Evaluation of Ultra-Low and Low-Dose CT Skeletal Survey for Fracture Detection in Assessments for Suspected Non-Accidental Injury: a Prospective Post-Mortem Study.

This prospective study aimed to evaluate the capabilities of two novel vertex-toes CT protocols (ultra-low and low-dose) compared to a standard high-dose CT protocol plus radiographic skeletal survey (SS) in a post-mortem setting. The study sought to assess perceived image quality (IQ) and the fracture detection capabilities of each CT protocol. Participants aged 2 years and under received three whole-body CT scans, followed by a SS. Radiologists reviewed images in the sequence of ultra-low, low-dose, to standard-dose CT, then SS. The reviewers graded perceived IQ on a Likert scale. Gwet’s agreement coefficient (AC)-2 and cumulative link mixed models quantified interobserver agreement and IQ differences, respectively. Fractures detected by each modality were compared across anatomical regions, with positive specific agreement (PSA) assessing interobserver agreement. Radiation doses were calculated for each CT protocol. Significant differences in IQ ratings were observed across CT protocols for all anatomical regions (ribs [ultra-low–low], p = 0.0001; femur [ultra-low–low], p = 0.0001; all others, p < 0.001), alongside low interobserver agreement in IQ (lumbar spine [low-dose]-AC = 0.78; humerus [low-dose]-AC = 0.77; all others, AC < 0.7). Among 29 participants, 69 fractures were identified across modalities, all being detected by the standard-dose CT. Ultra-low and low-dose CT identified 62 (90