Machine Learning Demonstrates High Accuracy for Disease Diagnosis and Prognosis in Plastic Surgery

Area of Science:

• Plastic Surgery
• Medical Artificial Intelligence
• Machine Learning Applications

Background:

• Machine learning (ML) offers pattern detection in large datasets for decision-making under uncertainty.
• This review examines ML's current role in plastic surgery, including clinical applications, accuracy, and future research directions.

Purpose of the Study:

• To review the applications of machine learning in plastic surgery.
• To outline the diagnostic and prognostic accuracies of ML models in clinical practice.
• To propose future directions for ML in plastic surgery research and application.

Main Methods:

• Searched EMBASE, MEDLINE, CENTRAL, and ClinicalTrials.gov (1990-2020).
• Included clinical studies reporting diagnostic and prognostic accuracies of ML models in plastic surgery.
• Collected data on clinical indication, model type, accuracies, and comparison with clinical evaluation.

Main Results:

• 51 articles were included from 1181 identified.
• ML applications included diagnosis prediction (88.80% accuracy), outcome prediction (86.11%), and pre-operative planning (80.28%).
• Neural networks were the most common models, followed by support vector machines, decision trees/random forests, and logistic regression.

Conclusions:

• Machine learning demonstrates high accuracy in diagnosing and prognosing conditions in plastic surgery, including burns, facial deformities, and cosmetic procedures.
• No studies directly compared ML performance against clinician performance.
• Future research should utilize larger datasets, data augmentation, and novel deep learning models across plastic surgery subspecialties.