Supervised learning approach for tracking quality determination of transmural and segmental time strain curves: A feasibility study

Abstract

Quantitative assessment of global and regional left ventricle function by means of myocardial strain estimation has been widely discussed as promising clinical diagnostic markers of left ventricular malfunction. These markers are provided to the clinicians without much feedback regarding their reliability, which may lead to erroneous diagnosis. Therefore, this study aims to classify the calculated strain curves into reliable or artefactual ones, before their clinical adaptation.

A supervised machine learning approach is utilized for the classification process. A total of 6,552 strain curves were used, for which a visual labeling protocol was defined and utilized by two experts.

An inter-observer labeling concordance of 93% was obtained, and a classification accuracy of 90% was achieved with a specificity of 92% and sensitivity of 78%.

This classification tool may enhance the reliability of the estimations of global, transmural and regional strain curves, by automatically classifying them into physiological or artefactual curves.