Determination of Osteopenia in Children on Digital Radiography Compared with a DEXA Reference Standard

Rationale and Objectives: The aim of this study was to explore the reliability of osteopenia diagnosis based on digital radiographs of appendicular skeleton obtained as part of routine clinical practice as compared with dual-energy x-ray absorptiometry (DEXA) gold standard (Z-score <-1). Materials and Methods: The study was an institutional review board-approved retrospective study of 58 children (mean age 12 years [4-18]). Digital radiographs of appendicular skeleton obtained within 6 months of DEXA scanning were presented in a blinded fashion to two musculoskeletal radiologists who were instructed to grade the level of mineralization. Sensitivity and specificity of each reviewer's osteopenia grading were calculated in comparison to lumbar DEXA Z-score values. Interobserver agreement was also calculated and significance evaluated with Bowker's test. Results: The reviewers correctly identified 28% of all patients with severe osteopenia (Z-score <=-2.0) as well as 26% of all patients with mild osteopenia (-2< Z-score <-1). Interobserver agreement for the correct diagnosis of bone mineral density (BMD) category ranged from 71% for normal BMD (Z-score >-1) to 0 and 25% for mild and severe osteopenia respectively. Conclusions: Visual diagnosis of osteopenia based on digital radiographs of appendicular skeleton has poor sensitivity and interobserver agreement. Clinical features and risk factors of pediatric patients should therefore guide DEXA evaluation and treatment recommendations.