virtual autopsy
Radiology (Stephen Gaeta 2011)
In recent years there has been increasing interest in the use of MRI for the diagnosis of disease and injury in the dead, in order to ascribe a cause of death.
The UK government has declared its interest in this application on several occasions, particularly following the damaging organ and tissue retention 'scandals' of recent years, including those involving Alderhey and Bristol hospitals.
BMA News (Saturday April 13th 2002) ran a front page headline 'Post mortem MRI scans could drain vital resources' following the government' declaration of interest, but highlighted the lack of human and financial resources available to undertake such imaging and interpretation. A full body MRI costs in the region of £1000 - considerably more than a conventional autopsy.
The following week a further story ran on page 6 'MRI moves could cut the scalpel out of autopsies', and quoted Dr Ian Roberts of Oxford as saying that pathologists had ' ... lost the PR battle', referring to the poor image that pathologists and the use of autopsy had in the minds of the general public.
The UK government have recently advertised a contract tender for investigation of the use of MRI autopsies, a contract worth £750,000.
One of the main criticisms of the use of MRI as an alternative to conventional autopsy is the lack of validation of this imaging technique. MRI post mortem examinations had previously been reported in the fields of neuropathology and paediatrics where it has had mixed results.
Alderstein et al (2003) compared MRI with autopsy in perinatal cases and were disappointed. In particular, as had been previously reported, MRI was not good at detecting major malformations such as cardiac anomalies. They calculated the positive predictive value of MRI for detecting such malformations as 80%, and concluded that although 15% more families consented to an MRI autopsy only, a normal MRI did not equate to a normal baby, and that parents giving consent for an MRI post mortem only needed to be given that information.
In Manchester, UK, a team investigated the use of MRI autopsies in Coroner's cases (Bisset et al 2002). They had access to a private scanner, paid for by the local Jewish community, and reported 53 cases. Of these only 6 cases received both an MRI scan and a full post mortem. The remaining cases were given a cause of death based on the medical notes and the MRI scan.
The short paper showed that MRI was not able to accurately assess ischaemic heart disease - the most common cause of death in this country. However, they concluded that the use of MRI was a credible alternative to autopsy, stating that doctors only accurately certify the cause of death in 31-75% of cases, and that MRI autopsies were at least as good as that!
Not surprisingly, this paper resulted in a flurry of responses and letters, all criticising the lack of validation and the inability of the technique to identify the most common cause of death.
A team led by Dr Roberts at Oxford has since reported on a further 10 cases (Roberts et al 2003). They concluded that the MRI was unable to;
- image coronary artery lesions
- differentiate thrombus from post mortem clot, and
- differentiate pulmonary oedema from pneumonic exudates.
In addition changes associated with decomposition, such as gas formation in the biliary tree, caused immense interpretative problems for radiologists. It was found that those radiologists with experience of interpreting post mortem images were no better than inexperienced colleagues in such cases, and it is recommended by the team that MRI autopsies are carried out within 24 hours of death. This obviously has immense resource implications.
Most of the literature comments on the utility of MRI post mortems as an adjunct to full autopsy, and Benbow and Roberts (2003 pp. 417-423) provide an interesting overview of other autopsy adjuncts, such as needle autopsies etc. However, Thali and his team in Switzerland have been carrying out MRI and multi-slice CT examinations in forensic cases, including victims of gunshot wounds. They have developed an imaging research programme 'Virtopsy' and have published a great deal of papers highlighting the ability to image features of wounds and carry out post-image processing in 3D.
The future of the autopsy seems to be under threat, but the overwhelming evidence suggests that it is still the 'gold standard' in diagnosis, and that the use of imaging techniques will remain the preserve of an autopsy adjunct rather than a replacement.
Institute of Forensic Medicine, University of Bern at NLM Visible Proofs exhibition
references
Virtual Autopsies from Visualiseringscenter C on Vimeo.
postmortem imaging overview
This chapter will provide an overview of the potential utility of post-mortem CT and MRI in the fields of paediatric/ perinatal pathology, sudden adult death and forensic casework, and introduce the reader to developments in the field of three dimensional (3D) image reconstruction and micro-imaging.
Jones R. Postmortem imaging: an update. Chapter 8 In: Kirkham N, Shepherd N. Progress in Pathology Vol 7, Cambridge University Press 2007
Essentials of forensic imaging
Menu
The Virtopsy approach: 3D optical and radiological scanning and reconstruction in forensic medicine
See at Google Books here
Virtopsy innovations
- The Virtopsy group has developed a prototype system using Micosoft Kinect allowing gesture control of post mortem CT datasets. Watch the video here... This kind of 'hands free' system will be ideal in the post mortem room, and represents an exciting development for autopsy practice.
3D CT skull
Click on the link below to download a 3D QuickTimeVR movie of a skull created from CT data in OsiriX radiology software...