4D Flow MRI, Cardiac Function, and T1-Mapping: Association of Valve Mediated Changes in Aortic Hemodynamics with Left Ventricular Remodeling

Abstract

BACKGROUND

Patients with bicuspid aortic valve (BAV) show altered hemodynamics in the ascending aorta that can be assessed by 4D flow MRI

PURPOSE

Comprehensive cardiac MRI was applied to test the hypothesis that BAV mediated changes in aortic hemodynamics (wall shear stress (WSS) and velocity) are associated with parameters of left ventricular (LV) remodeling.

STUDY TYPE

retrospective data analysis

POPULATON

49 BAV patients (mean age = 50.2±13.5, 62% male).

FIELD STRENGTH/SEQUENCE

Balanced steady-state free precession (bSSFP)-CINE, pre- and post-contrast T1 mapping with modified Look-Locker inversion recovery (MOLLI), time-resolved 3D phase-contrast (PC) MRI with three-directional velocity encoding (4D flow MRI) at 1.5 and 3T

ASSESSMENT

Quantification of LV volumetric data and myocardial mass, extracellular volume fraction (ECV), aortic valve stenosis (AS) and regurgitation (AR). 3D aortic segmentation, quantification of peak systolic velocities and 3D WSS in the ascending aorta (AAo), arch, and descending aorta (DAo).

STATISTICAL TESTS

Two-sided non-paired t-test to compare subgroups. Pearson correlation coefficient for correlations between aortic hemodynamics and LV parameters.

RESULTS

Of the 49 BAV patients, 35 had aortic valve dysfunction (AS (n=7), AR (n=16), both AS and AR (n=12)). Mean systolic WSS in the AAo, peak systolic velocities in the AAo and arch, and LV mass were significantly higher (p<0.001) in the AS/AR group compared to the patients without AS/AR. In the complete group, we observed significant relationships between increased LV mass and elevated peak systolic velocity (r=0.57, r=0.58; p<0.001) and WSS in the AAo and arch, respectively (r=0.54, r=0.46; p<0.001). We detected an association between ECV and WSS in the AAo (r=0.38, p=0.02). These relations did not held true for patients without AV dysfunction.

DATA CONCLUSION

AS and AR in BAV patients have a major impact on elevated aortic peak velocities and WSS which were associated with parameters of LV remodeling.

Introduction

Bicuspid aortic valve (BAV) is the most common congenital cardiovascular malformation in humans, affects 1–2% of the population, and frequently results in concomitant stiffening and aneurysms of the ascending aorta (AAo) with increasing risk of dissection or rupture (1, 2). The majority of patients, although often asymptomatic for many years, need regular imaging follow-up for standardized assessment of the AAo diameter to prevent risk of complications with aortic growth and for prophylactic surgical or interventional planning (3, 4). Common valvular complications associated with BAV with increasing age are aortic valve stenosis (AS) or aortic regurgitation (AR) which affect both aortic hemodynamics and cardiac function by increased loading of the left ventricle (LV) (1). Previous cardiovascular Magnetic Resonance studies have shown that BAV morphology is associated with altered aortic hemodynamics in terms of altered 3D blood flow patterns, increased peak velocities and eccentrically elevated wall shear stress (WSS) in the AAo (5, 6). Even patients with normally functioning BAV (i.e. no AS/AR) frequently develop changes in aortic hemodynamics (5, 7). Specifically, studies have shown that elevated peak velocities and systolic wall shear stress (WSS) are associated with development of BAV aortopathy (8) and aortic wall degeneration (9). However, the impact of these aortic changes on LV loading and subsequent risk for LV remodeling and thus impaired cardiac function is poorly understood.

Recent advances in MR allow for the simultaneous assessment of cardiac structure and function by CINE-imaging and T1-mapping. In addition, 4D flow MRI enables a comprehensive assessment of vascular hemodynamics, including the quantification of 3D WSS along the vessel lumen for the detection of pathological WSS (7, 10). A first effort was recently made to investigate relationships between blood flow patterns in the AAo and LV remodeling in patients with AS (11). This study, however, did not use pre- and post-contrast T1-mapping techniques for calculation of the Gadolinium extracellular volume fraction (ECV) which can help to identify diffuse structural myocardial changes such as interstitial fibrosis (12, 13).

The aim of this pilot feasibility study was therefore to use both 4D flow MRI, pre- and post-contrast T1-mapping, in addition to standard-of-care cardiovascular assessment for the analysis of potential associations between abnormal aortic hemodynamics (WSS and peak systolic velocities) and parameters of LV remodeling (volumes, mass and ECV). We hypothesized that there is a relationship between alterations in aortic hemodynamics and evidence of LV remodeling in BAV patients.

Material and Methods

Study Cohort

Fifty-two patients with BAV were retrospectively identified in a chart review from a larger group of BAV patients who underwent 4D flow MRI and T1-mapping as part of their cardiovascular MR assessment between January and August 2016. All patients were included in accordance with an Institutional Review Board (IRB) approval and waiver of consent. We initially excluded patients with diagnosed myocarditis or myocardial infarction from the study. Two patients were excluded during the MRI assessment due to associated repaired aortic coarctation. In addition, we excluded one patient after the first assessment because he proved to be an outlier in the group with no AV pathology. Forty-nine BAV patients were included in the final study cohort (age = 50.2±13.5 years, median 54 years, range 22–70 years, 62% male (see table 1). None of the patients met criteria of hypertrophic cardiomyopathy. Patients with a high AAo flow velocities above 6m/s and occurrence of aliasing were excluded a priori, as well. Hematocrit and creatinine was measured before the CMR, and all the patients had GFR>60 using the previously described equation (14). Patient hematocrit was collected within 48 hours of the cardiac MRI exam.

Table 1.

Patient demographics of the complete BAV cohort, patients with and without aortic valve (AV) dysfunction

Patient Characteristics
	Complete cohort	AV dysfunction	no AV dysfunction
BAV type (Sievers)	40 type I - RL, 6 type I - RN, 2 type 0, 1 type 2	28 type I – RL, 6 type I – RN, 1 type 2	12 type I – RL, 2 type 0
age (years)	50.2 ± 13.5	52.0 ± 13.8	45.8 ± 12.1
Gender (male/female)	30/19	22/13	8/6
height (cm)	175.6 ± 10.7	175.4 ± 10.9	176.0 ± 10.7
weight (kg)	79.8 ± 13.9	81.2 ± 13.6	76.6 ± 14.7
body surface area (m2)	2.0 ± 0.2	2.0 ± 0.2	1.9 ± 0.2
heart rate (beats/min)	65 ± 9	64 ± 10	67 ± 8

MRI protocol

All patients underwent a standard-of-care cardiac MRI exam on a 1.5T (36 patients) or 3T (13 patients) MRI system (Magnetom Avanto/Aera, Skyra, Siemens Medical Systems, Erlangen, Germany) including electrocardiogram (ECG) gated time resolved (CINE) 2D SSFP MRI in the short axis (full coverage from base to apex) for the evaluation of cardiac dimensions and global function parameters (figure 1a). 2D-CINE phase contrast (PC) sequences were performed at the level of the AV to assess peak flow velocities and to quantify regurgitant fraction. Contrast-enhanced MR angiography was acquired after intravenous contrast agent administration (bolus infusion, dose = 0.2 mmol/kg, Gadobutrol, Gadavist, Bayer Pharmaceuticals, Whippany, NJ).

Fig. 1.

MRI assessment to test associations between altered aortic flow and changes in LV structure and function in BAV. A) Assessment of left ventricular function and mass by 2D-CINE-SSFP, b) and c) pre- and post-contrast T1-mapping for calculation of extracellular volume fraction (ECV). D) 3D segmentation of the aorta based on 4D flow MRI for e) calculation of regional peak velocities shown as maximum intensity projection (MIP) and f) 3D wall shear stress (WSS) for the AAo, arch, and DAo. LV=left ventricular, AAo=ascending aorta, DAo=descending aorta.

T1-mapping was performed using a modified Look-Locker inversion recovery (MOLLI) technique, utilizing a 5(3)3 – two T1 experiments pre- and 4(1)3(1)2 post-contrast technique (15). Data for each slice (base, mid, apex) were acquired during breath holding pre- and 10–25 minutes following intravenous contrast agent administration. Imaging reconstruction included inline motion correction of the MOLLI images with different inversion times, and the calculation of parametric LV T1 maps (see figure 1b, c) as described previously (16). T1 mapping parameters were as follows: spatial resolution (pixel size) = 2.3 × 1.8mm, slice thickness = 5mm, TE/TR = 1.0–1.1msec/2.0–2.2msec; flip angle = 35°.

For the assessment of aortic hemodynamics, time-resolved 3D phase-contrast (PC) MRI with three-directional velocity encoding (4D flow MRI) was performed to measure 3D blood flow velocities with full volumetric coverage of the thoracic aorta. 4D flow MRI was performed approximately 25–35 minutes post contrast at the end of the MR study after T1-mapping and delayed enhancement imaging. 4D flow MRI was acquired during free breathing using respiratory and prospective ECG gating with a target efficiency of 80% as described previously leading to a scan time of about 8–12 min (17). Pulse sequence parameters were as follows: spatial resolution = 2.1–4.0×2.1–4.0×2.5–3.2mm³; temporal resolution = 37–40ms; field of view = 255–340 × 255–360mm², velocity sensitivity (venc) = 150–400cm/s; echo time (TE) = 2.2–2.5ms; repetition time (TR) 4.6–4.9ms; flip angle = 15°. The venc adjustment was made based on expected maximum velocities from 2D PC measurements. All 4D flow MRI scans were acquired with parallel imaging (GRAPPA) with a reduction factor of R = 2.

Data analysis: LV function and structure

Left ventricular end-diastolic and end-systolic volumes (LVEDV and LVESV), stroke volumes (SV) and ventricular mass were assessed for each patient using Syngo Argus (Siemens Medical Solutions, Erlangen, Germany) and normalized to the patient’s body surface area (BSA). The papillary muscles were not included in the myocardial mass calculation. Ejection fraction (EF) was automatically calculated. Epi- and endocardial LV contours (base, mid, apex) were manually delineated in the pre- and post-contrast T1-maps using Circle (CMR42, Circle Cardiovascular Imaging, Calgary, Canada). Myocardial ECV was automatically calculated using the formula $\text{ECV}=\frac{\mathrm{T}1\text{post}-\mathrm{T}1\text{pre blood}}{\mathrm{T}1\text{post}-\mathrm{T}1\text{pre myocardium}}\times \left(1-\text{hematocrit}\right)$(18). Basal, mid, and apical locations were averaged to obtain the global LV ECV.

Data analysis: aortic valve morphology

SSFP-CINE-images were evaluated by cardiovascular radiologists (JDS, JCC) with 15 and 20 years of experience with regard to AV morphology and Sievers type (19). Peak flow velocities and flow volumes for AS and AR assessment, and aortic valve orifice area were quantified by manual segmentation in 2D-CINE-PC images (Syngo Argus, Siemens Medical Solutions, Erlangen, Germany). AS and AR severity were graded on a 3-point scale (mild, moderate, severe) based of the orifice area and peak velocity or retrograde flow volume respectively according to international guidelines (20). Since it was not the aim of this study to investigate correlations between the degree of AV pathology and aortic hemodynamics, we defined all subgroups in general, either with stenosis or with regurgitation, as pathology, i.e. as additional AV dysfunction.

Data analysis: AAo diameter, aortic peak systolic velocities and 3D WSS

Cross-sectional aortic diameter was assessed on multiplanar reformatted MRA images in the mid AAo at the level of the midpoint between sinotubular junction and origin of the brachiocephalic trunk in accordance with international guidelines (21). The diameters were calculated from an average of two orthogonal measurements. The presence of AAo dilatation was defined as a diameter of the AAo > 4.0cm (21).

4D flow data preprocessing included noise filtering and correction for eddy currents, Maxwell terms, and velocity aliasing, as described previously (22). 3D PC MR-angiogram (MRA) images, weighted for systolic timeframes, were derived from 4D flow data by multiplication of the PC magnitude images with the absolute velocity images. The 3D PC-MRA images were subsequently used to semi-automatically segment the thoracic aorta using a commercial software package as illustrated in figure 1d (MIMICS, Materialise, Leuven, Belgium). The aorta segmentation was used to mask the 4D flow velocity data and to calculate a maximum intensity projection (MIP) of the peak systolic absolute velocities in sagittal orientation as shown in figure 1e. This method has recently been introduced and has been proven valuable for fast and improved detection of regional peak systolic velocities frequently found in BAV patients (23). Regions of interest in the peak velocity MIP covering the AAo, arch and DAo were used to automatically extract maximum velocities in all three aortic segments (23). In addition, peak systolic 3D WSS was calculated along the entire aortic wall using the same approach as described previously (see figure 1f) (10, 24). Regional mean peak systolic WSS values were obtained by segmentation of the thoracic aorta in three defined segments for the AAo, arch and DAo, analogous to the regional MIP assessment.

Statistical analysis

The assumption of normal distribution was assessed for each parameter using the Kolmogorov-Smirnov-Test. Quantitative values between different subgroups of the BAV cohort were compared using a two-sided non-paired t-test. Continuous data were expressed as mean ± SD. Correlations between aortic hemodynamics parameters (peak systolic velocity and WSS in the AAo and arch) and LV parameters were assessed using the Pearson correlation coefficient. Statistical significance was indicated by a p-value of <0.05. Statistical analysis was performed using statistical software SPSS (IBM, Armonk, New York).

Results

Aortic valve

According to the Sievers classification (19), 40 patients had a type I - RL BAV, 6 type I - RN, 2 type 0 and one type 2. The mean aortic valve area was 3.5±1.8 cm². After dividing the complete BAV cohort into patients with and without additional AV dysfunction, 35/49 patients had AV dysfunction in terms of AS (7 patients), AR (n=16) or both AS and AR (n=12). Complete details regarding the study cohort, AV morphology and cardiac function are found in tables 1 and 2.

Table 2.

Aorta and valve parameters, cardiac function and ECV for all BAV patients, patients with and without aortic valve (AV) dysfunction

Aorta & Valve Parameters
	Complete cohort	AV dysfunction	no AV dysfunction
Aortic valve stenosis	none 30, mild 5, moderate-severe 14	mild 5, moderate-severe 14	30
Aortic valve area (cm2)	3.50 ± 1.84	3.05 ± 1.90	4.59 ± 1.15
Aortic valve regurgitation	none 21, mild 16, moderate-severe 12	mild 16, moderate-severe 12	21
Diameter in mid AAo (cm)	4.0 ± 0.8	4.2 ± 0.7	3.7 ± 0.8
Diameter in arch (cm)	2.9 ± 0.4	3.0 ± 0.4	2.7 ± 0.3
Peak systolic velocity (m/s)
AAo	2.54 ± 1.09	2.87 ± 1.12	1.70 ± 0.22
arch	1.36 ± 0.43	1.47 ± 0.45	1.09 ± 0.21
DAo	1.23 ± 0.29	1.25 ± 0.31	1.18 ± 0.23
Peak systolic WSS (N/m2)
AAo	0.73 ± 0.21	0.78 ± 0.22	0.60 ± 0.10
arch	0.65 ± 0.16	0.66 ± 0.17	0.63 ± 0.15
DAo	0.69 ± 0.17	0.69 ± 0.17	0.69 ± 0.19
Cardiac Function and ECV
ECV (%)	25.9 ± 2.8	26.2 ± 2.4	25.2 ± 3.7
EDVI (ml/m2)	78.4 ± 27.9	81.6 ± 31.2	70.7 ± 16.7
ESVI (ml/m2)	30.0 ± 12.9	31.5 ± 14.4	26.7 ± 7.9
SVI (ml/m2)	48.3 ± 16.8	50.1 ± 18.9	44.1 ± 9.6
Myocardial mass index (g/m2)	74.3 ± 20.1	80.1 ± 21.0	60.6 ± 8.0
EF (%)	62.1 ± 5.6	61.9 ± 6.2	62.6 ± 3.9

AAo = ascending aorta, DAo = descending aorta, WSS = wall shear stress, ECV = extracellular volume fraction, EDVI = end-diastolic volume indexed to BSA, EDVI = end-systolic volume indexed to BSA, SVI = stroke volume indexed to BSA, EF = ejection fraction.

Cardiac assessment

As shown in tables 2 and 3, cardiac analysis resulted in normal systolic cardiac function, expressed by a mean EF of 62±6% (median 62%, 43–71%). Only three patients had an EF below 55%. LV dimensions (indexed EDV and ESV values) were within normal ranges, mean myocardial mass index of the complete cohort was 74±20g/m² (median 68g/m², 48–129g/m²), male: 82±20g/m², female: 61±20g/m². Mean ECV was not elevated (25.9±2.8%, median 26.0%, 20.9–34.3%). ECV of the basal cardiac segments was nearly the same (25.5±3.1%, median 25.0%, 20.3–34.9%).

Table 3.

Patient demographics, hemodynamic parameters, cardiac function and ECV for BAV patients for all subgroups

Patient subcohorts	AS/AR				no AS/AR	p-values
		AS	AR	AS + AR
number	35	7	16	12	14
age (years)	52.0 ± 13.8	59.6 ± 10.8	50.1 ± 12.5	50.1 ± 16.3	45.8 ± 12.1	* 0.13; † 0.02; ‡ 0.35; § 0.46
Gender (male/female)	22/13	3/4	8/8	14/1	8/6
diameter mid AAo (cm)	4.2 ± 0.7	4.1 ± 0.6	4.0 ± 0.8	4.4 ± 0.6	3.7 ± 0.8	* 0.06; † 0.22; ‡ 0.30; § 0.015
Hemodynamic Parameters
Peak systolic velocity (m/s)
AAo	2.87 ± 1.12	3.21 ± 0.56	1.95 ± 0.36	3.91 ± 1.02	1.70 ± 0.22	* <0.001; † <0.001; ‡ 0.03; § <0.001
arch	1.47 ± 0.45	1.51 ± 0.51	1.25 ± 0.33	1.73 ± 0.45	1.09 ± 0.21	* <0.001; † 0.07; ‡ 0.13; § <0.001
DAo	1.25 ± 0.31	1.13 ± 0.34	1.24 ± 0.30	1.32 ± 0.31	1.18 ± 0.23	* 0.74; † 0.77; ‡ 0.54; § 0.19
Peak systolic WSS (N/m2)
AAo	0.78 ± 0.22	0.83 ± 0.13	0.66 ± 0.16	0.92 ± 0.24	0.60 ± 0.10	* <0.001; † 0.003; ‡ 0.28; § <0.001
arch	0.66 ± 0.17	0.62 ± 0.23	0.63 ± 0.17	0.73 ± 0.13	0.63 ± 0.15	* 0.50; † 0.92; ‡ 0.98; § 0.09
DAo	0.69 ± 0.17	0.66 ± 0.20	0.70 ± 0.19	0.70 ± 0.14	0.69 ± 0.19	* 0.98; † 0.72; ‡ 0.89; § 0.86
Cardiac Function and ECV
ECV (%)	26.2 ± 2.4	26.2 ± 2.0	26.1 ± 2.3	26.3 ± 3.0	25.2 ± 3.7	* 0.58; † 0.44; ‡ 0.45; § 0.42
EDVI (ml/m2)	81.6 ± 31.2	65.1 ± 12.9	89.9 ± 38.6	80.7 ± 25.4	70.7 ± 16.7	* 0.13; † 0.41; ‡ 0.09; § 0.27
ESVI (ml/m2)	31.5 ± 14.4	21.9 ± 4.4	35.2 ± 15.6	32.6 ± 14.8	26.7 ± 7.9	* 0.15; † 0.09; ‡ 0.07; § 0.25
SVI (ml/m2)	50.1 ± 18.9	43.3 ± 9.5	54.8 ± 24.1	48.1 ± 14.4	44.1 ± 9.6	*0 .16; † 0.85; ‡ 0.19; § 0.44
LVmass index (g/m2)	80.1 ± 21.0	67.3 ± 10.5	78.6 ± 24.3	90.4 ± 16.8	60.6 ± 8.0	* <0.001; † 0.17; ‡ 0.01; § <0.001
EF (%)	61.9 ± 6.2	66.2 ± 4.5	59.6 ± 6.4	60.5 ± 7.9	62.6 ± 3.9	* 0.69; † 0.09; ‡ 0.13; § 0.43

AS=aortic valve stenosis, AR= aortic valve regurgitation, AAo=ascending aorta, DAo=descending aorta, EDVI=end-diastolic volume indexed to BSA, ESVI=end-systolic volume indexed to BSA, SVI=stroke volume indexed to BSA, LV=left ventricular, EF=ejection fraction.

^*Comparison AS/AR and no AS/AR,

^†comparison AS and no AS/AR,

^‡comparison AR and no AS/AR,

^§comparison AS+AR and no AS/AR; significant differences are printed in bold and italics.

We detected significantly increased LV mass (p<0.001) in the AS/AR group compared to the patients with normal AV function. After further subdividing the AS/AR group into the subgroups AS, AR and AS+AR, LV mass was larger in the AR (p=0.01) and AS+AR group (p<0.001) but not in the AS group (p=0.17) in comparison with the no AS/AR group. Global and basal ECV were not significantly elevated in the subgroup with AV dysfunction compared to the group without AV dysfunction (26.2±2.4% vs. 25.2±3.7%; 25.8±3.0% vs. 24.6±3.1%). However, when normalizing ECV to the muscle mass, we obtained significant differences between the patients with and without AV dysfunction (0.35±0.09 vs. 0.43±0.11, p=0.02).

Aortic dimensions, velocity and WSS

Results of the assessment of aortic dimension, peak velocity and WSS are summarized in table 3. The mid AAo diameter was enlarged (>4cm) in 25/49 patients, the mean diameter was 4.0±0.8cm, (median 4.1) with a range between 2.6cm and 5.4cm. Mean AAo diameters were similar in BAV patients without AV dysfunction compared to the group with additional AS/AR (3.7±0.8cm vs. 4.2±0.7cm; p=0.06), whereas mean aortic arch diameters were significantly smaller (2.7±0.3cm vs. 3.0±0.4 cm; p=0.01).

We detected significantly increased peak velocities in the AAo (p<0.001) and arch (p<0.001), peak systolic WSS in the AAo (p<0.001) in the group with AV dysfunction compared to the no AS/AR group.

Mean peak systolic velocities in the AAo were 2.9±1.1m/s (median 2.5m/s, 1.4–5.5m/s) in patients with AV dysfunction vs. 1.7±0.2m/s (median 1.7m/s, 1.3–2.1m/s) in the no AV dysfunction group. Mean peak systolic WSS in the AAo was 0.78±0.22 N/m² (median 0.72 N/m², 0.33–1.52 N/m²) in the AV dysfunction BAV subgroup compared to 0.6±0.10 N/m² (median 0.57 N/m², 0.47–0.82 N/m²) in those without AS/AR.

Significantly increased peak systolic AAo velocities (AS: p<0.001; AS+AR: p<0.001) and AAo WSS (AS: p=0.003; AS+AR: p<0.001) were detected for the AS and AS+AR group and elevated peak systolic velocity in the arch of the AS+AR cohort (p<0.001).

Examples for comprehensive cardiovascular assessment in patients with and without AV dysfunction are shown in figures 2 and 3. Figure 2 illustrates representative findings in a BAV patient with stenotic type I - RL BAV. Figure 3 depicts a side-by-side comparison of two patients with stenotic type 2 and non-stenotic type 0 BAV to demonstrate the cardiac and aortic results depending on AV dysfunction.

Fig. 2.

Comprehensive CMR assessment in a 68 yo female patient with type I - RL BAV showing mild LV hypertrophy in the 2D-CINE-SSFP image (a). b) Systolic peak velocity MIP reveals elevated velocities in the AAo. c) 3D aortic segmentation depicts the AAo dilatation (4.3cm). d) The ECV map shows ECV within normal limits (27%). e) 3D WSS depicts elevated WSS in the AAo. f) The 2D-CINE-SSFP image at the level of the aortic valve shows restricted valve opening due to severe stenosis (marked by arrows). LV=left ventricle, AAo=ascending aorta, DAo=descending aorta, MIP=maximum intensity projection, ECV=extracellular volume.

Fig. 3.

Side by side comparison of two patients with type 2 (with AV dysfunction) and 0 (without AV dysfunction) BAV. Top: 47 yo male patient with stenotic type 2 BAV: LV hypertrophy is depicted in the 2D-CINE-SSFP short axis image. Elevated AAo peak systolic velocity (white arrow) is displayed in the velocity MIP image. 3D WSS calculation demonstrates high systolic WSS, mainly in the distal outer AAo segment (black arrow, posterior view). The 2D-CINE-SSFP image depicts decreased aortic valve area (1.3cm²). Bottom: 34 yo male patient with type 0 BAV without additional AS or AR. The 2D-CINE-SSFP image shows normal LV mass. Velocity MIP and 3D WSS show regional peak systolic velocities and WSS within normal ranges. LV=left ventricle, AAo=ascending aorta, DAo=descending aorta, MIP=maximum intensity projection.

Ventricular-aortic associations

Ventricular-aortic analysis in the complete study cohort revealed significant relationships between increased LV mass and elevated peak systolic velocity (r=0.57, p<0.001) and increased WSS in the AAo (r=0.54, p<0.001) (Fig. 4 and 5 a), and between LV mass and elevated peak systolic velocity and WSS in the aortic arch (r=0.58, p<0.001, r=0.46, p<0.001).

Fig. 4.

Correlation plots for relationships between increased LV mass and elevated systolic peak velocity in the AAo for the complete cohort; the subcohorts are depicted with different symbols (a). b) shows the relations for the BAV patients with AV dysfunction, c) demonstrates that there is no correlation in patients without AV dysfunction. X- and y-axis have identical units as shown in a). LV=left ventricle, AAo= ascending aorta, AS=aortic valve stenosis, AR=aortic valve regurgitation.

Fig. 5.

Correlation plots for associations between LV mass and WSS in the AAo for the complete cohort; the subgroups are shown as different symbols (a). b) Demonstrates the relation between LV mass and AAo WSS for the group with AV dysfunction, whereas c) shows that there is no significant association between LV mass and AAo WSS in patients with normal valve function. X- and y-axis have identical units as shown in a). LV=left ventricle, AAo=ascending aorta, AS=aortic valve stenosis, AR=aortic valve regurgitation.

In addition, we detected associations between AAo WSS and ventricular volumes (EDV: r=0.33, p=0.02, ESV: r=0.4, p=0.005). WSS in the aortic arch and DAo showed significant associations (p<0.001) with mass (r=0.46, r=0.41), EDV (r=0.50, r=0.54), ESV (r=0.47, r=0.46), and SV (r=0.47, r=0.55).

Patients with AV dysfunction revealed significant associations between LV mass and WSS arch (r=0.58, p<0.001), between LV mass and AAo/arch velocity (r=0.46, p=0.006; r=0.53, p=0.002) (Fig. 4 and 5 b). In the subgroup without AS/AR, however, there were no significant relationships between LV mass and hemodynamic parameters in the AAo and arch (Fig. 4 and 5 c).

In addition, there were significant associations between elevated WSS and peak velocities in the arch with increased LVEDV (r=0.64, p=<0.001; r=0.35, p=0.04) and with increased LVESV (r=0.58, p<0.001; r=0.37, p=0.03) in patients with AV dysfunction, which were not present in the no AS/AR group. These relationships between increased WSS in the AAo, WSS and velocity in the arch and the LV volumetric results were also detected in the AR group (AAo WSS and LVEDV/LVESV: r=0.69, p=0.004; r=0.65, p=0.009), (arch WSS and LVEDV/LVESV: r=0.80, p<0.001; r=0.7, p=0.004), (arch velocity and LVEDV/LVESV: r=0.82, p<0.001; r=0.70, p=0.004).

Finally, a significant association was found between increased AAo WSS and elevated ECV for the complete cohort (r=0.38, p=0.01) and the AS/AR group (r=0.41, p=0.01) but not for BAV patients with normal AV function (r=0.34, p>0.05).

Discussion

The findings of our study demonstrate the potential of the comprehensive application of 4D flow MRI and cardiac MRI including T1-mapping techniques for the assessment of altered aortic hemodynamics and LV remodeling in BAV patients. As expected, relationships concerning LV parameters in terms of volumes and ECV and increased peak systolic velocities and WSS in the AAo and arch were strongly influenced by the presence of AV dysfunction.

Previous 4D flow MRI studies in BAV patients have focused on the characterization of altered aortic hemodynamics and have shown different flow patterns, valve opening angles, flow displacement and WSS distribution in the AAo depending on the BAV type (5, 7, 8, 25, 26). There is an ongoing debate on the genetic versus hemodynamic origin of the frequently observed progressive AAo dilatation. Currently, it is believed that BAV represents one component of a disease spectrum that includes abnormalities in the aortic wall and appears to have partially similar degenerative changes but less severe clinical consequences than Marfan syndrome (27, 28). Current work suggests that the pathophysiology in BAV is related to the interaction of abnormal WSS distribution, altered aortic wall structure, and an underlying genetic component (28, 29). This hypothesis is supported by a recent study in BAV patients: increased WSS corresponded with extracellular matrix protein dysregulation and elastic fiber degeneration in the AAo, implicating valve-related hemodynamics as a mediator of aortopathy (9).

A recent study investigated the impact of AS on aortic blood flow, WSS, and LV remodeling in a cohort of 37 patients, among those were 16 with BAV (11). In this study, stenotic BAV resulted in higher AAo WSS than stenotic TAV, and AV orifice area and flow displacement were associated with signs of LV remodeling (11). In the aforementioned study, T1-mapping techniques were not carried out to examine structural myocardial effects in addition to volumetric changes as a result of valve disease. Moreover, WSS was not analyzed in its 3D character but was limited to three defined 2D analysis planes placed along the AAo.

Our results revealed significantly elevated AAo WSS in patients with AV dysfunction compared to those with normally functioning BAV. This included patients with isolated AS and those with AS and AR, but not the AR group, indicating that the stenotic component causes these WSS differences in comparison to the BAV patients without additional AS/AR. High flow velocity and eccentric flow have previously been shown to result in progression of aortic diameter and rise in WSS (5, 6). Therefore, these findings are not surprising and can be explained by the underlying valve disease in addition to the congenital BAV component.

We used several currently available sequences for cardiovascular assessment to comprehensively analyze both aortic hemodynamics and LV performance within one examination. Cine-SSFP sequences in the short axis orientation for the assessment of LV volumes and function have been used for more than a decade (30, 31). MOLLI for T1-mapping of the heart is widely applied today with some protocol optimization and other adaptations (16, 32). ECV as a marker of interstitial fibrosis is particularly valuable for the characterization of more diffuse diseases that are more challenging to detect using conventional late gadolinium enhancement (LGE) (16, 18, 33, 34).

Comparing the BAV patients’ LV mass in our study to reference values (30, 31), patients without AV dysfunction and those with isolated AS were within normal limits while the other subgroups had a significantly larger LV mass, most likely caused by AR. LV hypertrophy in BAV patients without AS was also described in a study by Grotenhuis et al. (35). Aortic dilation and reduced elasticity were associated with AR and LV hypertrophy. Elevated ECV - only measured in the septal region - was described in adult patients with severe acquired AS (36). Correlations between diffuse myocardial fibrosis and LV mass were seen in a study in which endomyocardial biopsy was performed in the septum (37). This study revealed more interstitial fibrosis in patients with AR compared with AS. Similar to the study of Singh et al. on asymptomatic AS patients (38), we did not find differences between the ECV values in patients with AV dysfunction and those with normal AV function. In our patient cohort, we did not detect any associations between ECV and LV mass.

ECV correlated moderately with AAo WSS in all patients and those with AS+AR but not in patients with normal valve function. A more distinct LV reaction as a result of the valve-induced afterload is presumably the reason for these findings. We believe that BAV patients may develop more severe LV hypertrophy and signs of mild interstitial fibrosis, and thus more manifest LV remodeling in the presence of concomitant valve dysfunction. In this context, AS/AR had a marked impact on the observed relationships between LV mass with peak systolic velocities and WSS in the AAo. These findings are in line with the expected high pressure gradients across the dysfunctional aortic valve which are associated with rise in myocardial mass and higher flow velocities and thus, higher WSS.

There are several limitations in this work. The BAV patients were examined at 1.5 and 3T scanners. Since the field strength influences T1 values, we focused on ECV for the assessment of structural myocardial changes. Our BAV patient cohort consisted of a heterogeneous group regarding BAV type (Sievers) and concomitant AV pathology in 70% of the cases. This heterogeneity resulted in rather small subgroups after dividing the complete cohort according to the type of valve dysfunction. In particular, the AS subgroup including only seven patients with various AS severity, is underrepresented, and mainly the cause for the absence of any relationship between AS and LV mass. Besides, it was not the aim of this study to investigate the impact of different BAV fusion patterns.

Hypertension is also known to have impact on LV remodeling (39). As the data were assessed retrospectively, we did not have the patients’ blood pressure values but retrieved this information from the patients’ medical reports.

A limitation of our study is related to medical therapy in a sub-group of patients. Twelve of 49 patients (24%) received antihypertensive therapy, which may have impacted cardiac remodeling. Future studies should more systematically investigate the impact of antihypertensive therapy on LV changes and whether different types of drugs (e.g. ACE inhibiters vs. diuretics) have divergent effects. The analysis of our study cohort did not permit the assessment of these subgroups with statistical certainty.

Age is another factor affecting aortic hemodynamics (40). Even though the patients with AV dysfunction were older than those with normal AV function, this difference was only significant for the small group of patients with AS. The age between the complete AS/AR cohort and the no AV-pathology group did not differ significantly.

LV reference values provided in previous studies added the papillary muscles to the myocardial mass (30, 31), whereas we did not include the papillary muscles in the LV mass due to technical aspects in the analysis software. In this setting, we only measured global cardiac function; therefore, subclinical systolic dysfunction in terms of abnormal LV shape due to volume overload from AR could be missed.

In addition, the lack of a reference group, whether healthy volunteers or ideally, patients with TAV, limits our study’s validity. TAV patients with normal valve function do generally not have an indication for undergoing CMR and those with AV dysfunction would presumably not be a useful comparison group. Furthermore, IRB approval for a study with healthy volunteers receiving contrast medium is challenging in this context. Echocardiography was not performed in all subjects within the same period as the MRI scan in order to use it as a reference imaging modality. In addition, a retrospective study has its inherent biases with regard to the study setup.

In general, despite ongoing technical development, 4D flow MRI is still a rather time-consuming method for acquisition and data post-processing. The same critique point could be discussed for T1-mapping. However, these are advanced techniques that provide essential information on hemodynamics and structural myocardial changes compared to conventional MRI-sequences used in the routine clinical setting. This additional knowledge enables an improved understanding of complex cardiovascular relationships like in BAV patients.

In conclusion, AV dysfunction in terms of AS and AR has a major impact on cardiac structure, function and aortic hemodynamics in BAV patients. In our study, BAV patients with no additional AV pathology did not reveal signs of LV impairment due to altered aortic hemodynamics.