Autopsies using MRIs in babies and infants could help improve uptake rates

BMJ 2013; 346 doi: (Published 16 May 2013) Cite this as: BMJ 2013;346:f3205
  1. Zosia Kmietowicz
  1. 1BMJ

A large study has found that postmortem examinations in fetuses and infants up to 12 months using a combination of magnetic resonance imaging (MRI) and blood tests are as accurate as conventional methods.

MRI of an adult volunteer showing examples of the structures visible through imaging

The authors believe that the results provide sufficient evidence for minimally invasive autopsies to be offered routinely to parents whose babies have died, and that this could help to improve autopsy rates that have been falling for the past 40 years.

The use of postmortem examinations using MRI scans is currently restricted to only a handful of centres worldwide. Small scale studies have been reporting for more than a decade, but the accuracy of the method is uncertain.

Researchers from University College London and Great Ormond Street Hospital therefore analysed the results from 400 fetuses, babies, and children who were undergoing conventional postmortem examinations, and compared the cause of death with that obtained from whole body, postmortem MRI alone, or with other minimally invasive methods. These methods included blood samples taken by needle, visual examination of the body, and genetic and metabolic tests, but did not include tissue biopsies. The scans took place after 6 pm and before 8 am, so did not affect the access to MRI.

Reporting in the Lancet, the researchers found that autopsies using MRI alone resulted in the same cause of death as conventional autopsies in 222 of 400 cases (55.5%, 95% confidence interval 50.6% to 60.3%). But when MRI was combined with the other tests, the results concurred in 357 cases (89.3%, 85.8% to 91.9%).1

They also found that concordance varied with age, with the highest rate for fetuses (95-96%), newborn babies up to one month (81%), and infants up to 12 months of age (85%). For children aged between 12 months and 16 years, concordance was just 54%.

The differences were largely because deaths in fetuses and infants tended to be a result of structural abnormalities, while older children died from undetected pneumonia, myocarditis, and sepsis, said the researchers at a briefing in London on 15 May.

Sudhin Thayyil, consultant neonatologist at Great Ormond Street Hospital and University College Hospital and lead author, said that in 40% of cases an MRI was enough to decide to an accuracy of 99.4% whether a conventional postmortem examination would deliver anything interesting. In other cases, keyhole surgery could be used to retrieve samples of tissue.

He also said that autopsies using MRI could negate the need for microscopic brain examination when brain MRI is normal.

The Department of Health, which funded the study, has commissioned a report of how to establish a national service for minimally invasive postmortem examinations. Andrew Taylor, consultant radiologist at Great Ormond Street and a study author, believes that there could be six expert centres in England to which local pathology and radiology services could feed MRI scans for assessment.

Between 2000 and 2007, consent rates in the United Kingdom for fetal autopsy fell from 55% to 45%. The rates for neonatal autopsy fell from 28% to 21%, despite more parents being offered them, largely as a result of the retained organ scandal at Alder Hey hospital that was reported in 1999. Taylor told the briefing: “If we offer something less invasive, [parents] are more willing to take it up. Minimally invasive autopsy also has potential for religious groups that do not agree with conventional autopsy.”

He added: “Autopsies not only help us to establish the cause of death, but they often play an important role in advancing medical research and knowledge.” In a linked comment, Corinne Fligner and Manjiri Dighe of the University of Washington Medical Centre in Seattle, United States, said: “Crucial to success of an integrated post-mortem diagnostic programme will be clear performance standards, regular audits, physician training, and sufficient and stable funding to attract, train, and retain specialists and provide state-of-the-art resources for radiology and pathology.”2


Cite this as: BMJ 2013;346:f3205


View Abstract

Log in

Log in through your institution


* For online subscription