Diagnostic accuracy of postmortem MRI for musculoskeletal abnormalities in fetuses and children.

Objectives: The aim of this study was to compare the diagnostic accuracy of postmortem magnetic resonance (PMMR) imaging specifically for musculoskeletal pathology in fetuses and children, compared with conventional autopsy, with radiographic and histopathology assessment.
Methods: Institutional ethics approval and parental consent was obtained. A total of 400 cases underwent PMMR using a 1.5 T Siemens Avanto MR scanner before conventional autopsy. PMMR images and autopsy findings were reported blinded to the other data, respectively.
Results: A total of 400 cases were reported, with 277 (69%) fetuses (185 ≤24 weeks' gestation and 92 >24 weeks' gestation) and 123 children (42 newborns aged <1 month, 53 infants ≤12 months and 28 children ≤16 years). Musculoskeletal (MSK) abnormalities were found at autopsy in 47/400 (11.7%). Overall sensitivity and specificity (with 95% confidence interval) of PMMR for MSK pathology were 51.1% (37.0, 65.0) and 98.2% (96.2, 99.2), with positive and negative predictive values of 79.3% (61.6, 90.2) and 93.8% (90.8, 95.9), respectively. Overall accuracy between PMMR and autopsy for MSK abnormalities was 92.7% (89.7, 94.9). In some cases, PMMR detected MSK abnormalities not routinely examined for or detected at traditional autopsy.
Conclusion: Minimally invasive autopsy has good diagnostic accuracy for the exclusion of MSK abnormalities, but sensitivity is relatively poor. When PMMR is used with clinical examination and skeletal radiographs, all skeletal and soft tissue abnormalities of clinical significance are likely to be detected, even if not directly relevant to the cause of death.
(© 2014 John Wiley & Sons, Ltd.)