Postmortem CT compared to autopsy in children; concordance in a forensic setting.

Objectives: The aim of this study is to assess the accuracy of postmortem CT (PMCT) in determining the cause of death in children who underwent a forensic autopsy because of a suspected nonnatural death.

Methods: We selected forensic pediatric autopsies at the Netherlands Forensic Institute, whereby the subject underwent PMCT between 1-1-2008 and 31-12-2012. Cause of death was independently scored by a radiologist and a pathologist. Cause of death was classified (1) in categories being natural, unnatural, and unknown; (2) according to the ICD-10; and (3) according to institutional classification.

Results: In the study period, 189 pediatric forensic autopsies were performed. Fifteen were excluded because of putrefaction. Of the remaining 174 autopsies, 98 (56%) underwent PMCT. PMCT and autopsy identified the same category in 69/98 cases (70%, kappa 0.49). They identified the same cause of death in 66/98 cases (67%, kappa 0.5) using ICD-10; in 71/98 (72%, kappa 0.62) using a forensic classification. PMCT performed better in unnatural deaths (59-67 % agreement) than in natural deaths (0 % agreement). If no cause of death was detected with autopsy, PMCT failed to identify a cause of death in 98% (39/40).

Conclusions: Pediatric PMCT does identify the majority of unnatural causes of death, but does not identify new diagnoses (true positives) if no cause of death is found during autopsy. Diagnostic accuracy in natural deaths is low.

Key Points: • The case mix is an important predictor for the concordance between PMCT and autopsy. • In case of an unnatural death, 72--81% of PMCT results matches autopsy results. • In case of a natural death, 0% of PMCT results matches autopsy results. • If no cause of death is identified with autopsy, 98% of PMCT results concurs.