chen,Yubo Shi,Xinran Lin
DOI: https://doi.org/10.2147/jir.s484485
IF: 4.5
2024-10-25
Journal of Inflammation Research
Abstract:Weimin Cai, 1, &ast Xiao Wu, 1, &ast Kun Guo, 2 Yongxian Chen, 3 Yubo Shi, 4 Xinran Lin 1 1 Department of Gastroenterology and Hepatology, the First Affiliated Hospital of Wenzhou Medical University, Wenzhou, 325000, People's Republic of China; 2 Department of Cardiology, the First Affiliated Hospital of Wenzhou Medical University, Wenzhou, 325000, People's Republic of China; 3 Department of Chest Cancer, Xiamen Second People's Hospital, Xiamen, 36100, People's Republic of China; 4 Department of Pulmonary, Yueqing People's Hospital, Wenzhou, 325000, People's Republic of China &astThese authors contributed equally to this work Correspondence: Xinran Lin, Department of Gastroenterology and Hepatology, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, 325000, People's Republic of China, Tel +86 18857838243, Fax +86 0576 87755312, Email Background: Accurate prediction of treatment response in Crohn's disease (CD) patients undergoing infliximab (IFX) therapy is essential for clinical decision-making. Our goal was to compare the performance of the clinical characteristics, radiomics and deep learning model from computed tomography enterography (CTE) for identifying individuals at high risk of IFX treatment failure. Methods: This retrospective study enrolled 263 CD patients from three medical centers between 2017 and 2023 patients received CTE examinations within 1 month before IFX commencement. A training cohort was recruited from center 1 (n=166), while test cohort from centers 2 and 3 (n=97). The deep learning model and radiomics were constructed based on CTE images of lesion. The clinical model was developed using clinical characteristics. Two fusion methods were used to create fusion model: the feature-based early fusion model and the decision-based late fusion model. The performances of the predictive models were evaluated. Results: The early fusion model achieved the highest area under characteristics curve (AUC) (0.85– 0.91) among all patient cohorts, significantly outperforming deep learning model (AUC=0.72– 0.82, p=0.06– 0.03, Delong test) and radiomics model (AUC=0.72– 0.78, p=0.06– 0.01). Compared to early fusion model, the AUC values for the clinical and late fusion models were 0.71– 0.91 (p=0.01– 0.41), and 0.81– 0.88 (p=0.49– 0.37) in the test and training set, respectively. Moreover, the early fusion had the lowest value of Brier's score 0.15– 0.12 in all patient set. Conclusion: The early fusion model, which integrates deep learning, radiomics, and clinical data, can be utilized to predict the response to IFX treatment in CD patients and illustrated clinical decision-making utility. Keywords: deep-learning, radiomics, Crohn's disease, infliximab, prediction Crohn's disease (CD), a main form of inflammatory bowel disease (IBD), is characterized as a chronic, remitting, and relapsing inflammatory disorder of the gastrointestinal tract, requiring ongoing management and surveillance. 1 Infliximab (IFX), shows efficacy in inducing and maintaining remission, also is the first anti-tumor necrosis factor (TNF) α agent. Yet, primary loss of response (PLR) to IFX is observed in 13–40% of CD patients, while 23–46% patients experience a secondary loss of response (SLR). 2,3 The reduction of response to IFX, an anti-TNF monoclonal antibody, is believed to result from inadequate drug levels, attributed to the antibody's degradation and clearance. Consequently, when a decrease in the IFX's effectiveness is suspected, a comprehensive assessment involving the evaluation of active inflammation, potential IBD complications, clinical symptoms, biochemical markers, endoscopic findings, and imaging should be undertaken. Furthermore, consistent monitoring of disease activity during IFX treatment is crucial, enabling timely adjustments to therapeutic strategies to achieve and maintain long-term remission. For decades, the Crohn's Disease Activity Index (CDAI) has been regarded as a valuable tool for measuring the clinical activity of CD. 4–6 Nevertheless, the accurate calculation of CDAI requires over a seven-day period and the necessity for patient compliance limits its utility in daily clinical practice. 7,8 While colonoscopy offers an objective and accurate assessment of the intestinal condition, its invasiveness, high costs, and potential risks, including perforation, bleeding, and cardiovascular complications related to sedation, detract from its desirability. 9,10 Furthermore, a considerable number of CD patients are unsuitable candidates for endoscopy due to intestinal luminal strictures or a penetrating disease phenotyp -Abstract Truncated-
immunology