Accuracy of targeted post-mortem computed tomography coronary angiography compared to assessment of serial histological sections

Abstract

Post-mortem computed tomography with coronary angiography (PMCTA) could have a role in the investigation of sudden natural death. This study assesses the accuracy of targeted coronary PMCTA, using both air and iodinated contrast media, to provide sensitivity and specificity for critical stenosis (CS) on a segmental basis, based on a gold standard of 3-5 mm serial sections of the coronary arteries using macroscopic and histological techniques. Assessment of stenosis at 1 mm intervals on PMCTA was compared with the data from pathological analysis. Stenosis was defined as "critical" when the stenotic region reaches ≥75 %. Regions were defined every 20 mm or by a clear change in stenosis. Discrepancies were defined as significant if only one test showed CS. Five cases with 25 vessels with 124 regions were assessed. PMCTA was unable to identify plaque hemorrhage or dissection (but this was normally associated with CS). Eighteen segments had significant discrepancies, giving a sensitivity and specificity of 50 and 91.5 %. When an alternative gold standard was constructed by excluding regions beyond a CS (five cases), taking PMCTA as correct where a heavily calcified vessel opens under contrast injection (four cases), and correcting for misregistration of distance (one case), the sensitivity rose to 85.7 %. There was complete agreement when the right or left coronary arteries are assessed as a whole. This study shows that PMCTA is not a perfect replacement for histological examination of coronary vessels, but may have a role in routine post-mortem investigation.