Deqi Sensation to Predict Acupuncture Effect on Functional Dyspepsia: A Machine Learning Study

Area of Science:

• Integrative Medicine
• Computational Biology
• Gastroenterology

Background:

• Functional dyspepsia (FD) is a common gastrointestinal disorder with varied responses to treatment.
• Acupuncture is a potential therapy for FD, but predicting individual patient response remains a challenge.
• Understanding the factors influencing acupuncture efficacy is crucial for personalized treatment strategies.

Purpose of the Study:

• To develop and validate predictive models for acupuncture treatment outcomes in FD patients.
• To investigate the role of initial deqi sensations in predicting treatment response.
• To utilize Support Vector Machine (SVM) techniques for predicting acupuncture efficacy.

Main Methods:

• A retrospective study of 90 FD patients undergoing four weeks of acupuncture treatment.
• Support vector classification and regression models were trained using deqi sensations as features.
• Models were validated using 10-fold cross-validation and evaluated with metrics including accuracy and R ².

Main Results:

• The SVM models successfully predicted acupuncture response with 0.84 accuracy.
• The models achieved an R ² of 0.16 in predicting symptom improvement.
• Presence/absence and duration of deqi sensation, distention, and pain were significant predictors.

Conclusions:

• SVM algorithms effectively predict acupuncture response and symptom improvement in FD patients based on deqi sensations.
• These predictive models can potentially enhance acupuncture's clinical efficacy for FD.
• Optimizing treatment allocation and reducing healthcare costs are anticipated benefits.