ACL Injury Risk Screening Based on the Landing Error Scoring System: Smartphone-Based Classification and Comparison Across Vertical Drop Jump Tasks

This study focuses on screening for Anterior Cruciate Ligament (ACL) injury risk based on the Landing Error Scoring System (LESS). On the one hand, automatic high/low risk classification based on smartphones is developed, on the other hand, the impact on the screening outcome of gender and of performing one of two possible Vertical Drop Jump (VDJ) tasks-double support (DS-VDJ) and single support (SS- VDJ) -is assessed. For both aims, 102 participants were analyzed: 60 performing both jumps three times, to compare the jump techniques, and 42, performing only SS-VDJs, to develop the classifier. All jumps were evaluated through the traditional approach, where expert operators manually scored each frontal and sagittal videos of the jumps based on 17 LESS items. Comparison of LESS-risk score across jump techniques indicated no significant differences between gender within the same task, while DS- VDJ entailed a lower LESS-risk than SS- VDJ, suggesting that these jumps should not be used interchangeably. Participants tested to develop the classifier during SS- VDJs were also recorded using OpenCap, an iOS-based open-source and markerless motion capture system, which enabled the extraction of 16 biomechanical features. These features were used to train an optimized Fine Gaussian Support Vector Machine (SVM) classifier. The dataset was split with 80% (34 participants) for training and 20% (8 participants) for testing. achieving a 92 % accuracy in screening ACL injury risk based on the LESS score without expert intervention.