Diagnostic accuracy of post-mortem CT with targeted coronary angiography versus autopsy for coroner-requested post-mortem investigations: a prospective, masked, comparison study

Abstract

Background: England and Wales have one of the highest frequencies of autopsy in the world. Implementation of post-mortem CT (PMCT), enhanced with targeted coronary angiography (PMCTA), in adults to avoid invasive autopsy would have cultural, religious, and potential economic benefits. We aimed to assess the diagnostic accuracy of PMCTA as a first-line technique in post-mortem investigations.

Methods: In this single-centre (Leicester, UK), prospective, controlled study, we selected cases of natural and non-suspicious unnatural death referred to Her Majesty's (HM) Coroners. We excluded cases younger than 18 years, known to have had a transmittable disease, or who weighed more than 125 kg. Each case was assessed by PMCTA, followed by autopsy. Pathologists were masked to the PMCTA findings, unless a potential risk was shown. The primary endpoint was the accuracy of the cause of death diagnosis from PMCTA against a gold standard of autopsy findings, modified by PMCTA findings only if additional substantially incontrovertible findings were identified.

Findings: Between Jan 20, 2010, and Sept 13, 2012, we selected 241 cases, for which PMCTA was successful in 204 (85%). Seven cases were excluded from the analysis because of procedural unmasking or no autopsy data, as were 24 cases with a clear diagnosis of traumatic death before investigation; 210 cases were included. In 40 (19%) cases, predictable toxicology or histology testing accessible by PMCT informed the result. PMCTA provided a cause of death in 193 (92%) cases. A major discrepancy with the gold standard was noted in 12 (6%) cases identified by PMCTA, and in nine (5%) cases identified by autopsy (because of specific findings on PMCTA). The frequency of autopsy and PMCTA discrepancies were not significantly different (p=0·65 for major discrepancies and p=0·21 for minor discrepancies). Cause of death given by PMCTA did not overlook clinically significant trauma, occupational lung disease, or reportable disease, and did not significantly affect the overall population data for cause of death (p≥0·31). PMCTA was better at identifying trauma and haemorrhage (p=0·008), whereas autopsy was better at identifying pulmonary thromboembolism (p=0·004).

Interpretation: For most sudden natural adult deaths investigated by HM Coroners, PMCTA could be used to avoid invasive autopsy. The gold standard of post-mortem investigations should include both PMCT and invasive autopsy.

Funding: National Institute for Health Research.

Figure 1

Images from post-mortem CT with targeted coronary angiography in a case of myocardial infarction Post-mortem CT with targeted coronary angiography (PMCTA) of a male ex-smoker aged 62 years with borderline type 2 diabetes who died suddenly and unexpectedly. PMCTA reconstructed images with straightened curved multiplane reconstructions of the positive (A) and air (B) contrast runs, and a 3D-volume reformat (C), all showing a critical stenosis of the proximal left anterior descending artery, as indicated by an asterisk. The myocardium showed an anteroseptal perfusion deficit, indicated by a dashed line (D). In the absence of artery calcification, these findings were diagnosed as a soft plaque occlusion leading to acute myocardial infarction. RCA=right coronary artery. LAD=left anterior descending. LCx=left circumflex. LV=left ventricle.

Figure 2

Images from post-mortem CT with targeted coronary angiography in three cases of haemorrhage (A) Axial brain image in a 73-year-old woman who collapsed and then had a cardiac arrest after a short interval. The autopsy report described a small amount of subarachnoid blood and normal cerebral cortex, but did not describe the cerebellum, whereas post-mortem CT with targeted coronary angiography (PMCTA) showed a clear clinically significant cerebellar haemorrhage, as indicated by an asterisk. Autopsy gave coronary artery disease as the cause of death, which, although also detected by PMCTA, was clearly incorrect. (B) Axial brain image of a 33-year-old man with type 1 diabetes and alcohol addiction. Toxicology showed evidence of clinically significant diabetic ketoacidosis, and both PMCTA and autopsy give the primary cause of death as diabetic ketoacidosis, but autopsy failed to report the clear subarachnoid haemorrhage, as indicated by an asterisk, which although not extensive enough to definitely cause death, might have substantially contributed to death. (C, D) An 84-year-old woman who was taking anticoagulation treatment with documented declining haemoglobin concentrations in the days leading to her death. She died from myocardial insufficiency secondary to hypovolaemia and anaemia agreed on both autopsy and PMCTA. However, autopsy did not find a bleeding source and attributed it to gastric erosions. PMCTA clearly showed a left scapula fracture (arrow) with approximately 1 L of blood in the left chest wall (dashed line).

Figure 3

Images from post-mortem CT with targeted coronary angiography in two cases of trauma (A, B) An 86-year-old woman who was found dead in the rear doorway to her home on a cold day in February. PMCTA agreed with autopsy on the presence of ischaemic heart disease, but autopsy failed to report the trauma, which was potentially relevant in this case. (A) 3D-bone reconstruction with anterior dislocation of the shoulder (arrow shows the direction of dislocation). (B) Coronal brain multiplanar reconstruction image with subcutaneous haematoma (*). (C, D) A 91-year-old woman with an agreed primary cause of death of myocardial insufficiency due to aortic stenosis. However, PMCTA recorded in part 2 of the death certificate (associated conditions) an (C) acute pathological fracture of the left femur (asterisk) and (D) lung metastases (arrows), not reported on autopsy, which were thought likely to have acutely exacerbated her chronic cardiac condition. PMCTA=post-mortem CT with targeted coronary angiography.