Whole-body MRI in suspected infant abuse.

Abstract

The purpose of our study was to examine the utility of whole-body MRI (WB-MRI) in the detection of skeletal and soft-tissue injuries in suspected infant abuse. Twenty-one infants (0-12 months) underwent WB-MRI for evaluation of suspected child abuse. WB-MRI at 1.5 T was performed using coronal and sagittal STIR sequences within 5 days of initial skeletal survey. Follow-up skeletal survey was performed in 16 cases. The "truth" was determined by integrating the initial and follow-up skeletal surveys, where available, into a summary of skeletal injuries (summary skeletal survey). Statistics included analysis of counts and proportions, concordance rate, sensitivity, and specificity. Summary skeletal survey and WB-MRI identified 167 fractures or areas of skeletal signal abnormality: 46 (27.5%) by both techniques, 68 (40.7%) by summary skeletal survey only, and 53 (31.7%) by WB-MRI only. WB-MRI had high specificity (95%) but low sensitivity (40%) for identifying fractures or signal abnormalities compared with summary skeletal survey. Thirty-seven classic metaphyseal lesions or metaphyseal signal abnormalities were identified: 11 (29.7%) by both techniques, 24 (64.8%) by summary skeletal survey only, and two (5.4%) by WB-MRI only. WB-MRI had very low sensitivity (31%) for identifying signal abnormality where classic metaphyseal lesions were seen with skeletal survey. WB-MRI had low sensitivity (57%) for identifying signal abnormality in areas where rib fractures were seen on skeletal survey. WB-MRI identified soft-tissue injuries such as muscle edema and joint effusions that, in some cases, led to identifying additional fractures. WB-MRI is insensitive in the detection of classic metaphyseal lesions and rib fractures, high specificity indicators of infant abuse. WB-MRI cannot replace the skeletal survey but may complement it by identifying soft-tissue abnormalities.