RESUMO
OBJECTIVE: To determine the diagnostic capabilities of unenhanced postmortem computed tomography (UPMCT) in detecting traumatic abdominal injuries. MATERIAL AND METHODS: Cases of traumatic death with both UPMCT and classical autopsy were collected retrospectively from our institution "virtopsy" database in a period of 5 years. Cadavers with gunshot injuries were excluded. Sensitivity, specificity, accuracy, negative (NPV) and positive (PPV) predictive values of PMCT globally and for hemoperitoneum, liver, spleen, pancreas and kidney injuries individually were estimated using the autopsy report as gold standard. RESULTS: Seventy-one cadavers were included. UPMCT had a sensitivity of 80% and a specificity 94%, with an accuracy of 83%, a PPV of 98% and a NPV of 59% for the diagnosis of traumatic abdominal injuries. The highest sensitivity was obtained for the detection of hepatic injuries (71%) and the lowest for pancreatic injuries (12%). UPMCT had a specificity of 100% for the detection of hemoperitoneum. A NPV of 98% was found for the detection of perihepatic hematomas. CONCLUSION: The low sensitivity and low NPV do not support the use of UPMCT as an alternative to conventional autopsy to diagnose and/or rule out traumatic abdominal injuries. Nevertheless, UPMCT remains a helpful tool as it helps detect hemoperitoneum and virtually exclude presence of perihepatic hematomas.