A new approach to kinematic feature extraction from the human right ventricle for classification of hypertension: a feasibility study

Abstract

This work presents a novel approach to analyze the function of the human right ventricle (RV) by deriving kinematic features of the relative change in shape throughout the cardiac cycle. The approach is anatomically consistent, allows direct comparison across populations of individuals, and potentially provides new metrics to improve the diagnosis and understanding of cardiovascular diseases such as pulmonary hypertension (PH). The details of the approach are presented, which includes a variation of harmonic topological mapping and proper orthogonal decomposition techniques, with particular focus on their applicability with respect to untagged cardiac imaging data. Results are shown for the decomposition of a collection of clinically obtained human RV endocardial surfaces segmented from cardiac computed tomography imaging into the fundamental shape change features for individuals both with and without PH. The features are shown to be consistent and converging towards intrinsically physiological components for the heart, and may potentially represent a new set of features for classifying the progressive change in RV function caused by PH, particularly in comparison to traditional clinical metrics.