A feasibility study on kinematic feature extraction from the human interventricular septum toward hypertension classification

Abstract

The current work is intended to present the first steps in examining the feasibility of improving the diagnosis and understanding of cardiovascular diseases, particularly pulmonary hypertension (PH), by analyzing the shape of the right side of the interventricular septum. Although there clearly exists a connection between the septum shape and PH, there is still no quantitative and clear understanding of how this shape change relates to the disease state. Therefore, there exists a great need for techniques that are capable of quantitatively analyzing variations in septum shape and to perform a study that analyzes these shape changes relating to various states of PH. Such advancements could directly improve the diagnostic process for this deadly disease, including early detection overcoming the early nonspecific symptoms of PH and better understanding the progression of the disease by identifying behaviors that relate to either beneficial or deleterious cardiac adaptation. The details of a potential shape analysis approach for the interventricular septum is presented, including a surface parameterization approach and direct application of proper orthogonal decomposition for feature extraction. To show the feasibility of implementing this approach, the decomposition of a single patient's interventricular septum at several phases in a single cardiac cycle is shown. The results show the capability to obtain distinct features of the septal wall shape in a generally applicable way, and provide a path for future large-scale analyses of groups of patients. © 2014 Springer International Publishing.