Structural fatigue modeling of cumulative mechanical stress in the athletic elbow using routine radiographs.

Abstract

Elbow overuse injuries are common in athletic populations and are typically diagnosed only after symptom onset, when cumulative mechanical stress has already resulted in structural compromise. Conventional radiographic evaluation remains largely qualitative and lacks objective indicators for assessing stress accumulation under repetitive loading, limiting opportunities for early preventive intervention. To develop and externally validate an imaging-based structural fatigue modeling framework for objectively quantifying cumulative mechanical stress in the athletic elbow using routine radiographs, enabling early overuse injury risk stratification prior to clinical manifestation. The proposed framework was evaluated using two independent datasets. The primary dataset comprised elbow radiographs from athletes exposed to varying levels of repetitive elbow loading, accompanied by workload history and longitudinal clinical follow-up. An external validation dataset was obtained from an independent athletic cohort. Patient-specific anatomical reference landmarks were extracted to establish normalized elbow geometry. Fatigue-sensitive structural indicators-including regional density imbalance, articulation continuity deviation, and directional micro-deformation gradients across the humeroulnar and radiocapitellar joints-were quantified without pixel-wise segmentation. These indicators were integrated using orthopaedic biomechanical reasoning to derive a cumulative mechanical stress index. Model performance was assessed using cross-validation on the primary dataset and independent testing on the external validation dataset. On the primary dataset, the framework achieved an overall accuracy of 91.6%, with a sensitivity of 90.8% for identifying high-risk pre-symptomatic athletes and a specificity of 93.2% for low-risk cases. On the external validation dataset, performance remained consistent, achieving an accuracy of 89.4%, sensitivity of 88.1%, and specificity of 91.7%. Structural fatigue indicators demonstrated strong correlation with subsequent injury occurrence (r = 0.69 in both the primary and external datasets; p < 0.01), outperforming conventional qualitative radiographic assessment and texture-based analysis methods. Imaging-derived structural fatigue modeling enables objective assessment of cumulative mechanical stress in the athletic elbow using routine radiographs. The proposed framework demonstrates robust performance across independent datasets and provides a clinically interpretable, segmentation-free decision-support tool for early risk stratification, supporting proactive workload management and injury prevention in sports medicine and orthopaedic practice.