Treatment Algorithm for the Resorption of Calcific Tendinitis Using Extracorporeal Shockwave Therapy: A Data Mining...

Area of Science:

• Orthopedics
• Medical Informatics
• Biomedical Engineering

Background:

• Extracorporeal shockwave therapy (ESWT) shows promise for calcifying shoulder tendinitis but has variable outcomes.
• Data mining offers potential for predicting treatment success in healthcare settings.

Purpose of the Study:

• To identify the optimal data mining technique for predicting ESWT-induced shoulder calcification resorption.
• To determine the most accurate algorithm for clinical application in predicting treatment outcomes.

Main Methods:

• A case-control study involving 248 patients with calcified shoulder tendinitis treated with ESWT.
• Seven clinical factors were analyzed: sex, age, affected side, symptom duration, Constant-Murley score, calcification size, and type.
• Five data mining techniques were evaluated: multilayer perceptron, naïve Bayes, sequential minimal optimization, logistic regression, and J48 decision tree.

Main Results:

• The J48 decision tree achieved 89.5% accuracy using 10-fold cross-validation.
• Shorter symptom duration (≤10 months) and smaller calcification size (≤10.82 mm) were the strongest positive predictors of resorption.
• Symptom duration, calcification size, and type were the most influential input attributes.

Conclusions:

• The J48 decision tree method provides high precision and accuracy for predicting shoulder calcification resorption following ESWT.
• Clinical factors like symptom duration and calcification size can guide treatment expectations and patient selection for ESWT.