Application of Digital Analysis for Assessment of Coronary Sub-Occlusions in Autopsy Pathology: It Is Time to Move beyond Histology Alone

Abstract

Background: Coronary artery disease (CAD) underlies most cases of myocardial infarction (MI), causing or at least contributing to oxygen supply–demand mismatch and myocardial injury, so a careful and reliable evaluation of the main coronary arteries and large branches is a key moment of autopsy in order to establish the cause of death. The aim of this study is to evaluate the application of digital image analysis in the assessment of coronary artery sub-occlusions. Methods: A total of 50 coronary sections sampled during 11 consecutive autopsies, regardless of the cause of death, were analyzed. The ideal lumen and the percentage of the residual lumen were evaluated by digital pathology using QuPath v 4.3 and by an expert pathologist. The evaluations performed were compared using Lin’s concordance correlation coefficient. Results: The Lin agreement index between the two evaluation methods for all measurements showed an excellent agreement rate [0.923, with confidence interval (0.866, 0.956)]. However, in the case of critical stenosis, from 60% to 80% and from 65% to 75%, the Lin agreement index between the two evaluation methods was, respectively, 0.798 [0.603, 0.904], corresponding to good agreement, and 0.516 [0.071, 0.725], corresponding to slight agreement. The digital system has superior performance in cases where lumen occlusion falls between 60% and 80% and provides an objective assessment of the residual lumen area. Conclusions: According to the widespread availability and ease of use of these technologies, we suggest that image analysis should be considered a routine tool and established as the diagnostic gold standard in this field.