Visual analysis of regional myocardial motion anomalies in longitudinal studies. (October 2019)

• Record Type: Journal Article
• Authors: Sheharyar, Ali; Ruh, Alexander; Aristova, Maria; Scott, Michael; Jarvis, Kelly; Elbaz, Mohammed; Dolan, Ryan; Schnell, Susanne; Lin, Kai; Carr, James; Markl, Michael; Bouhali, Othmane; Linsen, Lars
• Is Part Of: Computers & graphics. Volume 83(2019)
• Page Start: 62
• Page End: 76
• Publication Date: 2019-10
• Subjects: Biomedical visualization -- Myocardial motion -- Heart transplant; Computer graphics -- Periodicals; 006.6
• DOI: 10.1016/j.cag.2019.07.004
• Languages: English
• ISSNs: 0097-8493

Abstract

Highlights: Early detection of regional anomalies in myocardial motion for patients after the heart-transplant operation. Visualization of the variability within a cohort using the concept of functional boxplots. Visual encoding of the spatio-temporal location and severeness of regional anomalies in an individual dataset by comparing it to the healthy cohort. Visualization of the evolution of the anomalies within a sequence of scans in a longitudinal setting. Graphical abstract: Abstract: A heart-transplanted patient is at risk of developing several complications such as rejection, which is one of the leading causes of deaths in the first year after the transplant. The regional myocardial motion is known to be depressed early on during rejection before the reduction in global systolic function. Therefore, early detection of regional anomalies is crucial. We use a magnetic resonance (MR) imaging method called tissue phase mapping (TPM) to capture regional myocardial motion of heart-transplanted patients in a longitudinal study. We compare the individual scans of the longitudinal study to a cohort of healthy volunteers to detect anomalies. We use a spatio-temporal visualization based on a radial layout where myocardial regions are laid out in an angular pattern similar to the American Heart Association (AHA) model and where the temporal dimension increases with increasing radius. We compute nested envelopes of central regions for the time series of each region and each of the three velocity directions using the concept of functional boxplots. We propose visual encodings to analyze regional anomalies of a scan of an individual patient dataset and perform a qualitative user study with medical experts. We extend this layout to the visual analysis of longitudinal data to monitor changes in regional anomalies of a patient for multiple scans taken at different times. We apply our approach to data from a longitudinal study of patients under observation after a heart-transplant procedure and evaluate this mechanism with medical and non-medical experts.