Post-mortem diagnosis: evolving a team approach.

The decline in autopsy rates in the past four decades in developed countries (to less than 5% in the USA) has paralleled continued discrepancies between clinical and autopsy diagnoses of up to 20–30%. The interconnected causes of low autopsy rates include absence of predictable funding, clinical overconfi dence in diagnostic modalities, reluctance to request and undertake autopsies, decreased expertise in autopsy, a scarcity of auditable standards and mandated autopsy rates, and reluctance of families to consent to autopsy. The need for mortality data based on accurate post-mortem diagnosis of disease and identifi cation of cause of death has increased interest in minimally or less-invasive procedures to replace or augment conventional autopsy. However, uncertainty about the accuracy of less-invasive diagnostic techniques for post mortems and about integration with conventional autopsy has raised concerns about whether reliable population-based determination of cause of death can be maintained. In The Lancet, Sudhin Thayyil and colleagues report on the fi rst large, prospective, validation study of fetuses, infants, and children comparing post-mortem MRI with conventional autopsy and a uniquely defi ned minimally invasive autopsy. The MRI protocol, optimised to acquire non-contrast T1-weighted and T2-weighted images of the brain, spine, and body on a 1·5 Tesla magnet, needed about 90 min scan time in fetuses and 60 min in children, and used three-dimensional sequences for image reconstruction in diff erent planes. The conventional autopsy complied with UK national guidelines, including in-situ, macroscopic, and histological assessment of the brain and internal organs, and ancillary assessments, consisting of clinical history; ante-mortem diagnostic studies; post-mortem plain-fi lm radiography; external examination; placental histopathological examination for fetuses; and laboratory tests, including genetic, metabolic, and microbiological studies. The minimally invasive autopsy consisted of these common ancillary assessments and post-mortem MRI, with no postmortem sampling of tissues or body fl uids other than blood, and no histological assessments. MRI alone versus conventional autopsy was con cordant in 222 (55·5%, 95% CI 50·6–60·3) of 400 cases for identifi cation of cause of death or major pathological change. By contrast, minimally invasive versus conventional autopsy was concordant in 357 (89·3%, 85·8–91·9) cases. However, concordance varied substantially by age. A high concordance for fetuses (95–96%), and relatively high rates for newborn babies (81%) and infants (85%), contrasted with that of only 53·6% for children aged between 12 months and 16 years. Concordance between fetuses and infants was substantially greater than that reported in several previous smaller series, with one reporting that a minimally invasive autopsy provided information of at least equivalent clinical signifi cance to that of conventional autopsy in 32 (73%) of 44 fetuses. Post-mortem MRI is particularly valuable for delineation of anatomical abnormalities, and might be better than autopsy for identifi cation of structural brain abnormalities, especially when substantial post-mortem autolysis restricts pathological assessment. The high concordance in fetuses and infants shows the importance of common ancillary studies in this age group. Conversely, the low concordance in older childhood deaths refl ects the diff erent processes that caused those deaths. In children, acquired natural diseases were frequently missed, especially myocarditis, pneumonia, and sepsis. Identifi cation of such diseases usually requires macroscopic and histopathological assessment; specifi c diagnosis would not be expected from imaging or a blood test. This poorer concordance in children was similar to a large series of adult deaths referred for coroner’s autopsy, in which the major discrepancy rate for cause of death Published Online May 16, 2013 http://dx.doi.org/10.1016/ S0140-6736(13)60890-9