4D Flow MRI in Assessment of True Severe Low-Gradient Aortic Stenosis

Degenerative aortic stenosis is a progressive valvular heart disease affecting the aging population and when
hemodynamically significant is associated with serious adverse outcomes, such as heart failure, syncope, and
death. Although, symptomatic classical severe aortic stenosis (AS), wherein aortic valve anatomical or effective
area is ≤ 1.0 𝑐𝑚2 and mean gradient ≥ 40 mmHg raises no diagnostic dilemma, patients with low gradient aortic
stenosis (i.e., mean gradient < 40 mmHg) that may be severe (i.e., aortic valve area ≤ 1.0 𝑐𝑚2) remain a
challenge for determining the true severity of aortic stenosis and present a significant unmet clinical need. The
discordance between aortic valve area of ≤ 1.0 𝑐𝑚2 and mean gradient < 40 mmHg is encountered in as many
as 30 to 40% of patients with aortic valve area ≤ 1.0 𝑐𝑚2 by transthoracic echocardiography (TTE). In patients
with these discordant findings, the true severity is uncertain and have led to the use of dobutamine stress
echocardiography (DSE). Based on DSE, one can classify the potentially severe low-gradient subjects into 4
distinct subgroups a) Low gradient severe aortic stenosis (LGS): aortic valve area of ≤ 1.0 𝑐𝑚2 and a mean
gradient of ≥ 40 mmHg with DSE. b) Low gradient pseudo-severe aortic stenosis (LGPS): aortic valve area of >
1.0 𝑐𝑚2 and a mean gradient of < 40 mmHg with DSE. c) Low gradient indeterminate aortic stenosis severity
(LGI): aortic valve area of ≤ 1.0 𝑐𝑚2 and a mean gradient of < 40 mmHg with DSE. d) Finally, aortic valve area
of > 1.0 𝑐𝑚2 and a mean gradient of ≥ 40 mmHg with DSE are classified as moderate or less severe aortic
stenosis (LSA). Nearly one-third of patients classified as LGS are classified as LGPS with DSE (6). Further,
compounding the concern in patients classified as LGPS, LGI, LSA is that they often manifest with symptoms
potentially attributable to severe AS suggesting an initial misclassification perhaps due to measurement errors.
Compared to echo, with MRI, the velocities can be measured in all 3D directions and due to its tomographic
nature, no geometric assumptions are required. The hypothesis of the study is that our planned CMR methods
which we will develop based on previous work will be able to better stratify these AS subjects. Our specific aims
are: 1) We will develop efficient rest and dobutamine stress 4D Spiral flow imaging protocols based on k-space
dependent respiratory gating. 2) We have recently developed a Deep Learning framework which based on
Computational Fluid Dynamics (CFD) simulations of training data learns to directly map velocities to pressures.
This will be adapted to measure the transvalvular pressure gradients (TVPG) in human subjects and validated.
3) In 40 subjects with potentially severe low gradient AS (10 in each of the LGS, LGPS, LGI, LSA), a TTE study,
and a CMR study will be performed both at rest and under dobutamine stress. To validate, TVPG will be
measured with cath in a group with n=10 of moderate AS (MAS) subjects undergoing cath for other purposes.
This group will also undergo TTE and CMR studies. Subjects with severe AS were not selected for cath for safety
reasons. Classification across modalities, flow, velocity, pressure, and orifice area will be statistically correlated.

Grant Number: 5R21AG080859-02
NIH Institute/Center: NIH
Principal Investigator: AMIR AMINI