Forensic postmortem computed tomography in suspected unnatural adult deaths

Our study sought to evaluate validity of forensic postmortem CT in establishing cause of death (COD) in suspected unnatural adult death based on the reference standard of autopsy.

In our prospective, single-center study, 64 of 94 consecutive corpses (70.7 % male, mean age: 47.4 years) who underwent CT and autopsy between November 2013 and April 2019 were included in the analysis. Primary objective was agreement between CT and autopsy on primary COD using kappa statistics. Secondary objectives were competing COD and specific pathological findings.

Agreement on primary COD between forensic CT and autopsy without or in consideration of toxicological and histological findings was strong (85.9 % [55 of 64 corpses]; κ = 0.83 [95 %CI: 0.74 to 0.93] and 95.3 % [61 of 64 corpses]; κ = 0.94 [95 %CI: 0.84-1.04], respectively, McNemar p = 0.03). Sensitivity and specificity of CT in identification of acute heart failure, intracranial bleeding, burns and heat shocks, gunshot wounds, polytrauma, cranio-cerebral trauma, and strangulation or hanging was 100 %, each. Acute respiratory failure was detected with a sensitivity and specificity of 100 % and 96.8 %, cuts and stab wounds with 95.2 % and 100 %, and intoxication, pneumonia, or gastrointestinal bleeding with 60.0 % and 100 %, respectively. Agreement on competing COD was moderate (51.6 %, [33 of 64 corpses]; κ = 0.47 [95 %CI: 0.40 to 0.53], p < 0.001).

Forensic postmortem CT, complemented by external, toxicological, and histological examination was sufficiently valid to assess primary COD in the majority of suspected unnatural deaths with few restrictions.