Injury Detection in Traumatic Death: Postmortem Computed Tomography vs. Open Autopsy

Abstract

Abstract Objective To compare postmortem computed tomography (PMCT) vs. autopsy in detecting and analyzing injuries due to traumatic deaths. Materials and methods In this retrospective study, a cohort of 52 subjects were purposively sampled to reflect a broad range of injuries. Injuries from autopsy and PMCT reports were coded using the Abbreviated Injury Scale (AIS) and the level of agreement of AIS 2+ and 3+ injuries were compared. Results A combined total of 353 AIS-coded injuries were detected - PMCT detected 63% and autopsy detected 74% of injuries. PMCT identified 92 (26%) additional injuries missed by autopsy. PMCT missed 131 (37%) injuries. The kappa value for agreement between the two modalities for presence of injuries was moderate for the majority of anatomic regions [head (κ = 0.53), thoracic organs (κ = 0.58), upper extremity (κ = 0.53), pelvis (κ = 0.45), skeletal chest injuries (κ = 0.57)]. Kappa value was least among abdominal (κ = 0.07) and vascular injuries (κ = 0.10). Substantial agreement (κ = 0.69) was present in lower extremity injuries. PMCT outperformed autopsy for bony injuries, in particular, pelvic injuries, base of skull fractures and upper extremity injuries. PMCT better detected pneumothoraces. Soft tissue injuries, particularly abdominal organ injuries, lung contusions and vascular injuries were better detected by autopsy. When injuries missed by PMCT were re-assessed, subtle but inconclusive imaging findings were identified. Conclusion A combination of PMCT and autopsy can detect more injuries than either modality in isolation. PMCT detected a considerable number of injuries missed by autopsy and vice versa.