Radiologists' Diagnostic Performance in Differentiation of Rickets and Classic Metaphyseal Lesions on Radiographs: A...

BACKGROUND. Despite evidence supporting the specificity of classic metaphyseal lesions (CML) for the diagnosis of child abuse, some medicolegal practitioners claim that CML result from rickets rather than trauma. OBJECTIVE. The purpose of this study was to evaluate radiologists' diagnostic performance in differentiating rickets and CML on radiographs. METHODS. This retrospective seven-center study included children younger than 2 years who underwent knee radiography from January 2007 to December 2018 and who had either rickets (25-hydroxyvitamin D level < 20 ng/mL and abnormal knee radiographs) or knee CML and a diagnosis of child abuse from a child abuse pediatrician. Additional injuries were identified through medical record review. Radiographs were cropped and zoomed to present similar depictions of the knee. Eight radiologists independently interpreted radiographs for diagnoses of rickets or CML, rated confidence levels, and recorded associated radiographic signs. RESULTS. Seventy children (27 girls, 43 boys) had rickets; 77 children (37 girls, 40 boys) had CML. Children with CML were younger than those with rickets (mean, 3.7 vs 14.2 months, p < .001; 89.6% vs 5.7% younger than 6 months; 3.9% vs 65.7% older than 1 year). All children with CML had injuries in addition to the knee CML identified at physical examination or other imaging examinations. Radiologists had almost perfect agreement for moderate- or high-confidence interpretations of rickets (κ = 0.92) and CML (κ = 0.89). Across radiologists, estimated sensitivity, specificity, and accuracy for CML for moderate- or high-confidence interpretations were 95.1%, 97.0%, and 96.0%. Accuracy was not significantly different between pediatric and nonpediatric radiologists (p = .20) or between less experienced and more experienced radiologists (p = .57). Loss of metaphyseal zone of provisional calcification, cupping, fraying, and physeal widening were more common in rickets than CML, being detected in less than 4% of children with CML. Corner fracture, bucket-handle fracture, subphyseal lucency, deformed corner, metaphyseal irregularity, and subperiosteal new bone formation were more common in CML than rickets, being detected in less than 4% of children with rickets. CONCLUSION. Radiologists had high interobserver agreement and high diagnostic performance for differentiating rickets and CML. Recognition that CML mostly occur in children younger than 6 months and are unusual in children older than 1 year may assist interpretation. CLINICAL IMPACT. Rickets and CML have distinct radiographic signs, and radiologists can reliably differentiate these two entities.