Serological Biomarker-Based Machine Learning Models for Predicting the Relapse of Ulcerative Colitis

Area of Science:

• Gastroenterology
• Biomedical Informatics
• Computational Biology

Background:

• Ulcerative colitis (UC) relapse poses a significant challenge in patient management.
• Predicting UC relapse is crucial for timely intervention and personalized treatment strategies.

Purpose of the Study:

• To evaluate the efficacy of machine learning models utilizing serological markers for predicting Ulcerative colitis relapse.
• To compare the performance of different machine learning algorithms in UC relapse prediction.

Main Methods:

• A cohort study involving 292 UC patients, collecting serological markers at hospital discharge.
• Development and comparison of four machine learning models: random forest (RF), logistic regression, decision tree, and neural network.
• Performance evaluation using accuracy, sensitivity, specificity, and AUC.

Main Results:

• The random forest model demonstrated the highest predictive performance with an AUC of 0.889 in the test group.
• The best-performing model achieved 76.4% accuracy, 78.5% sensitivity, and 76.4% specificity.
• Key predictors identified include age, hemoglobin concentration, white blood cell count, and platelet distribution width.

Conclusions:

• Machine learning models incorporating serological markers provide a highly accurate, noninvasive method for predicting UC relapse.
• These models can aid in optimizing patient outcomes and guiding personalized therapeutic decisions.