Journal DOI 10.68956/journal.human-performance-futures

Creative Commons source reading · Prediction and forecasting

Artificial intelligence in ACL injury prediction and prevention: a systematic review

Molavi MR; Mottaghitalab M; Samaei P; Zarei H

Originally published in J Orthop Surg Res · 2026-04-07

From the article

Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.

Abstract

AbstractBackgroundAnterior cruciate ligament (ACL) injuries are prevalent in sports, with significant physical, economic, and long-term health impacts. Artificial intelligence (AI) offers promising solutions for predicting and preventing ACL injuries through advanced data analysis. This systematic review evaluates AI applications in ACL injury prediction and prevention, focusing on techniques, performance metrics, sports contexts, and intervention effectiveness.MethodsFollowing PRISMA 2020 guidelines, we searched PubMed, Scopus, Google Scholar, and Web of Science from inception to November 1, 2025 for peer-reviewed studies in English using AI for ACL injury prediction or prevention. Studies were excluded if they focused on other injuries or were non-original research. Two reviewers independently screened articles, extracted data…

Results

A total of 80 records were identified through database searches. After removing 20 duplicates, 60 records were screened by title and abstract, of which 35 were excluded for not meeting the inclusion criteria. The remaining 25 full-text articles were assessed for eligibility, and 18 were excluded due to non-English language, lack of focus on ACL or AI, or insufficient data. Seven studies were included in the final review. The study selection process is presented in the PRISMA flow diagram (Fig. 1).

Results of individual studies

All seven studies utilized AI techniques to predict or prevent ACL injuries, with applications focusing on either risk prediction or biomechanical assessment for prevention. Four studies [36–38, 40] primarily aimed to predict ACL injury risk, employing ML models such as Linear SVM (accuracy: 96%, F1-score: 95%) [36], Fine Gaussian SVM (accuracy: 79.5%, AUC: 0.85) [37], random forest regressor (MAE: 0.81°–8.73°) [38], and random forest with logistic regression (AUC-ROC: 0.63) [40]. Two studies [39, 42] focused on biomechanical assessments to inform prevention strategies, using Sparrow Search Algorithm-optimized ELM/LSTM (R2: 0.9947–0.9992) [39] and MediaPipe Pose (good concurrent validity with VICON) [42], respectively. One study [41] used a deep learning model (Convolutional Neural Network) to predict knee joint moments for risk…

Editorial reading guide

  • What data created and tested the model, and was evaluation independent?
  • What is the cost of a false positive, false negative or out-of-distribution decision?