loading
loading
NIH grant 1R01HL182092-01A1: Developing 3D Real-Time CMR for Arrhythmia Imaging
Institute: {'code': 'HL', 'abbreviation': 'NHLBI', 'name': 'National Heart Lung and Blood Institute', 'admin_org_id': '12500068', 'admin_funding_url': ''}. PI: Rizwan Ahmad. Award: $729984.
Start: 2026-07-22T00:00:00.
PROJECT SUMMARY
Cardiac arrhythmias affect 1.5 to 5% of the general population. Due to the high prevalence of cardiac
comorbidities in this group, patients with arrhythmia constitute a significant fraction of those undergoing cardiac
imaging. Among widely available modalities, cardiovascular magnetic resonance imaging (CMR) offers the most
comprehensive assessment of cardiac structure, function, and tissue characterization. CMR is particularly well
suited to detect myocardial abnormalities, which are common in patients with arrhythmias. However, imaging
this population using slice-by-slice two-dimensional (2D) breath-held segmented CMR results in motion artifacts.
Two-dimensional real-time (RT) CMR overcomes these artifacts but poses challenges for functional analysis as
isolating and analyzing matching beats across slices is challenging in patients with frequent ectopy or when the
goal is to assess the arrhythmic beats themselves. Three-dimensional (3D) CMR addresses some of these
limitations, but current techniques rely on gating or binning strategies that collapse beat-to-beat variation and
degrade image quality in irregular rhythms. As a result, existing 3D methods often fail to generate diagnostic-
quality images in patients with frequent arrhythmias. Three-dimensional RT CMR could overcome the limitations
of both 2D RT CMR and conventional 3D CMR, but it has remained out of reach due to the extremely high
acceleration rates required.
To address this need, we propose a novel unsupervised reconstruction framework called generative prior with
multi-dynamic modeling (GeMM) to enable highly accelerated 3D RT CMR from a free-running scan. While
broadly applicable, 3D RT imaging is particularly advantageous in patients with frequent premature ventricular
contractions (PVCs). First, it is inherently robust to cardiac and respiratory motion irregularities. Second, it
eliminates the need to identify and align sinus beats across slices for functional assessment. Third, it is
insensitive to through-plane motion, allowing reliable computation of segmental metrics such as regional strain
and dyssynchrony, which often change before left ventricular ejection fraction declines. Fourth, it enables
morphology-specific analysis of sinus and arrhythmic beats, supporting a more comprehensive assessment of
mechanical impairment.
The project includes three specific aims: (1) develop, optimize, and validate GeMM for 3D RT cine and late
gadolinium enhancement imaging using data from 3D digital cardiac phantoms; (2) validate GeMM in healthy
subjects and clinical patients without arrhythmia; and (3) validate GeMM in PVC patients to quantify fibrosis as
well as beat-specific left ventricular mechanics and dyssynchrony. These studies will demonstrate the feasibility
and added value of 3D RT imaging in patient groups where conventional 3D CMR techniques routinely fail to
provide diagnostic-quality images.document fetched 2026-08-02 08:24