Enhancing Pediatric Adnexal Torsion Diagnosis: Prediction Method Utilizing Machine Learning Techniques.

Area of Science:

• Pediatric Surgery
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Pediatric adnexal torsion and acute appendicitis present similar symptoms, complicating diagnosis.
• Accurate and timely diagnosis is crucial for effective treatment and preventing complications in pediatric patients.

Purpose of the Study:

• To develop and evaluate a machine learning-based diagnostic approach for differentiating pediatric adnexal torsion from acute appendicitis.
• To assess the accuracy of support vector classifiers in distinguishing these two conditions.

Main Methods:

• Retrospective analysis of 41 female pediatric patients (21 with adnexal torsion, 20 with acute appendicitis).
• Utilized clinical presentation, pain duration, white blood cell counts, and ultrasound findings as features for machine learning models.
• Employed support vector classifiers (SVMs) with various kernels to predict the diagnosis.

Main Results:

• No significant age difference between groups.
• Adnexal torsion group showed shorter pain duration, less vomiting, lower fever, reduced leukocytosis, and lower CRP elevation compared to the appendicitis group.
• SVM models achieved a predictive accuracy of 87%–97% in distinguishing adnexal torsion from appendicitis.

Conclusions:

• Machine learning, particularly SVMs, can significantly enhance diagnostic accuracy for pediatric adnexal torsion versus appendicitis.
• The proposed AI-driven diagnostic approach shows promise for clinical application.
• Further extensive validation and exploration of diverse machine learning models are recommended.