Predicting Radial Head Dislocation in Hereditary Multiple Osteochondromatosis

BACKGROUND: Forearm deformities are common in children with hereditary multiple osteochondromatosis (HMO) and may progress during growth, leading to functional impairment. Among these deformities, radial head dislocation (RHD) represents one of the most clinically significant and potentially preventable complications. Existing classification systems are primarily descriptive and show limited reliability in guiding surveillance and treatment decisions. This study aimed to identify radiologic predictors of RHD in HMO and to develop a quantitative, clinically applicable classification system to support risk stratification during skeletal growth. METHODS: A retrospective review was conducted of 143 patients (186 forearms) with HMO treated and followed between 2006 and 2024. Standard anteroposterior and lateral radiographs available at the time of evaluation were evaluated for lesion distribution, proportional ulnar length (PUL), radial bowing, ulnar variance, distal radial epiphyseal angle, carpal slip, and the presence of RHD. Forearms were classified using the Masada and Jo systems. Multivariate analysis was performed to identify independent predictors of RHD and to construct a risk-based classification framework. RESULTS: Radial head dislocation was identified in 36% of forearms. Patients with RHD demonstrated substantially lower PUL compared with those without dislocation, and a PUL threshold of ≤0.89 effectively distinguished high risk from low-risk cases. Absence of a distal radius lesion was also associated with an increased likelihood of RHD, whereas radial bowing and ulnar variance were not independent predictors. Existing classification systems showed limited applicability, with a considerable proportion of forearms remaining unclassifiable. On the basis of radiologic and statistical findings, a 3-tier risk classification was developed: type A1 (PUL ≤0.89 without distal radius lesion), type A2 (PUL ≤0.89 with distal radius lesion), and type B (PUL >0.89). CONCLUSION: Proportional ulnar length is the most reliable radiographic predictor of radial head dislocation in children with HMO. Incorporating distal radius lesion status improves risk stratification and enables a practical, growth-oriented framework for surveillance and surgical decision-making in pediatric patients. This study presents one of the largest single-center HMO cohorts to propose a reproducible risk stratification system that addresses unclassifiable cases in Masada/Jo criteria and integrates distal radius lesions as a novel modifier. LEVEL OF EVIDENCE: Level III.