5 - 4D Flow vCath: A novel virtual catheter technique for assessing hemodynamics in aortic valve disease

Friday, February 8

1:59 PM - 2:10 PM

Location: Grand AB

Presenter(s)

Mohammed S.M. Elbaz, PhD

Postodoctoral Fellow
Northwestern University Feinberg School of Medicine

Author Only(s)

Michael Scott, MSc, BSc

Graduate Student Researcher
Northwestern Feinberg School of Medicine Department of Radiology
AB

Alex Barker, PhD

Associate Professor
University of Colorado Denver - Anschutz Medical Campus
PM

Patrick M McCarthy, MD

Professor
Northwester University Feinberg School of Medicine
SM

S. Chris Malaisrie, MD

Associate Professor of Surgery
Northwestern University
JC

Jeremy D. Collins, MD

Associate Professor
Mayo Clinic, Rochester, MN, USA
RB

Robert O. Bonow, MD

Professor
Northwestern University Feinberg School of Medicine
James Carr, MD

Knight Family Professor of Cardiac Imaging
Northwestern University
MM

Michael Markl, PhD

Lester B. and Frances T. Knight Professor of Cardiac Imaging
Northwestern University

Background:

4D Flow CMR has emerged as a useful technique for evaluating altered aortic blood flow in patients with aortic valve disease such as congenital Bicuspid Aortic Valve (BAV). However, current 4D flow analysis frameworks are limited by 1) cumbersome and manual analysis, 2) adoption of a multi 2D plane analysis approach not fully utilizing the 4D nature (3D + time) of the data, 3) challenging time-varying aortic segmentation for the quantification of flow metrics across time. To address these limitations, we propose a novel 3D virtual catheter (vCath) method that mathematically mimics the invasive catheter technique but for the noninvasive and automated assessment of several advanced metrics of aortic hemodynamics from 4D Flow CMR.

Methods:

A total of 106 4D Flow CMR scans were analyzed from 87 subjects (57 BAV patients [age: 46±12yrs], 30 healthy controls [age: 44±13yrs, p=0.45]) retrospectively enrolled from an IRB-approved study. To test the reproducibility, 15 healthy subjects underwent test-retest 4D Flow scanning with 2 weeks between the two scans. Aortic 4D Flow CMR was performed at 1.5T (Siemens) during free breathing with respiratory navigator gating (spatial resolution = 2-3mm³, temporal resolution = 33–43ms, venc = 150–450 cm/s).
From a 3D segmentation of the thoracic aorta, vCath construction was fully automated (Fig 1). For each time point in the cardiac cycle, 3D hemodynamic kinetic energy (KE), viscous energy loss rate (EL) and vorticity were quantified from 4D Flow only over the vCath volume [3]. Systolic vCath peak KE (KE_{vcath_peak}), peak EL (EL_{vcath_peak}) and Vorticity_vol_vCath-peak were computed and normalized by the vCath volume (Fig2).

Results: In all 106 4D flow scans, automated aortic centerline detection was successful. Only in 10% of the cases, minor manual correction was needed to exclude supra-aortic branches. As Fig. 2 shows, BAV patients had a significantly larger vCath radius and volume than controls (p<0.001). While no significant differences were found in systolic KE_{vcath_peak}between controls and BAV patients (p=0.63), systolic EL_{vcath_peak} and Vorticity_vol_vCath-peak were significantly elevated in BAV patients (p<0.001). Good test-retest reproducibility was found for all vCath-derived hemodynamic metrics (Fig. 3).

Conclusion: This study demonstrated the feasibility and reproducibility of a novel automated 3D virtual catheter (vCath) approach for quantifying alterations in advanced 3D time-resolved aortic hemodynamics from 4D Flow CMR in BAV patients. Notably, vCath modeling automatically adjusted to each patient’s aortic shape and ensured adequate vCath placement in the vessel center throughout the cardiac cycle, enabling robust time-resolved assessment even from a static aortic segmentation. Hence, the vCath technique might facilitate better clinical translation of the powerful but currently complex 4D Flow CMR framework to assess advanced hemodynamics in large cohorts of aortic disease patients.