MINI-SYMPOSIUM: NON-INVASIVE RADIOLOGICAL AUTOPSY

An image of sudden death:utility of routine post-mortemcomputed tomographyscanning in medico-legalautopsy practiceChris O’Donnell

a Unlike England and Wales but similar to many areas of the world, all

AbstractPost-mortem computed tomography (CT) is an established technique at

Victorian Institute of Forensic Medicine (VIFM) used to assist pathologists

in determining cause and manner of death. It also plays an important role

in identification of deceased individuals as exemplified by the 2009

“Black Saturday” Victorian bushfires in which the remains of 164 individ-

uals were subjected to disaster victim identification procedures. CT scan-

ning is now explicitly incorporated into the Victorian Coronial legislation

(Coroners Act 2008), and is an important component of the preliminary

examination process whereby a pathologist reviews the circumstances

of a death, any pre-existing medical history, whole body CT images, the

external appearances of the body and expedited (overnight) toxicological

screen results, so that a recommendation to the Coroner may be formu-

lated regarding the likely cause of death and necessity for autopsy. This

process has seen a reduction in autopsies from a mean of 62% over

the last 5 years to 47% of admissions in less than 12 months. VIFM

pathologists perform the primary interpretation of CT images in consulta-

tion with a radiologist. A process of quality audit has been instituted in

order to detect systematic errors in this interpretation, in addition to

a structured education programme directed to correct those errors. New

imaging techniques, notably whole body CT angiography and dual energy

CT have the prospect of even more substantial forensic application.

Keywords angiography; cause of death; computed tomography;

coroner; post-mortem

Introduction

In mid 2005, a computed tomography (CT) scanner was installed

into the mortuary of the Victorian Institute of Forensic Medicine

(VIFM). Since that time all deceased persons admitted to the

Institute have been CT scanned from head to toe, and images

permanently stored on a picture archiving and communication

system (PACS). Well over 15,000 cases have now been exam-

ined. CT images do not replace autopsy but assist pathologists in

determining cause of death, manner of death, mechanism of

injury and documentation of injury for presentation as evidence

in court.1 CT is also useful in identification of the deceased where

Chris O’Donnell MBBS FRANZCR MMed GradDipForMed is the Principal

Consultant Radiologist at the Victorian Institute of Forensic Medicine

and Department of Forensic Medicine, Monash University, Australia.

DIAGNOSTIC HISTOPATHOLOGY 16:12 552

other techniques are not available or difficult to attain.2 In a mass

disaster scenario, CT is used as a triage tool.3 In the 2009 “Black

Saturday” Victorian bushfires, radiologists used CT imaging to

separate human from non-human remains, determine the

number of individuals within a body bag, assist odontologists

and anthropologists in assigning gender and age, detect disease

processes or medical procedures, and at the time of autopsy,

localize metallic items within the body for retrieval by the

pathologist. Routine CT scanning prior to autopsy or other bodily

interventions produces images that are available at any time for

retrospective review by pathologists or other interested parties,

for example following the discovery of additional information by

police in a process known as “digital exhumation”.

Post-mortem CT interpretation is not the same as clinical CT.

Images that would normally be considered unacceptable to the

clinical radiologist are routine in post-mortem scanning. Notable

contributory factors include a lack of oral or intravenous radio-

graphic contrast, malpositioning of the body in the gantry, and

metallic artefact due to foreign bodies located in or on the

deceased person. Image analysis requires an understanding of

the changes to visceral anatomy that routinely occur after death,

features of the agonal process, artefacts of death including

autolysis and putrefaction, as well as the consequences of

resuscitation including external cardiac massage. Failure to

recognize or understand these features can lead the reader of CT

images into errors of analysis with the potential for inaccurate

evidence or conflict with the pathologist’s autopsy findings.

Radiological findings of visceral pathology on post-mortem CT

are similar but invariably more extreme than CT performed in the

clinical environment given that individuals have succumbed to

that pathological process.

The interpreter of post-mortem CT must have a thorough

understanding of forensic autopsya practice in particular the

nature of injury and outcome of trauma, realizing that

the mechanism or manner of death may be as important to the

forensic pathologist as the consequences of inflicted injury on

a particular organ. Ultimately the individual who produces

a written radiological report must understand the relevant

legislation, be prepared to provide verbal expert evidence to

a court and endure the demands of a legal system that rightly

tests the validity and interpretation of presented facts in

a rigorous manner. Substantial background information

including (but not restricted to) the circumstances surrounding

the death should be available at the time of image analysis.

Ideally CT imaging should be undertaken in a co-operative

environment with pathologists and radiologists working together

in a forensic institute or situation where the CT scanner is co-

located with pathologists such as in a hospital department. It is

not the province of an occasional user working in isolation.

This issue of CT image interpretation has recently been

explored by Filograna.4 She has discussed the three sources of

medico-legal deaths including so-called routine coroner’s cases, are

examined in a single Institute by forensic pathologists. Thus where the

word forensic is used within the text it is taken to include all deaths, both

routine and forensic i.e. it includes deaths that would be examined by

histopathologists in England and Wales.

� 2010 Elsevier Ltd. All rights reserved.

MINI-SYMPOSIUM: NON-INVASIVE RADIOLOGICAL AUTOPSY

error in diagnostic imaging; notably perceptual (finding present

but not recognized), cognitive (incorrect interpretation of

a finding) and system factors (organizational issues in the

institution) and related them to the conduct of post-mortem CT

imaging. The relative novelty of post-mortem CT imaging

means that all three factors are commonly encountered.

Perceptual errors are inevitable when the observer is inexperi-

enced with the imaging modality or there is a new application of

that modality i.e. CT imaging of the deceased. Cognitive error is

also likely if the paradigm of clinical CT interpretation is applied

directly to the evaluation of the deceased and system error will

flourish if the imaging is being performed and interpreted in an

environment that is divorced from the forensic pathological

context.

Post-mortem CT images should therefore be assessed by those

with a background in forensic medicine and understanding of

post-mortem pathology and cross-sectional imaging. This has

spawned a new subspeciality termed “necro-radiology”.5 At VIFM

the sheer number of admissions and limited access to radiolog-

ical expertise precludes the written interpretation of all CT scans

by a radiologist (as routinely occurs in clinical practice).

Pathologists have taken primary responsibility for the viewing of

images with a radiologist providing consultation on a case by

case basis. This pragmatic approach has been criticized yet

forensic pathologists have many of the desirable attributes

described above i.e. an in-depth understanding of traumatic

effects on the body and experience in provision of expert

evidence as well as court practices. Pathologists are well aware of

the artefacts of death and the mistakes of interpretation that can

occur in forensic practice, described so eloquently in the classic

paper by Moritz.6 Pathologists at the VIFM have been educated

on the CT findings of such artefacts as well as the radiological

correlates of pathological processes e their particular area of

expertise. Unlike the radiologist who is often constrained by

limited provision of background information, the pathologist has

considerable data available at the time of CT reporting notably

circumstances surrounding death provided by police, previous

medical history, toxicology and ultimately the ability to perform

external examination of the deceased person even if no autopsy

is forthcoming. Any CT findings identified on the pre-autopsy CT

scan can if necessary be viewed directly at the time of autopsy.

This process of ‘validation’ has been invaluable and an ongoing

educative exercise to pathologists.

Although initially reluctant, pathologists have embraced the

new technology notwithstanding the potential for interpretive

error due to lack of experience7 or understanding of CT image

acquisition. An ongoing teaching programme has been supple-

mented by a quality audit whereby 10% of all cases are retro-

spectively reviewed by the radiologist. Radiologist’s findings are

matched with the pathologist’s written CT observations and any

discrepancies graded as substantial (i.e. a CT finding that might

reasonably be expected on radiological grounds alone to be the

cause of death), minor (i.e. a CT finding that of its own might not

reasonably be expected to be the cause of or responsible for the

death, but might possibly be considered to be a contributory

factor to the death) or incidental (i.e. a CT finding that is inci-

dental to the cause of death but is of particular medical or

pathological interest). Substantial and minor discrepancies are

discussed with the pathologist and in appropriate cases, images

DIAGNOSTIC HISTOPATHOLOGY 16:12 553

reviewed at a weekly departmental pathology meeting. This

process allows for the identification of systematic error and

redress by education.

The audit has secondary consequences. It reveals technical

issues i.e. CT hardware failure or findings considered by the

radiologist to be the result of error by technical staff. Any such

finding is relayed to the mortuary management and corrections

made. If systematic technical error is revealed then re-education

of technologists is instituted. The process also acts as an ongoing

learning process for the reviewing radiologist since CT results are

directly correlated with the pathologist’s autopsy findings and

interpretations.

In the 4 years following installation, CT was used very much as

an adjunct to the routine procedures of forensic pathology at VIFM.

For example if CT revealed an obvious cause of death e.g. ruptured

atheromatous, abdominal aortic aneurysm, and the circumstances

were not deemed suspicious, then a recommendation might be

made to the coroner that autopsy was not required as a ‘natural’

disease process responsible for the death had been identified. If the

cause of death was not obvious on circumstantial grounds but

there was a significant history of medical co-morbities, non-

suspicious circumstances, and no specific CT findings (including

no evidence of significant trauma) then autopsy might not be

performed if an objection to such a procedure had been raised by

the senior next of kin and the coroner was of the view that death

was most probably the result of natural causes (otherwise not

specified). Alternatively the circumstances of deathmight not have

been suspicious yet CT revealed a concerning finding such as

a subdural haematoma or unexplained healing rib fractures in

a child prompting further investigations including a full autopsy.

CT was routinely used by pathologists following autopsy to clarify

an autopsy finding e.g. demonstration of a pelvic fracture to

explain detected pelvic haematoma or provide an image for

presentation to investigators e.g. multiple, rounded, depressed

skull vault fractures due to injury inflicted with a hammer. On

occasion pathologists might overlook the examination or docu-

mentation of a particular body region at the time of autopsy yet

retrospectively be able to review CT images prior to completion of

their report e.g. diameter of the aortic valve.

All previous applications of post-mortem CT have continued,

however, recent changes to the Victorian Coroners Act8 have

reinforced the use of CT in everyday practice. The Act creates the

concept of a preliminary examination entailing (a) visual

including dental examination of the body (b) review of personal

and health information, (c) taking of bodily fluids, (d) imaging of

the body including X-rays, CT, MRI, US and/or photography, (e)

taking of surface swabs and (f) fingerprinting of the body. A

preliminary examination (or component part thereof) is per-

formed on all deceased persons reported to the coroner and

admitted to VIFM. The Act allows the coroner to determine that

a reportable death requires no further investigation “if a medical

investigator conducts a medical examination on the deceased

person and provides a report to the coroner that includes an

opinion that the death was due to natural causes”. At VIFM a so-

called duty pathologist is the nominated medical investigator

created in response to the Act. Using the facility of the prelimi-

nary examination, the duty pathologist forms a judgement on

whether the death was natural and provides to the coroner

a recommendation as to the necessity for an autopsy. The

� 2010 Elsevier Ltd. All rights reserved.

Figure 2 Coronal MPR view of the heart through the root of the aorta

showing a large volume haemopericardium towards the apex of the lateral

ventricle with a crescent of hyperdense blood in the wall of the ascending

aorta (arrow). Findings are indicative of ruptured Type A dissection of the

ascending aorta.

MINI-SYMPOSIUM: NON-INVASIVE RADIOLOGICAL AUTOPSY

coroner then makes a determination on whether autopsy should

progress, usually in consultation with the deceased’s senior next

of kin.

The duty pathologist reviews all aspects of the preliminary

examination including CT images prior to formulation of an

opinion on whether a death is natural and/or recommendation to

the Coroner for autopsy. On that basis CT analysis is now

formalized and integral to the conduct of forensic pathology

services at VIFM; albeit the issues addressed at the time of

preliminary CT assessment are not necessarily complete or well-

defined. For example details on the depth and direction of

a penetrating metallic object or the number and location of

fractures are less important than their presence, prompting the

pathologist to brief the coroner on the need for an autopsy. The

specific CT findings can later be analyzed to assist the pathologist

in performing that formal autopsy or subsequently at the time of

constructing the autopsy report in consultation with an expert

radiologist. In contrast the detection of subarachnoid haemor-

rhage and a basilar aneurysm (Figure 1) by the duty pathologist

provides a likely cause of natural death that may obviate the

need for further coronial investigation by way of an autopsy.

The Act also allows the coroner to “impose conditions on the

manner in which an autopsy is to be performed”8 such as

limiting the number of body cavities to be explored or the organs

removed. CT scanning can be useful in this regard specifically if

the duty pathologist finds an abnormality in a particular

anatomical region but is unable to determine the cause of that

abnormality e.g. a cerebral mass lesion but no other abnormality,

or spontaneous haemopericardium but no cause evident. The

two likely causes of such a condition are ruptured thoracic aortic

dissection (Figure 2) or myocardial infarct. The exact determi-

nation of which pathology is responsible could have profound

consequences for the family of the deceased, especially if the

deceased is young, as thoracic dissection may be associated with

Figure 1 Midline sagittal MPR view of the brain in a deceased individual

who collapsed and died suddenly. Extensive subarachnoid and intra-

ventricular blood is associated with a large, hyperdense ovoid mass

(arrow) in the pre-pontine cistern compressing the medulla and pons

posteriorly. Appearances are due to a large basilar artery aneurysm that

has spontaneously ruptured into the subarachnoid space.

DIAGNOSTIC HISTOPATHOLOGY 16:12 554

familial genetic conditions including Marfan syndrome. Limited

autopsy prompted by the CT scan findings (including the exclu-

sion of significant pathology elsewhere) may be more acceptable

to next of kin when they are considering the possibility of

autopsy, especially if the value of such a procedure (particularly

in accurately determining the cause of death) is explained to the

family members.

Subsequent to the adoption of a duty pathologist at VIFM in

mid 2009, analysis of autopsy rates has shown a substantial

reduction to 47% of admissions from a mean of 62% over the

previous 5 years. These autopsy rates had been stable since

2004/05 despite the installation of the CT scanner in mid 2005

(Figure 3). Although CT is not necessarily entirely responsible for

this decline, the formalized process of preliminary examination

has encouraged pathologists to use CT images as part of their

formulation and advice to the coroner.

The future of post-mortem CT imaging is assured. Newer

techniques of whole body, minimally invasive CT angiography9

promise an even greater contribution to forensic practice and

technical advances such as dual energy CT, the prospect of

quantitative analysis as well as improved detection of pathology

including subcutaneous haematoma.10 Magnetic Resonance

Imaging (MRI) with its improved delineation of soft tissues will

also make a substantial contribution despite the cost of instal-

lation and technical difficulties in the mortuary environment.11

Summary

In summary CT scanning at VIFM has three main roles: (a) a tool

for triage by the duty pathologist to determine if autopsy should be

recommended to the Coroner (in conjunction with review of

circumstances, medical history, toxicology and external exami-

nation), (b) an adjunct to autopsy by predicting findings for the

pathologist, clarification of observations and better understanding

of trauma mechanisms and (c) assistance in identification of the

� 2010 Elsevier Ltd. All rights reserved.

Admisions to VIFM

Years

Pe

rce

nta

ge

100

80

60

2001/02

2002/03

2003/04

CT installed Duty path.

Non-autopsy

Autopsy

2004/05

2005/06

2006/07

2007/08

2008/09

2009/10

40

0

20

Figure 3 Graph demonstrating the ratio of autopsy to non-autopsy cases

admitted to the VIFM since 2001/02. CT scanner was installed into the

mortuary in mid 2005 yet the ratio of autopsy to non-autopsy cases

changed very little until the introduction of the duty pathologist at the

end of 2009. Note a drop in the autopsy rate in 2009/10 to 47% from an

average of 62% in the preceding 4 years.

MINI-SYMPOSIUM: NON-INVASIVE RADIOLOGICAL AUTOPSY

deceased. In the future more sophisticated applications of CT

including whole body angiography and dual energy CT as well as

MRI will enhance these roles assisting pathologists in better

understanding of the cause and mechanism of death. A

DIAGNOSTIC HISTOPATHOLOGY 16:12 555

REFERENCES

1 O’Donnell C, Rotman A, Collett S, Woodford N. Current status of

routine post-mortem CT in Melbourne, Australia. J Forensic Sci Med

Pathol 2007; 3: 226e32.

2 Blau S, Robertson S, Johnstone M. Disaster victim identification: new

applications for postmortem computed tomography. J Forensic Sci

2008; 53: 956e61.

3 Rutty GN, Robinson CE, BouHaidar R, Jeffery AJ, Morgan BJ. The role

of mobile computed tomography in mass fatality incidents. J Forensic

Sci 2007; 52: 1343e9.

4 Filograna L, Tartaglione T, Filograna E, Cittadini F, Oliva A, Pascali VL.

Computed tomography (CT) virtual autopsy and classical autopsy

discrepancies: radiologist’s error or a demonstration of post-mortem

multi-detector computed tomography (MDCT) limitation? Forensic Sci

Int 2010; 195: e13e7.

5 O’Donnell C, Woodford N. Post-mortem radiology e a new sub-

specialty? Clin Radiol 2008; 63: 1189e94.

6 Moritz AR. Classical mistakes in forensic pathology. Am J Clin Pathol

1956; 26: 1383e97.

7 Kremer C, Racette S, Marton D, Sauvageau. Radiographs interpreta-

tion by forensic pathologists: a word of warning. Am J Forensic Med

Pathol 2008; 29: 295e6.

8 Victorian Coroners Act 2008. Available at: www.legislation.vic.gov.au/

(accessed Feb 2010).

9 Ross S, Spendlove D, Bolliger S, et al. Postmortem whole-body CT

angiography: evaluation of two contrast media solutions. AJR Am J

Roentgenol 2008; 190: 1380e9.

10 Persson A, Jackowski C, Engstrom E, Zachrisson H. Advances of dual

source, dual-energy imaging in postmortem CT. Eur J Radiol 2008; 68:

446e55.

11 Yen K, Lovblad KO, Scheurer E, et al. Post-mortem forensic neuro-

imaging: correlation of MSCT and MRI findings with autopsy results.

Forensic Sci Int 2007; 173: 21e35.

� 2010 Elsevier Ltd. All rights reserved.