Post-Mortem Computed Tomography (PMCT ... - espr.org · Post-Mortem Computed Tomography (PMCT) Diagnostic Accuracy in Children Shelmerdine SC, Davendralingam N, Palm L, Minden T,

Post-Mortem Computed Tomography (PMCT) Diagnostic Accuracy in Children

Shelmerdine SC, Davendralingam N, Palm L, Minden T, Cary N, Sebire NJ, Arthurs OJ

Departments of Clinical Radiology & Pathology, Great Ormond Street Hospital, London, UK

ESPR 2019 – Thursday, 16th May 2019 – Outreach and Post Mortem Session

SCS is supported by a RCUK/ UKRI Innovation Fellowship and Medical Research Council (MRC) Clinical Research Training Fellowship

(Grant Ref: MR/R00218/1).

This award is jointly funded by the Royal College of Radiologists (RCR).

Disclosures

Aim: Determine diagnostic accuracy of whole body post-mortem CT in children

Background

• Global decline in parental consent for childhood autopsy • Need to identify ‘cause of death’ remains

➔Driving need for non-invasive alternatives (imaging)

• CT = widely accessible, utilised in adults

• Much debate for appropriate use in children• Small studies, heterogenous patient group

• Guidelines for CT (RCPath, NODO) on limited evidence

Ethically approved, single centre, retrospective observational studyParental consent for research obtained cases

Inclusion Criteria:All CTs in children, 6 year period (2012-2018)Matching autopsy reports available

Exclusion Criteria:Perinatal deaths and pathology specimen imaging

Autopsy Protocol

1/7 specialist paediatric pathologistsRCPath and European autopsy guidelines

Aware of CT findings prior to autopsy

Imaging Protocol

Non-contrast CT before autopsy, MDCT systemBone/Soft Tissue kernel, 0.625mm collimation

Reported by 1/3 attending paediatric radiologists

Methods

Primary Outcome:Main pathological lesions/ cause of death

Secondary Outcome:Correct identification of any pathological lesion per body systemIrrespective of cause of death

Findings expressed as diagnostic accuracy rates:Sensitivity, specificity, PPV, NPV, concordance rates

Data Analysis

223 Post-mortem CTs performed between January 2012 -2018

136 Post-mortem CTs in infants and children with matching standard autopsy

87 cases excluded:71 Perinatal deaths11 Pathology reports unavailable3 Pathological specimens2 Minimally-invasive tissue biopsy autopsy (no full body dissection)

Median age: 7 months (range: 2 days – 14 years 8 months), Male gender: 74/136 (54.4%) Time from death to CT : 4 days (range: 1 – 22 days) Time from CT to autopsy: 0 day (range: 0 – 8 days)Commonest indication: Found ‘unresponsive’ (65/135, 42.6%)

Results

Results

POSITIVE(Abnormal Autopsy)

NEGATIVE(Normal Autopsy)

TOTAL

POSITIVE(Abnormal CT)

55 agreed 11 ‘overcalls’ 66

NEGATIVE(Normal CT)

22 ‘misses’ 48 normal/unexplained deaths

70

TOTAL 77 (77/136, 56.6%)

59(59/136, 43.4%)

136

• Sensitivity: 71.4% [60.5, 80.3] • Specificity: 81.4% [69.6, 89.3]• PPV: 83.3% [72.6, 90.4]• NPV: 68.6% [57.0, 78.2]• Concordance: 75.7% [67.9, 82.2]

CT found a cause of death for 55/136

(40.4%) of all cases.

55 ‘agreed’ findings

25 Traumatic brain/ body injuries (2 cranio-facial injuries, 1 thoracic injury, 15 traumatic brain injuries, 3 multisystem injuries (abdomen/chest/brain) and 4 multiple skeletal injuries)

6 Pulmonary edema

3 Postoperative complications (2 relating to congenital heart disease; 1 relating to congenital diaphragmatic repair failure)

4 Aspiration pneumonia

2 Metabolic bone disease

Results

2 Asthma attack with pneumothorax

2 Dilated cardiomyopathy

11 others (1 muscle wasting/ malnutrition; 1 neck laceration; 1 multiple insufficiency fractures; 1 VACTERL; 1 cerebral edema and cerebellar tonsillar coning; 1 bronchial obstruction/mucus plugging; 1 bowel perforation and peritonitis; 1 lung disease of prematurity; 1 pelvic mass with lung metastases; 1 hydrocephalus; 1 pulmonary hemorrhage/consolidation )

22 ‘misses’

• 8 Bronchopneumonia

• 2 Traumatic brain injuries

• 12 ‘others’:

• 6 cardiothoracic related (1 pulmonary hemorrhage; 1 pulmonary interstitial edema

from drowning; 1 pulmonary edema from cardiac valvular

disease; 1 chronic lung disease of prematurity; 1 asthma

attack, constricted airways and aspiration; 1 hypertrophic

cardiomyopathy and aspiration)

• 4 GI related (1 gastroenteritis with splenic and hepatic infarction;

1 bowel infarction/hemorrhage; 1 intestinal malrotation; 1

metabolic disorder, abnormal fat deposition)

• 2 neurologically related (1 acute infarction; 1 motor neuron disease)

11 ‘overcalls’

• 8 Pulmonary edema/infection

• 2 GI sepsis/infarction

• 1 4th ventricle effacement

Results

Many ‘discrepancies’ were pulmonary or cardiac related

Sensitivity

(%)

Specificity

(%)

PPV

(%)

NPV

(%)

Concordance (%)

Neurological (n = 136)

75.6

[60.7, 86.2]

96.7

[90.8, 98.9]

91.2

[77.0, 97.0]

89.8

[82.2, 94.4]

90.2

[83.9, 94.2]

Cardiac (n = 136)

31.3

[14.2, 55.6]

95.8

[90.6, 98.2]

50.0

[23.7, 76.3]

91.3

[85.0, 95.1]

88.2

[81.7, 92.6]

Thoracic (n = 136)

64.7

[51.0, 76.4]

82.4

[72.9, 89.0]

68.8

[54.7, 80.1]

79.5

[70.0, 86.7]

75.7

[67.9, 82.2]

Abdominal (n = 136)

53.8

[29.1, 76.8]

88.5

[81.7, 93.0]

33.3

[17.2, 54.6]

94.7

[89.0, 97.6]

85.2

[78.2, 90.2]

Musculoskeletal (n = 136)

98.4

[91.4, 99.7]

91.8

[83.2, 96.2]

91.0

[81.8, 95.8]

98.5

[92.1, 99.7]

94.8

[89.7, 97.5]

Results – Body Systems

Pneumonia

Pneumonia

Pulmonary haemorrhage

Drowning

a c

b d

e

f

Normal

Pneumonia

1 year old - unresponsive 1 month old – unexplained death 1 month old - collapse

3 months old, unresponsive 10 month old, abuse 7 year old - unresponsive

Pulmonary Examples

c d

a b

3 year old girl.Riding a push-along

scooter on pavement home.

Man was reversing out of driveway and

did not see child.

Example Case

Discussion

• Largest reported pediatric post-mortem CT population to dateHighest concordance = intracranial & MSK pathologiesPulmonary and cardiac pathologies remain challenging

• Cause of death still remains unexplained in majority cases

• Limitations: Pathologists were not blinded to CT resultsCT findings not re-reviewed for this study

• Further research of newer post-mortem CT techniques in children are warranted.

AJR 2019; 212:1–13

If you want to read more…

Minimally Invasive Autopsy

(MIA) Research Team

Funders

CRTF (MR/R00218/1)

Inset: Prof Neil Sebire (top left), Prof Andrew Taylor (bottom left), Dr Michael Ashworth (top right), Dr Tom Jacques (bottom right)

Group image (left to right): Dr Susan Shelmerdine, Jade Parmenter, Dr Celine Lewis, Anna Guy, Lakeisha Ward, Hannah McGarrick, Wendy Norman, Rod Jones, Dr Owen Arthurs, Toby Hunt, Dr Ciaran Hutchinson, Ian Simcock, Dr Alistair Calder