Four-dimensional flow cardiovascular magnetic resonance aortic cross-sectional pressure changes and their associations with flow patterns in health and ascending thoracic aortic aneurysm

Abstract

Background

Ascending thoracic aortic aneurysm (ATAA) is a silent and threatening dilation of the ascending aorta (AscAo). Maximal aortic diameter which is currently used for ATAA patients management and surgery planning has been shown to inadequately characterize risk of dissection in a large proportion of patients. Our aim was to propose a comprehensive quantitative evaluation of aortic morphology and pressure-flow-wall associations from four-dimensional (4D) flow cardiovascular magnetic resonance (CMR) data in healthy aging and in patients with ATAA.

Methods

We studied 17 ATAA patients (64.7 ± 14.3 years, 5 females) along with 17 age- and sex-matched healthy controls (59.7 ± 13.3 years, 5 females) and 13 younger healthy subjects (33.5 ± 11.1 years, 4 females). All subjects underwent a CMR exam, including 4D flow and three-dimensional anatomical images of the aorta. This latter dataset was used for aortic morphology measurements, including AscAo maximal diameter (iD_MAX) and volume, indexed to body surface area. 4D flow MRI data were used to estimate 1) cross-sectional local AscAo spatial (∆P_S) and temporal (∆P_T) pressure changes as well as the distance (∆D_PS) and time duration (∆T_PT) between local pressure peaks, 2) AscAo maximal wall shear stress (WSS_MAX) at peak systole, and 3) AscAo flow vorticity amplitude (V_MAX), duration (V_FWHM), and eccentricity (V_ECC).

Results

Consistency of flow and pressure indices was demonstrated through their significant associations with AscAo iD_MAX (WSS_MAX:r = −0.49, p < 0.001; V_ECC:r = −0.29, p = 0.045; V_FWHM:r = 0.48, p < 0.001; ∆D_PS:r = 0.37, p = 0.010; ∆T_PT:r = −0.52, p < 0.001) and indexed volume (WSS_MAX:r = −0.63, V_ECC:r = −0.51, V_FWHM:r = 0.53, ∆D_PS:r = 0.54, ∆T_PT:r = −0.63, p < 0.001 for all). Intra-AscAo cross-sectional pressure difference, ∆P_S, was significantly and positively associated with both V_MAX (r = 0.55, p = 0.002) and WSS_MAX (r = 0.59, p < 0.001) in the 30 healthy subjects (48.3 ± 18.0 years). Associations remained significant after adjustment for iD_MAX, age, and systolic blood pressure. Superimposition of ATAA patients to normal aging trends between ∆P_S and WSS_MAX as well as V_MAX allowed identifying patients with substantially high pressure differences concomitant with AscAo dilation.

Conclusion

Local variations in pressures within ascending aortic cross-sections derived from 4D flow MRI were associated with flow changes, as quantified by vorticity, and with stress exerted by blood on the aortic wall, as quantified by wall shear stress. Such flow-wall and pressure interactions might help for the identification of at-risk patients.

Graphical Abstract

1. Background

Ascending thoracic aortic aneurysm (ATAA) is an asymptomatic dilation of the ascending aorta (AscAo) that can ultimately lead to aortic dissection and consequently to an increased morbi-mortality. AscAo dilatation is frequently observed in specific diseases, such as bicuspid aortic valve as well as Marfan or Turner syndromes, but also in the general population [1]. Clinical recommendations for the evaluation of risk of aortic rupture and prophylactic surgery planning are mainly based on the measurement of maximal aortic diameters from imaging data [2], such as echocardiography, computed tomography (CT) and cardiovascular magnetic resonance (CMR) angiography. Such diameter measures are often performed on pre-selected two-dimensional (2D) slices. An ATAA patient requires follow-up once maximal aortic diameter exceeds 42 mm and when such diameter exceeds 55 mm, the patient is referred to surgery. However, a patient may undergo aortic dissection despite an aortic diameter below surgical threshold [3]. Accordingly, new more sensitive and specific indices are urgently needed to precisely identify patients at risk of aortic rupture.

Automated three-dimensional (3D) segmentation methods have been proposed to extract morphological parameters, such as aortic length, volume, and curvature for all thoracic aorta regions from 3D CMR images [4], [5] or CT [6] angiograms. In addition to such 3D aortic morphology, CMR can provide a functional evaluation of the aorta, including quantitative analyses of aortic blood flow using 4D flow sequences. Several studies have highlighted the development of specific flow patterns in aortic dilatation [7], [8]. Such flow patterns have been characterized using 4D flow CMR through various quantitative parameters [7], [9], including 1) wall shear stress (WSS) [10], [11]; 2) vorticity patterns [5], [12]; 3) intra-arterial pressure changes through the aortic length which has demonstrated added value in valve stenosis or aortic coarctation [13]. Of note, to the best of our knowledge, pressure differences within aortic cross-sections, which can provide insight into how changes in geometry locally affect blood flow, have never been investigated. Such cross-sectional component can be of particular interest since aneurysm morphology may be variable exhibiting asymmetric or localized dilation.

The aorta has been characterized using the multiple above-mentioned morphological and velocity-based quantitative parameters. However, the interaction between these various complementary parameters in health and aneurysmal disease remains underexplored. Thus, our objectives are 1) to estimate aortic cross-sectional pressure differences as well as flow vorticity and WSS non-invasively in the AscAo using 4D flow CMR data; 2) to study the relationship between these flow parameters and aortic 3D-derived morphological parameters, namely maximal diameter and volume; 3) to investigate the associations and interactions between velocity-derived features (pressure, WSS, vorticity). To achieve these objectives, we performed our analyses on 1) patients with AscAo dilatation and a tricuspid aortic valve, 2) healthy volunteers matched for age, sex, and blood pressure to the dilated patients, and 3) healthy and young volunteers providing normality for all aortic parameters. The above-mentioned matching is crucial because age, sex, and blood pressure are major confounding factors for the targeted aortic measurements.

2. Methods

2.1. Study population and data acquisition

We retrospectively studied 17 ATAA patients (64.7 ± 14.3 years, 5 females) with a tricuspid aortic valve without severe stenosis, regurgitation, or previous surgery. ATAA was defined by a maximal diameter over the AscAo ≥41 mm or ≥22 mm/m² when indexed to body surface area (BSA). We also studied 17 healthy subjects (59.7 ± 13.3 years, 5 females) matched for age, sex, and systolic blood pressures to the ATAA patients, as well as 13 younger and healthy individuals (33.5 ± 11.1 years, 4 females) to derive representative “normal” quantitative measures and associations. MRI data of these healthy individuals are part of a local protocol focusing on the non-invasive evaluation of myocardial stiffness by ultrafast echo as compared to CMR (NCT02537041). Accordingly, the present ancillary work objectives are independent from the goals of the primary protocol. Approval from the local ethics committee and subjects informed consent were obtained.

CMR was performed on a 3T scanner (MR750w GEM, GE Healthcare, Waukesha, Wisconsin) with a 32-channel cardiac phased-array coil, after injection of gadolinium-based contrast agent (0.1 to 0.2 mmol/kg gadoteridol [Guerbet, Aulnay-sous-Bois, France]). 4D flow data were acquired during free-breathing with retrospective electrocardiogram gating in a sagittal oblique volume encompassing the thoracic aorta, using the following scan parameters: echo time = 1.7 ms, repetition time = 4.3–4.4 ms, flip angle = 15°, voxel size = 1 × 1.48 × 2.38 mm³, and velocity encoding = 250 cm/s in all directions. Velocities were encoded in each direction with a balanced four-point scheme obtained every 4 repitition time. The number of views per segment was fixed to 2 and total acquisition duration was around 10 min. Datasets had an effective temporal resolution of 34.8 ms and were reconstructed with 50 cardiac phases after using view sharing independent of heart rate. Brachial pressures were recorded simultaneously to CMR acquisitions using the SphygmoCor Xcel device (AtCor Medical, Sydney, Australia), while immediately after, applanation tonometry of the right carotid artery was performed by an experienced operator (10 years) with the PulsePen device (DiaTecne, Milan, Italy). Applanation tonometry acquisitions were performed in a lying position and in a temperature-controlled semi-dark room, to reproduce MRI exam conditions, resulting in central systolic blood pressure (SBP) and diastolic blood pressure (DBP).

Additionally, to assess morphology, all patients underwent a sagittal oblique volume acquisition encompassing the entire thoracic aorta using a 3D spoiled gradient recalled (SPGR) sequence with the following scan parameters: flip angle = 24°, repetition time = 3.1 ms, echo time = 1.3 ms, voxel size = 0.67 × 0.67 × 3.2 mm³.

2.2. Image pre-processing and aortic segmentation

The SPGR images were used for semi-automated 3D aortic segmentation through an explicit active contours algorithm implemented within the Mimosa software (LIB, Sorbonne University, Paris), which was previously shown to reproducibly quantify aortic morphology [4]. Such segmentation enabled the quantitative measurement of AscAo maximal diameter (D_MAX) perpendicular to the centerline and volume (Vol_AscAo). AscAo maximal diameter and volume were then indexed to body surface area (iD_MAX, iVol_AscAo).

4D flow MRI data were corrected for phase offset and phase wrapping according to guidelines [7]. A time-averaged phase-contrast MR angiography (PC-MRA) was then derived from the 3 directional velocities weighted by the modulus images, when considering 5 time phases around the systolic peak, which was defined as the temporal phase with maximal velocity in the AscAo [13]. Finally, the same aforementioned segmentation algorithm [4] was applied to the resulting PC-MRA dataset to segment the thoracic aorta volume, define the aortic centreline and thus isolate aortic velocity fields.

In addition, for each subject, PC-MRA segmentation was used to further define 64 cross-sectional planes perpendicular to the aortic centreline from the most proximal level of the AscAo, to the celiac trunk bifurcation. AscAo segment was delimited by the brachiocephalic bifurcation (Fig. 1) and comprised 5 to 11 cross-sectional planes, depending on individual AscAo length. Algorithms and user interface were implemented in MATLAB (The MathWorks, Natick, Massachusetts).

Fig. 1.

Aortic 4D flow MRI data processing pipeline. Upper panel: 4D flow CMR modulus and velocity components images and aortic segmentation as well as aortic cross-sectional slicing. Bottom panel: estimation of relative 3D pressure map using iterative Poisson equation and definition of temporal and spatial changes in pressure amplitude (∆P_T, ∆P_S) as well as temporal and spatial distance between pressure extrema in the ascending aorta (∆T_PT, ∆D_PS); estimation of wall shear stress (WSS) map and definition of the spatio-temporal WSS peak within the ascending aorta (WSS_MAX); estimation of vorticity using the λ₂-method and definition of the vortex eccentricity (V_ECC), maximal amplitude (V_MAX), and duration (V_FWHM) within the ascending aorta.

2.3. Quantitative indices of intra-aortic pressures

The 3D+t aortic pressure maps relative to the zero reference set at the most proximal level of the AscAo were obtained while applying pressure Poisson equations to the 4D flow CMR velocity fields delimited by the aortic segmentation of the PC-MRA dataset as previously described [14]. To capture the effects of changes in blood flow patterns on intra-aortic pressure distribution, several quantitative cross-sectional indices reflecting either spatial or temporal pressure variations were derived from the time-resolved relative pressure maps within the AscAo segment (Fig. 1). Spatial pressure quantitative indices included 1) maximal pressure difference within each slice (magnitude between pressure extrema within such single cross-sectional location) for each time phase (∆P_S(t)), which was then averaged for all AscAo cross-sectional planes, to minimize the effect of noise (Fig. 2), and its maximal value through time was named ∆P_S (mmHg); 2) the distance between pressure extrema within each slice and for each time phase (∆D_PS(t)), which was then averaged for all AscAo cross-sectional planes and its maximal value through time was named ∆D_PS (mm). Finally, relative pressure time curve averaged across all AscAo cross-sectional planes was calculated and used to estimate the following pressure quantitative indices: 1) the difference in magnitude between systolic and diastolic relative pressure peaks (∆P_T, mmHg); 2) the duration between these two peaks (∆T_PT, ms).

Fig. 2.

Ascending aorta time-resolved cross-sectional maximal pressure changes. Individual curves of maximal pressure changes within each ascending aorta cross-sectional slice (gray) and the average curve over all slices (red).

2.4. Quantitative indices of vorticity

The λ2-method that was initially proposed by Jeong and Hussain [15] and validated quantitatively in several studies [12], [16] was used for vortex identification in our segmented aortic 4D flow data. The following parameters were then estimated in each cross-sectional plane of the AscAo (Fig. 1): 1) the peak vortex amplitude across the cardiac cycle (V_MAX) derived from the vorticity time curve, which was averaged over the whole AscAo segment for each time phase; 2) average vortex duration in the AscAo, which was defined as the full width at half maximum (V_FWHM, ms) of the aforementioned vorticity time curve. In addition, vortex eccentricity was estimated as the distance between the vortex center and the aortic centreline indexed by local aortic diameter (V_ECC), where vortex center was defined as the center of mass of the largest cluster of pixels identifying a vortex.

2.5. Quantification of wall shear stress

WSS was calculated using the method previously described by Potters et al. [17], which minimizes errors related to the estimation of the spatial variations of blood flow velocities at the aortic wall boundaries. Briefly, space rotation was performed to be able to locally define the vector normal to the aortic wall according to a single coordinate, namely z. This rotation was individually applied to each point of the aortic segmentation where WSS was calculated.

WSS was evaluated for all aortic segmentation points and at all cardiac cycle time-phases (Fig. 1). Then spatial maxima of WSS were estimated at systolic peak for each plane of the AscAo segment and were averaged resulting in maximal AscAo WSS (WSS_MAX).

2.6. Statistical analysis

Continuous variables were provided as mean ± standard deviation for each group, after testing for the normality of their distribution using the Shapiro-Wilk test. Linear regressions were used to study associations between AscAo 4D flow CMR-derived indices, namely relative pressure, WSS, and vorticity as well as their associations with indices of AscAo morphologies, including maximal diameter and volume. For each linear regression, correlation coefficient and p-value were provided and further adjustment for age and SBP, as well as iD_MAX when appropriate, was performed using a multivariate regression model. Statistical differences between groups were tested using the nonparametric Wilcoxon rank-sum test. All reported p-values were two-sided and statistical significance was indicated by a p-value <0.05. Analyses were performed using Stata software (StataCorp, College Station, Texas).

3. Results

Fig. 3 shows two examples of 3D blood flow velocity, relative pressure, and WSS maps derived from 4D flow CMR of a young healthy subject and a patient with dilated AscAo along with time-resolved velocity, WSS, vorticity, and relative pressure curves in the AscAo averaged over young and elderly healthy subjects as well as ATAA patients. The 3D maps highlighted a drop in velocity magnitude, spatial pressure difference, as well as WSS in the dilated AscAo, as compared to the AscAo of the healthy subject. In line with these qualitative observations in single cases, time-resolved curves illustrated that both physiological (aging) and to a greater extent pathological (ATAA) aortic dilatations were associated with 1) a decrease in blood flow velocity magnitude, 2) a decrease in WSS magnitude, 3) a decrease in vortex magnitude along with an increase in its duration, and 4) a decrease in relative pressure magnitude along with an earlier occurrence of its second peak.

Fig. 3.

Aortic 4D flow CMR-derived velocity, relative pressure, WSS, and vorticity. Examples of quantitative hemodynamic maps for a young healthy volunteer (top) and a patient with a dilated ascending aorta (bottom), along with mean (solid lines) and standard deviation (shades) for time-resoled velocity, WSS, vorticity, and relative pressure curves in the ascending aorta of young healthy subjects (gray), elderly healthy subjects (yellow), and ATAA patients (blue). Time curves were rescaled between 0 and 1000 ms for all patients before averaging them and the averaged curve was then truncated at 700 ms. CMR, cardiovascular magnetic resonance; WSS, wall shear stress.

Consistently with the illustrated trends, Table 1 summarizes patients and healthy subject characteristics along with CMR-derived quantitative AscAo morphological and hemodynamic measures. Four ATAA patients had mild aortic regurgitation (AR), 4 had moderate AR while the remaining 9 patients had no AR. According to our study design, there were no significant differences in age, SBP, DBP, and BSA between ATAA patients and their matched controls. SBP was slightly higher in the elderly controls as compared to the younger healthy controls. We found a physiological enlargement of the aorta with age, as revealed by the significant increase in AscAo maximal diameter and volume in the elderly controls as compared to the younger healthy controls. Such aortic enlargement was expectedly even more pronounced in ATAA patients.

Table 1.

Subjects characteristics and CMR measurements in the ascending aorta. Data are provided as numbers or mean±standard deviation.

Parameter	Young controls	Elderly controls	ATAA patients
n (men/women)	13 (9/4)	17 (12/5)	17 (12/5)
Age (years)	33.5 ± 11.1	59.7 ± 13.3+	64.7 ± 14.3
BSA (m2)	1.7 ± 0.20	1.8 ± 0.20	1.9 ± 0.34
SBP (mmHg)	111.6 ± 13.4	116.1 ± 11.0	117.8 ± 16.3
DBP (mmHg)	79.9 ± 9.1	81.8 ± 5.4	80.6 ± 9.6
DMAX (mm)	27.8 ± 2.4	31.4 ± 3.3+	42.9 ± 4.9*
iDMAX (mm/m2)	16.2 ± 1.5	17.4 ± 2.2	23.1 ± 4.7*
VolAscAo(mL)	26.4 ± 7.5	36.9 ± 10.9+	95.13 ± 40.19*
iVolAscAo(mL/m²)	15.2 ± 3.6	20.2 ± 5.7+	49.4 ± 17.6*
WSSMAX (Pa)	0.87 ± 0.15	0.69 ± 0.20+	0.54 ± 0.15*
VMAX (s−1)	121.7 ± 64.6	95.6 ± 52.6	67.2 ± 46.1*
VECC	0.24 ± 0.05	0.25 ± 0.03	0.19 ± 0.09*
VFWHM (ms)	334.1 ± 120.9	440.8 ± 118.0+	631.1 ± 197.4*
∆PS (mmHg)	1.7 ± 0.6	1.6 ± 0.6	1.6 ± 0.7
∆DPS (mm)	13.7 ± 2.6	15.7 ± 3.2+	19.3 ± 4.3*
∆PT (mmHg)	2.0 ± 1.1	1.7 ± 0.5	1.5 ± 0.5
∆TPT (ms)	268.5 ± 65.6	213.1 ± 71.8	132.8 ± 58.1*

AscAo ascending aorta, BSA body surface area, D_MAX/iD_MAX non-indexed/indexed maximal AscAo diameter, SBP/DBP systolic/diastolic blood pressure, V_ECC AscAo vortex eccentricity, V_FWHM AscAo vortex duration, V_MAX AscAo maximal vorticity magnitude, Vol_AscAo/iVol_AscAo non-indexed/indexed AscAo volume, WSS_MAX AscAo maximal wall shear stress, ∆P_S cross-sectional maximal pressure difference through the AscAo, ∆D_PS distance between pressure extrema within the AscAo,∆P_T difference in magnitude between systolic and diastolic relative pressure peaks within the AscAo, ∆T_PT duration between systolic and diastolic pressure peaks within the AscAo, +indicates significance of comparisons between elderly and young healthy subjects, *indicates significance of comparisons between ATAA patients and elderly controls .

Furthermore, while WSS was significantly lower in the elderly as compared to younger controls and even lower in ATAA patients, vortex duration was significantly higher in ATAA patients and lowest in healthy younger controls. Vortex amplitude has a decreasing trend between young and elderly controls and was significantly lower in ATAA patients as compared to elderly controls. Among pressure indices, duration between systolic and diastolic relative pressure peaks, ∆T_PT, was lower in the elderly controls as compared to younger controls and even lower in the ATAA patients. The distance between relative pressure extrema within the AscAo was significantly higher in the ATAA patients as compared to their matched controls, as well as between elderly and younger controls.

3.1. Associations between ascending aorta hemodynamic and morphological indices

Associations between hemodynamic and both indexed or non-indexed AscAo maximal diameter and volume for the entire study group are summarized in Table 2. We found a significant and negative association between WSS_MAX and both AscAo diameter and volume. Vortex eccentricity in the AscAo (V_ECC) was negatively and significantly associated with AscAo diameter and volume. Vortex duration (V_FWHM) showed a significant increase with AscAo diameter and volume. In addition, the distance between cross-sectional pressure extrema within the AscAo (∆D_PS) was significantly and positively correlated with AscAo diameter and volume. Finally, the duration between systolic and diastolic relative pressure peaks within the AscAo (∆T_PT) decreased significantly with AscAo diameter and volume. Most of these associations remained significant after adjustment for age and SBP (Table 2).

Table 2.

Associations between ascending aorta hemodynamic and morphological indices (whole group, n = 47). Data are provided as (correlation coefficient r, corresponding p value) or ns.: non-significant.

	DMAX (r, p)	iDMAX (r, p)	VolAscAo (r, p)	iVolAscAo (r, p)
WSSMAX	(−0.66, <0.001)*	(−0.49, <0.001)*	(−0.60, <0.001)*	(−0.63, <0.001)*
VMAX	(−0.30, 0.040)	ns.	ns.	(−0.29, 0.048)
VECC	(−0.44, 0.002)*	(−0.29, 0.045)	(−0.51, <0.001)*	(−0.51, <0.001)*
VFWHM	(0.56, <0.001)*	(0.48, <0.001)*	(0.47, <0.001)	(0.53, <0.001)
∆PS	ns.	ns.	ns.	ns.
∆DPS	(0.64, <0.001)*	(0.37, 0.010)	(0.56, <0.001)	(0.54, <0.001)*
∆PT	ns.	ns.	ns.	ns.
∆TPT	(−0.69, <0.001)*	(−0.52, <0.001)*	(−0.62, <0.001)*	(−0.63, <0.001)*

D_MAX/iD_MAX non-indexed/indexed maximal AscAo diameter, ns. non-significant results, V_ECC AscAo vortex eccentricity, V_FWHM AscAo vortex duration, V_MAX AscAo maximal vorticity magnitude, Vol_AscAo/iVol_AscAo non-indexed/indexed AscAo volume, WSS_MAX AscAo maximal wall shear stress, ∆P_S cross-sectional maximal pressure difference through the AscAo, ∆D_PS distance between pressure extrema within the AscAo, ∆P_T difference in magnitude between systolic and diastolic relative pressure peaks within the AscAo, ∆T_PT duration between systolic and diastolic pressure peaks within the AscAo. *Remained significant after adjustment for age and systolic blood pressure (p < 0.05).

3.2. Associations between pressure and flow in normal aging

Table 3 summarizes associations of intra-aortic pressure indices with flow parameters such as vorticity and WSS in healthy subjects (30 subjects, 9 women, mean age: 48.3 ± 18.0 years), revealing that when local spatio-temporal pressure differences increase there is an increase in WSS along with high magnitude but quickly vanishing vortex. Indeed, maximal cross-sectional spatial pressure difference within the AscAo (∆P_S) was significantly and positively correlated with WSS_MAX (Fig. 4A) and maximal vorticity magnitude V_MAX (Fig. 4B), while it was significantly and negatively correlated with vortex duration V_FWHM. Fig. 5 illustrates examples of extreme values of ∆P_S along with the corresponding changes in ascending aorta velocity, WSS, and vorticity throughout the cardiac cycle. Of note, associations with WSS_MAX and V_MAX were independent of age, SBP, and AscAo indexed maximal diameter (p = 0.03 and p = 0.006, respectively). Furthermore, difference in magnitude between systolic and diastolic relative pressure peaks (∆P_T) was significantly and positively correlated with WSS_MAX, while it was significantly and negatively correlated with vortex eccentricity V_ECC.

Table 3.

Associations between pressure and flow in healthy subjects. (30 subjects, 9 women, 48.3±18.0 years). Data are provided as (correlation coefficient r, corresponding p value) or ns.: non-significant.

	∆PS (r, p)	∆PT (r, p)
WSSMAX	(0.59, <0.001)*	(0.40, 0.03)
VMAX	(0.55, 0.002)*	ns.
VECC	ns.	(−0.40, 0.03)
VFWHM	(−0.40, 0.03)	ns.

AscAo ascending aorta, ns. non-significant results, V_ECC AscAo vortex eccentricity, V_FWHM AscAo vortex duration, V_MAX AscAo maximal vorticity magnitude, WSS_MAX AscAo maximal wall shear stress, ∆P_S cross-sectional maximal pressure difference through the AscAo, ∆P_T difference in magnitude between systolic and diastolic relative pressure peaks within the AscAo. *Remained significant after adjustment for age, systolic blood pressure and iD_MAX (p < 0.05).

Fig. 4.

Pressure and flow associations in healthy volunteers. Associations between cross-sectional spatial pressure difference (∆P_S) and flow indices within the ascending aorta of healthy subjects (young controls in gray and healthy elderly controls matched to ATAA patients in orange): WSS_MAX (A) as well as maximal vorticity magnitude V_MAX (B).

Fig. 5.

Examples of extreme values of ascending aorta cross sectional pressure changes. High (left) and low (right) ascending aorta cross-sectional pressure changes along with the corresponding velocity, WSS, and vorticity curves throughout the cardiac cycle.

3.3. Associations between pressure and flow in aortic dilation

Fig. 6 illustrates the above-mentioned “normal” linear regressions of pressure with WSS and vorticity magnitude with further superimposition of patients with ATAA, revealing that, thanks to their remodeled dilated lumen, more than half of ATAA patients had low spatial intra-ascending aorta cross-sectional pressure differences, with similar magnitude as elderly subjects. However, one might highlight that some ATAA patients (6/17 patients) had elevated intra-aortic pressure differences despite their AscAo dilatation (5 of these 6 patients had a non-indexed AscAo maximal diameter >41 mm).

Fig. 6.

Pressure and flow associations in aortic dilation. Patients with aortic dilation (blue) are superimposed to linear regression dot plots obtained on young (gray) and elderly (orange) healthy subjects for associations between AscAo pressure cross-sectional spatial difference and WSS (A) and between AscAo pressure cross-sectional spatial difference and vorticity magnitude (B). Patients with elevated spatial pressure differences despite pronounced AscAo dilation are highlighted with red circles. WSS_MAX, maximum wall shearstress; V_MAX, maximum flow vorticity amplitude; ΔP_S, ascending aorta spatial pressure change.

4. Discussion

Novel indices to quantify cross-sectional and temporal pressure changes within the ascending aorta from 4D flow CMR images were proposed and were shown to be mechanistically associated with aortic geometry, WSS as well as vortex amplitude, duration, and eccentricity in health and aortic dilation.

Intra-aortic pressure gradients can be decomposed into 1) longitudinal pressure gradient along the aorta, which measures pressure gradients through the aortic length. It is valuable for the assessment of the overall effect of geometrical changes, including pathological dilation or narrowing and even physiological tapering [14], [18], [19], on pressure distribution. Such longitudinal pressure gradient can help evaluate the uniformity/non-uniformity as well as the severity of the geometrical changes. Indeed, several studies focused on intra-aortic pressure mapping from 4D flow CMR images either in patients or healthy volunteers [13], [18], [20], [21], highlighting the ability of the resulting maps to capture both disease-related and subclinical changes in local pressure differences through the aortic length. Associations of such longitudinal changes with pressure wave reflection and left ventricular remodeling in aging have also been reported [14]. 2) Transverse or cross-sectional pressure gradient focuses on pressure differences within aortic cross-sections providing insight into how a local change in geometry locally affects blood flow. Indeed, aortic curvature is shown to affect pressure patterns, inducing higher pressures in the external part of the curved aortic segment [18], as is the case in the ascending aorta. Aortic cross-sectional pressure gradients can be particularly valuable in aneurysms with variable morphology exhibiting asymmetric or localized dilation. As such, regions of high- or low-pressure within the aneurysm can be assessed through such cross-sectional pressure gradient and potentially help identifying areas at risk of rupture or dissection.

4D flow CMR has been used in several recent studies in aging and in aortic diseases to individually estimate aortic stiffness [22], [23], WSS [24], [25], intra-aortic relative pressures along with their changes through the aortic length [13], [14], [18], [19] as well as vorticity [5], [26]. Aortic WSS has been widely documented in terms of methodological [27], [28], [29], [30] and clinical [31], [32], [33], [34], [35] evaluations with a specific focus on patients with a bicuspid aortic valve [31], [33]. Fewer studies focused on vorticity per se with a very pioneering work from Kilner et al. [36], and few recent studies described and quantified blood flow vortical organization in the thoracic aorta of healthy subjects [5], [12], as well as vanishing and altered patterns in patients, while using the λ2 method [12]. Of note, WSS and vorticity values obtained in our population are in the same range as values provided in the above-mentioned WSS [32], [37], [38] and vorticity [12] studies. One might highlight however that changes in flow patterns in the thoracic aorta have been widely described through the quantitative evaluation of backward flow from 2D phase-contrast MRI [39] and more recently from 4D flow CMR [40]. Such backward flow, which starts during late systole and persists until late diastole, has been attributed to the local reversal of pressure gradients [41], [42], inducing a local reversal in blood flow, captured by the velocity sign on either 2D or 4D phase contrast CMR images. Flow reversal patterns are mostly located in the curved segments of the aorta [40], which is in line with the higher convective relative pressures observed in the external part of such segments [18] and with cross-sectional changes in aortic pressure gradients found in our study.

In the present study, consistency of the comprehensive quantification of 3D anatomical and 4D flow images was evaluated through associations between cross-sectional pressure or flow parameters and morphological indices of the aorta. As such, we found a decrease in maximum WSS in dilation, which is in line with literature findings [32]. Besides, a reduction in the magnitude of the vortex as well as an increase in its duration were observed with increasing aortic size, in agreement with visual or quantitative findings in the literature [5], [43]. Furthermore, our data revealed that vortices tended to be eccentric in normal-sized aortas and more centralized in dilated aortas. This result is consistent with observations in aging, since backward flow area in the ascending aorta is confined to its inner curvature in young subjects and tends to grow and appear in a larger proportion of the lumen in elderly subjects with larger aortic stiffness and dimensions [39]. Finally, we found that the duration between systolic and diastolic pressure peaks tends to be shorter in larger aortas. This is consistent with previous findings [14] and might be explained by the early return of wave reflections in larger and stiffer aortas.

Regarding flow and pressure associations in healthy aging, we found a positive association of WSS and vortex magnitudes with spatial and temporal cross-sectional pressure differences. These associations might be explained by the fact that blood flow orientation and velocity are locally governed by pressure gradients. Indeed, a locally important pressure difference seems to be associated to a high magnitude vortex and high WSS exerted on the inner aortic wall. These preliminary associations highlight that spatio-temporal knowledge on intra-aortic cross-sectional pressure distribution could give indications on local flow pattern changes and their interaction with the aortic wall.

Finally, superimposition of ATAA patients data to “normal” trends associating pressure with flow indices revealed that more than half of the dilated patients had the same distribution as elderly subjects. These patients seem to present regularized and pseudo-normalized pressure gradients, vorticity, and WSS through the augmentation of their ascending aorta size. However, one-third of the patients had elevated intra-aortic pressure differences despite AscAo dilatation. These latter patients combine substantial aortic dilatation, which may weaken their aortic wall, with strong spatio-temporal cross-sectional pressure differences, linked to stronger vortices and potential subsequent extra-dilation. Of course, this is an hypothesis that further specific studies should confirm.

5. Limitations

The main limitation of our study is group size. Although the targeted associations were significant despite such small numbers, including more subjects is necessary in order to strengthen conclusions of this preliminary work. In addition, a longitudinal follow-up of the patients is mandatory in future studies using all these innovative 4D flow CMR quantitative indices to evaluate their added value, as compared to conventional geometrical characterization of the aorta in the identification of at-risk patients.

6. Conclusion

A non-invasive and comprehensive characterization of aortic geometry, blood flow patterns, flow-wall interactions, and inner cross-sectional pressure gradients is now rendered possible while using 4D flow CMR. Such evaluation allowed for studying flow-pressure complementarity in healthy aging and for potential identification of patients with dilated ascending aorta concomitant with high cross-sectional pressure gradients, as compared to age-matched healthy individuals. Application to a larger database with longitudinal follow-up may aid in elucidating the usefulness of such interplay in patients with aortic diseases.

Funding

We would like to acknowledge the FRM project ING20150532487 for funding Kevin Bouaou and ESME-Sudria for funding Sophia Houriez–Gombaud-Saintonge as well as the ECOS-SUD project number A15S04 (France-Argentina) exchange grant for funding a fruitful exchange around 4D flow MRI data processing.

Author contributions

Sophia Houriez–Gombaud-Saintonge: Formal analysis, Methodology, Software, Writing – original draft, Writing – review and editing. Ioannis Bargiotas: Methodology, Software, Writing – review and editing. Elie Mousseaux: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – original draft, Writing – review and editing. Thomas Dietenbeck: Methodology, Software, Supervision, Writing – review and editing. Emilie Bollache: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review and editing. Kevin Bouaou: Conceptualization, Data curation, Formal analysis, Methodology, Software, Visualization, Writing – original draft, Writing – review and editing. Didier Lucor: Formal analysis, Methodology, Supervision, Validation, Writing – original draft, Writing – review and editing. Emmanuel Messas: Data curation, Funding acquisition, Investigation, Writing – review and editing. Elena Jiménez: Methodology, Software, Writing – original draft, Writing – review and editing. Nadjia Kachenoura: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. Alain Giron: Methodology, Writing – original draft, Writing – review and editing. Umit Gencer: Data curation, Resources, Writing – review and editing. Alain De Cesare: Methodology, Software, Writing – review and editing. Gilles Soulat: Conceptualization, Data curation, Funding acquisition, Resources, Writing – original draft, Writing – review and editing.

Ethics approval and consent

This study conformed to local ethics committee regulations. Ethics committee approval was waived for this study type. Signed informed consent was given by all participants.

Consent for publication

Not applicable.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Nadjia Kachenoura reports financial support was provided by Fondation pour la recherche Médicale. Sophia Houriez–Gombaud-Saintonge reports a relationship with ESME Sudria that includes employment. Nadjia Kachenoura reports a relationship with ECOS SUD that includes funding grants. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.