General Information
Conflict of Interest
Request Clinical Trials
Associated Trial(s): What type of data are you looking for?: Individual Participant-Level Data, which includes Full CSR and all supporting documentationRequest Clinical Trials
Data Request Status
Status: Approved Pending DUA SignatureResearch Proposal
Project Title: Analysis of Risk Factors for Loss of Response to Biologic Therapy in Patients with Inflammatory Bowel Disease and Establishment of a Predictive Model
Scientific Abstract:
Background: Biologics, as a crucial treatment for IBD, can have adverse effects on patients' health and quality of life if loss of response occurs. Machine learning has been applied to various clinical models, but there is currently a lack of research on factors influencing loss of response to ustekinumab in ulcerative colitis and predictive models based on common clinical indicators.
Objective: This study aims to retrospectively analyze risk factors associated with loss of response to biologics in IBD patients and to construct a predictive model using machine learning.
Study Design：Patients' General data, disease data, laboratory examinations, imaging examinations are the primary predictors. This study will analyze the influencing factors of the response to ustekinumab in the treatment of ulcerative colitis and construct a predictive model for the response.
Participants：Regardless of whether they have undergone induction studies, patients who are treated with ustekinumab for ulcerative colitis will be included.
Primary and Secondary Outcome Measure：endoscopic remission ,endoscopic improvement, histological remission, histological improvement, clinical remission,rectal bleeding, bowel movement frequency, Mayo score, serum and fecal inflammatory markers and colectomy.
Statistical Analysis：Binary logistic regression analysis will be used to identify independent risk factors for loss of response to biologic therapy in IBD patients.The identified independent risk factors will be used to construct predictive models using various machine learning algorithms.
Brief Project Background and Statement of Project Significance:
Inflammatory bowel disease (IBD) is a chronic, nonspecific inflammatory condition of the gastrointestinal tract with an unknown cause, characterized by recurrent episodes. It includes ulcerative colitis (UC) and Crohn's disease (CD). An epidemiological study based on nine countries in the Asia-Pacific region revealed that China has the highest incidence rate of IBD (3.44 cases per 100,000 people) [1]. Over the past two decades, the number of diagnosed cases in China has shown a rapid upward trend . IBD is a prolonged condition prone to recurrent flare-ups, imposing both disease and economic burdens on patients. The use of biologics can effectively achieve clinical response and remission in moderate to severe CD and UC, while reducing hospitalization and surgery rates among IBD patients .
Despite their high cost, biologics remain the treatment of choice for most patients with moderate to severe IBD due to their efficacy. However, if loss of response occurs during biologic therapy, subsequent treatment options are limited, including switching to other biologics, combining with immunosuppressants, or switching to JAK inhibitors. Therefore, failure in biologic therapy for IBD can impose significant burdens on patients in terms of disease progression, financial costs, and quality of life.
This study aims to retrospectively analyze the risk factors associated with loss of response to biologics in IBD patients and to construct a predictive model for loss of response using machine learning. The goal is to provide a reliable method for selecting biologics in clinical practice for IBD patients. By identifying high-risk patients for loss of response through the predictive model, this study seeks to offer scientific references for clinical decision-making by healthcare professionals, ultimately improving patients' quality of life and maintaining disease remission.
Specific Aims of the Project:
(1) Conduct statistical analysis based on case data and literature review to screen and summarize risk factors for loss of response to biologic therapy in IBD patients.
(2) Based on the summarized risk factors, construct a predictive model for loss of response to biologic therapy in IBD patients using machine learning algorithms.
(3) Provide a reliable predictive tool for loss of response to biologic therapy in IBD patients, and offer individualized treatment plans for high-risk patients identified by the predictive model. These plans may include combining with immunosuppressants, switching to JAK inhibitors, establishing follow-up systems, and close monitoring. Ultimately, this will assist and provide references for clinical decision-making by healthcare professionals, while also offering practical evidence for related research.
Study Design:
Other
Explain:
Using machine learning, a predictive model for loss of response to biologic therapy in UC will be constructed based on independent risk factors.
What is the purpose of the analysis being proposed? Please select all that apply.: Research on clinical prediction or risk prediction
Software Used: R
Data Source and Inclusion/Exclusion Criteria to be used to define the patient sample for your study:
Inclusion Criteria:
1.Patients receiving regular treatment with UST for UC.
2.Patients with complete clinical data.
Exclusion Criteria:
1.Patients with incomplete clinical data.
2.Patients who discontinued UST treatment for UC due to various reasons.
In summary,regardless of whether they have undergone induction studies, patients who are treated with ustekinumab for ulcerative colitis will be included.
Primary and Secondary Outcome Measure(s) and how they will be categorized/defined for your study: The primary outcomes are clinical remission during the active phase, normalize serum or fecal inflammatory markers and endoscopic mucosal healing. Histological remission and histological improvement are secondary outcomes.Clinical remission is defined as the resolution of rectal bleeding and normalization of bowel movement frequency, or a modified Mayo score of less than 2 points with no individual subscore greater than 1. Normalization of inflammatory markers refers to the normalization of CRP or a reduction in fecal calprotectin (FC) to an acceptable range (100--250 μg/g). Mucosal healing is defined as a Mayo endoscopic subscore of 0. Histological remission and histological improvement are defined as Geboes highest grade＜2.0 and ＜3.2. Endoscopic improvement is defined as mayo endoscopic score＜2.
Main Predictor/Independent Variable and how it will be categorized/defined for your study:
This study Gender, Age at Onset, Age at Diagnosis, Disease Duration, BMI, Smoking, Lesion Location, Intestinal Endoscopy, Intestinal Imaging, Complications, Previous Biologic Use, Previous Surgical History, IBD Medication, Erythrocyte sedimentation Rate(ESR), C-reactive protein（CRP）, Hemoglobin, White Blood Cells, Neutrophils, Lymphocytes, Monocytes, Platelets, Mean Platelet Volume, Albumin, Prealbumin, Alkaline Phosphatase, Bilirubin, Alanine aminotransferase(ALT)，Aspartate aminotransferase（ AST）, Total Cholesterol (TC), Triglycerides (TG), Serum Creatinine (Scr), Estimated Glomerular Filtration Rate (eGFR), Blood Urea Nitrogen, Cystatin C, Fecal Calprotectin.Normally distributed continuous variables were expressed as mean +/- standard deviation and analyzed for differences using independent samples t-tests. Skewed continuous variables were expressed as median or interquartile range and analyzed for differences using rank-sum tests. Categorical variables were expressed as frequencies and percentages and analyzed for differences using chi-square tests.
Other Variables of Interest that will be used in your analysis and how they will be categorized/defined for your study: This section is currently not planned.
Statistical Analysis Plan: Data will be recorded and categorized using Microsoft Excel 2021. Data processing and calculations will be performed using IBM SPSS Statistics. For datasets with missing values exceeding 20%, the deletion method will be applied. For datasets with missing values less than 20%, imputation or substitution methods will be used. Continuous variables following a normal distribution will be expressed as mean +/- standard deviation and analyzed using independent samples t-tests. Continuous variables with a skewed distribution will be expressed as median or interquartile range and analyzed using rank-sum tests. Categorical variables will be expressed as frequency and percentage and analyzed using chi-square tests. Binary logistic regression analysis will be used to identify independent risk factors for loss of response to biologic therapy in IBD patients.In this study, patients will be randomly divided into a modeling set and a validation set at a ratio of 7:3. Commonly used supervised learning algorithms will be utilized to construct the prediction models, such as Decision Tree (DT), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN), Logistic Regression (LR), and Support Vector Machine (SVM). Finally, the model performance will be evaluated using metrics such as precision, accuracy, recall, F1 score, and ROC curve, and the optimal model will be selected. A web-based or application-based visual assessment tool will ultimately be developed.
Narrative Summary:
Biologics, as a crucial treatment for IBD, can have adverse effects on patients' health and quality of life if loss of response occurs. Machine learning has been applied to various clinical models, but there is currently a lack of research on factors influencing loss of response to ustekinumab in ulcerative colitis and predictive models based on common clinical indicators.This study aims to retrospectively analyze risk factors associated with loss of response to biologics in IBD patients and to construct a predictive model using machine learning.
Project Timeline:
2025.3-202-6 Collect patient data and organize materials,obtain all data required for subsequent experiments.
2025.7-2025.9 Preprocess the data and perform univariate and multivariate analyses,ummarize independent risk factors for loss of response to biologic therapy in IBD patients.
2025.10-2025.12 onstruct predictive models using different algorithms and evaluate their performance,obtain the optimal predictive model.
2026.1-2026.6 Organize the experimental process and write papers
Dissemination Plan: The outputs of this project will serve as my master's thesis and will be submitted to Chongqing Medical University, as well as displayed on academic platforms such as CNKI (China National Knowledge Infrastructure). Additionally, a manuscript will be submitted for publication in a peer-reviewed journal. Acknowledgments to the YODA project will be included in all products derived from this research.
Bibliography: