Three-dimensional assessment of ascending aortic stiffness, motion, and growth in ascending thoracic aortic aneurysm

Abstract

Aims

Aortic wall stiffening in ascending thoracic aortic aneurysm (aTAA) is common. However, the spatial and temporal relationships between stiffness, aortic size, and growth in aTAA remain unclear.

Methods and results

In this single-centre retrospective study, we utilized vascular deformation mapping to extract multi-directional aortic motion, aortic distensibility, and aortic growth in a multi-planar fashion from multi-phasic ECG-gated computed tomography angiograms. Aortic displacement and stiffness metrics were compared between patients with sporadic ascending aortic dilation (Dilated), individuals without thoracic aortic dilation, and patients with Marfan syndrome. A total of 96 patients were included. Total and axial aortic root motion was significantly decreased in the Dilated group (n = 49) compared with the Non-dilated group (n = 38) and Marfan group (n = 16). Aortic distensibility was significantly lower in the Dilated group compared with the Non-dilated group and exhibited a more diffuse pattern of stiffening compared with the Marfan group in which stiffening was localized to the root. In Dilated group, aortic distensibility was moderately and positively associated with aortic growth rate (R = 0.34, P = 0.02). The moderate-to-strong association between age and aortic stiffness in non-dilated segments was either significantly blunted or absent in dilated segments.

Conclusion

Vascular deformation mapping provides multi-level stiffness assessments of the ascending aorta using multi-phasic computed tomography angiography. Ascending aortic stiffening is a spatially heterogeneous process with stiffening tending to increase with degree of regional dilation and age, whereas lower stiffness was associated with faster growth of the mid-ascending aorta in those with sporadic aTAA.

Graphical Abstract

Open in new tabDownload slide

Introduction

Ascending thoracic aortic aneurysm (aTAA) is characterized by degradation of the normally robust and elastic aortic wall, leading to chronic dilation and potentially catastrophic events such as dissection and rupture.^1–3 Central aortic stiffening in the absence of aortic dilation has been well documented as a risk factor for a range of adverse cardiovascular outcomes (e.g. stroke, myocardial infarction, congestive heart failure).⁴ In aTAA, wall stiffening localized to the aneurysm itself is common, largely owing to increased collagen content and disorganization in combination with the accelerated elastin fibre fragmentation that is a hallmark of the aortopathy seen in Marfan syndrome. However, spatial and temporal relationships between stiffness, aortic size, and progressive growth in aTAA remain incompletely understood. As a result, aortic stiffness has not become a key metric for informing disease severity and estimating in aTAA, despite frequent calls for biomechanical metrics to improve on current diameter-based assessments.^5–7

Current limitations in our understanding of aortic stiffness in patients are in part related to shortcomings in techniques for assessing stiffness in vivo. For example, the most common imaging techniques for quantifying aortic stiffness are based in echocardiography and cine magnetic resonance imaging (MRI), approaches that measure pulsatile changes in aortic diameter and/or area on two-dimensional (2D) images, typically localized to one aortic region (e.g. mid-ascending aorta for MRI and proximal tubular ascending aorta for echocardiography).^8–11 Results from such analyses may differ based on both operator dependence and be subject to variability based on regional variations in stiffness along the aortic length. Further, methods of assessing global aortic stiffness though measurement of pulse wave velocity (PWV) by carotid-to-femoral applanation tonometry do not directly assess the ascending aortic segment and thus have limited value for understanding the behaviour of ascending aortic aneurysm.^5,12,13

Vascular deformation mapping (VDM) is an emerging image analysis technique that uses ECG-gated computed tomography angiography (CTA) images to quantify three-dimensional (3D) changes in aortic geometry between two different scans or points in the cardiac cycle. VDM was originally described as a technique to map growth of the thoracic aorta between two CTAs during imaging surveillance (termed VDM-G for ‘growth’) and has been shown to outperform expert raters for quantifying thoracic aortic growth.¹⁴ More recently, VDM analysis has been adapted to quantify pulsatile ascending aortic motion owing to left ventricular contraction as well as regional stiffness (termed VDM-D for ‘dynamic’).^14–18 Given the high-resolution, four-dimensional nature of modern CTA image acquisition, VDM analysis affords a unique opportunity to better understand the relationships between aortic motion and stiffness (VDM-D) and subsequent aortic growth over surveillance (VDM-G).

In this study, we aimed to combine multi-phasic CTA data and VDM analysis to perform a comprehensive assessment of ascending aortic motion, regional stiffness, and longitudinal growth in patients with both non-dilated and dilated ascending aortas, including subgroups with sporadic disease and Marfan syndrome, to better understand the temporal and spatial relationships between these metrics. We hypothesized that aortic stiffness is variable across the length of the ascending aorta and that VDM-D will allow depiction of disease-specific stiffness patterns.

Methods

Ethical statement

This retrospective study was performed as part of an institutional review board-approved study (HUM00133798), and informed consent was waived.

Patient population

We performed a single-centre retrospective cohort study spanning from 2011 to 2023. Patients were selected from an institutional research database of patients that contains patients with dilated ascending aortas who underwent clinical CTA imaging, as well as a population of patients without thoracic aortic dilation who underwent clinical CTA for reasons unrelated to thoracic aortic disease. We specifically aimed to assemble three groups of patients to examine spatial and temporal relationships between aortic root motion, aortic stiffness, baseline aortic size, and patient characteristics (e.g. age, sex) as follows: (i) patients with sporadic ascending thoracic aortic dilation (Dilated), defined as aortic root and/or ascending aorta ≥4.0 cm diameter, with the absence of any syndromic features and no family history of thoracic aortic disease. To allow for longitudinal assessment of aortic motion and aortic growth, Dilated patients were required to have two multi-phasic ECG-gated CTA with a minimum interval of ≥2 years; (ii) patients with an established clinical diagnosis of Marfan syndrome meeting Ghent criteria, and who had the availability of at least one multi-phasic ECG-gated CTA prior to root repair; and (iii) The third group comprised individuals without thoracic aortic dilation (<4.0 cm) based on guideline definitions, who had the availability of one multi-phasic ECG-gated CT angiogram. We excluded patients across all groups with (i) history of aortitis or intramural haematoma; (ii) prior aortic root or ascending aortic repair; (iii) poor aortic enhancement (<250 Hounsfield units) limiting clear imaging of peak systole and peak diastole; (iv) motion artefact (pulsation/respiratory) limiting accurate segmentation of the ascending aorta; and (v) lack of blood pressure data within 1 month of the CT examination. Baseline patient characteristics including clinical risk factors for aortic disease (e.g. smoking status, valvular disease) and blood pressure data were extracted through medical chart review. In total, medical records and CT angiograms of 551 subjects were screened for eligibility. After applying the above inclusion and exclusion criteria, a total of 96 subjects were included for final analysis, 49 subjects with a diagnosis of a dilated aorta, 38 subjects with a non-dilated aorta, and 16 subjects with Marfan. An outline of the patient selection is illustrated in Figure 1.

Vascular deformation mapping

VDM is a validated image analysis technique, based in deformable image registration, which allows for accurate, 3D assessment of changes in aortic geometry between two different states. VDM was first described as a method to map growth of the thoracic aorta in a patient over time through comparison of aortic geometry between two CTA scans (i.e. at baseline and a follow-up study in the future) to quantify the degree of aortic growth (termed VDM-G for ‘growth’ henceforth), demonstrating superior, submillimetre measurement accuracy compared with expert raters in validation work.^14–17 More recently, the VDM technique has been applied to quantify 3D, pulsatile displacement of the ascending aorta between systole and diastole using multi-phase/dynamic CTA acquisitions (termed VDM-D for ‘dynamic’ henceforth). Specifically, VDM-D can be used to quantify multi-directional (i.e. axial, in-plane) motion of the aortic root caused by left ventricular contraction, in addition to changes in cross-sectional area of the ascending aorta owing to pulse pressure.¹⁸  Figure 2 summarizes the CTA analysis workflow used in this study for both VDM-G and VDM-D.

Metrics derived from VDM-D

VDM-D allows for quantification of the pulsatile 3D displacement of the ascending aorta between diastole and systole, which includes components of aortic root motion (i.e. downward motion related to traction from left ventricular contraction) in addition to pulsatile changes in cross-sectional aorta owing to pulse pressure.

Aortic root motion was extracted from baseline and follow-up CTAs. The analysis plane for extracting root motion metrics was located at the sinuses of Valsalva in a plane perpendicular to the tangential direction of the aortic centreline, at the level of the coronary arteries as this location provides a distinct anatomic feature for deformable registration. Subsequently, a normal vector to this 2D analysis plane is obtained via cross-product of in-plane unit vectors. Aortic root motion was quantified by the following metrics: (i) total displacement, obtained by averaging the 3D displacement field as given by VDM-D over the entire analysis plane; (ii) axial displacement, defined as the projection of the total displacement vector in the direction parallel to the diastolic aortic centreline; and (iii) in-plane displacement, defined as the projection of the total displacement in the direction perpendicular to the diastolic aortic centreline.

Distensibility is a metric that reflects the local aortic stiffness and is defined as the relative change in aortic cross-sectional areas over the cardiac cycle per unit of pressure. In our study, the ratio between diastolic and systolic cross-sectional areas was obtained from multi-phase CTA and distensibility was calculated using the following formula:

where $Asys$ and $Adia$ refer to the systolic and diastolic aortic cross-sectional areas, respectively, and PP represents the pulse pressure (in millimetres of mercury). To examine distensibility gradients across the ascending aorta, we extracted and analysed aortic motion at three distinct locations along the ascending aortic length: (i) the aortic root, at the level of the coronary artery ostia in the sinuses of Valsalva, perpendicular to the aortic centreline (i.e. the same level as root motion extraction); (ii) at the level of the mid-ascending aorta [i.e. mid-point between the sinotubular junction (STJ) and the proximal arch]; and (iii) at the proximal aortic arch, specifically at 2 cm proximal to the base of the innominate artery to avoid analysis plane being affected by a gradual take-off of the artery (Figure 3).

Aortic measurements and aortic growth rate extraction

Maximal aortic diameter measurements covering the ascending aorta and proximal arch were obtained from clinical radiology reports in the following location: sinuses of Valsalva, STJ, mid-ascending aorta, and proximal arch. All aortic measurements at our institution are performed in a standardized fashion in a dedicated 3D laboratory by trained technicians using 3D analysis software (Vitrea 7.14, Vitrea; Cannon Medical Systems) and a curved planar reformat technique. Aortic growth rate was extracted from VDM-G output meshes using Paraview (Kitware Inc.)¹⁹ in two different regions of interest: (i) aortic root: from the aortic annulus to the STJ and (ii) ascending aorta: from the STJ to 2 cm before the base of the innominate artery (Figure 4). Given that VDM-G generates growth rate values at each mesh vertex, the 90th percentile value of growth rate from both aortic regions was used for analysis given that in validation work we have found this value to reliably capture maximal growth while avoiding the effect of outlier values related to image noise.

Statistical analysis

Continuous variables reported as mean ± standard deviation (SD) for normally distributed data. Normality of data was assessed using the Shapiro–Wilk test. χ² test was used for categorical variables. An independent t-test was used to compare the mean differences between two separate groups to assess significant differences. A paired t-test was used to evaluate the mean differences within the same group at two different time points. To compare differences between two independent groups with non-normally distributed data, the Mann–Whitney U test was employed. Multivariable linear regression models were used to identify independent associations between patient characteristics, aortic distensibility (VDM-D), and aortic growth (VDM-G). Statistical analysis was conducted using Python with specific libraries such as SciPy for statistical tests and Pandas for data manipulation.^20,21 To summarize subgroup-specific trends in aortic geometry, ascending aortic motion, and distensibility, we used a statistical shape modelling (SSM) approach to compute the mean aortic geometry and displacement for all individual patient models within each subgroup (i.e. Non-dilated, Dilated, and Marfan). Our SSM technique was adapted from previously published methods and was completed using in-house Python scripts and vascular modelling toolkit.^22,23

Results

Patient demographics

Patient demographics, risk factors, and aortic diameters are detailed in Table 1. The mean age was significantly higher in the Dilated group (60.2 ± 12.7 years) compared with the Non-dilated and Marfan groups (48.3 ± 11.9 and 38.6 ± 15.4 years, respectively, P = <0.001). The mean diastolic blood pressure was higher in Dilated compared with both the Non-dilated and Marfan groups (78 ± 11.8 vs. 70.0 ± 10.9 vs. 66 ± 10.4 mmHg, respectively, P < 0.001), and the mean systolic blood pressure was higher in Dilated compared with Marfan group (130 ± 15.1 vs. 114 ± 12.1 mmHg, P < 0.001). Hypertension (HTN) and hyperlipidaemia (HLD) were more frequent in Dilated group compared with Non-dilated and Marfan groups (HTN—57 vs. 10.5 vs. 6.3%; HLD—43 vs. 14 vs. 6.3%). Bicuspid aortic valve (BAV) (n = 14, 29%) and aortic stenosis (AS) (n = 7, 14%) were observed only in Dilated patients. The mean ascending aortic diameter in Dilated group was significantly larger than Marfan patients at the STJ, mid-ascending, and proximal arch levels; however, sinuses diameters were not significantly different. Marfan patients demonstrated significantly larger mean diameters only at the sinuses and STJ compared with Non-dilated group. No significant differences in sex, smoking status, and pulse pressure between groups were observed. The mean length of interval between baseline and follow-up CT in the Dilated group was 4.6 ± 2.2 years.

Aortic root motion metrics

Table 2 summarizes the aortic root motion metrics and distensibility values at the levels of the aortic root, mid-ascending aorta, and proximal arch for all subgroups. Total displacement was significantly decreased in the Dilated group compared with both the Non-Dilated group (5.27 ± 1.73 mm vs. 6.74 ± 1.93, P < 0.001) and the Marfan group (5.27 ± 1.73 mm vs. 6.99 ± 2.58, P = 0.003). Axial displacement was significantly lower among patients with Dilated compared with the Non-dilated group (3.09 ± 1.76 vs. 4.58 ± 1.72, P < 0.001) and in comparison with the Marfan group (3.09 ± 1.76 vs. 5.33 ± 2.43, P < 0.001), but no significant differences seen for in-plane displacement between groups. Root motion metrics did not change significantly between baseline and follow-up CTA in the Dilated group. Furthermore, we did not observe any significant sex differences in root motion metrics (i.e. total, axial, and in-plane displacement), nor by the presence of BAV, AS, aortic insufficiency (AI), or hypertension.

Regional differences in distensibility

As illustrated in Figure 5, individuals with Dilated demonstrated significantly lower distensibility values at all three ascending aortic locations in comparison with those with a normal aorta. Moreover, although measures of distensibility at the aortic root were comparable between the Dilated and the Marfan groups, Dilated patients demonstrated significantly lower distensibility at the mid-ascending aorta and the proximal arch than those with Marfan.

Within the Marfan group, distensibility was significantly lower at the aortic root compared with the mid-ascending aorta (2.12 × 10⁻³ vs. 3.70 × 10⁻³ mmHg⁻¹, P = 0.004) and the proximal arch (2.12 × 10⁻³ vs. 4.27 × 10⁻³ mmHg⁻¹, P < 0.001). In contrast, the Dilated and Normal aorta groups did not exhibit significant differences in distensibility across the different ascending aortic regions assessed. No significant differences in regional distensibility were observed by sex, BAV, AS, AI, or hypertension.

Distensibility correlations with age and aortic size

Figure 6 illustrates the effect of age on distensibility at the aortic root, the mid-ascending aorta, and the proximal arch across three distinct patient groups. In the Non-dilated subgroup, there was a moderate, negative correlation between age and distensibility at all ascending aortic levels (aortic root R = −0.56, P < 0.01; mid-ascending R = −0.72, P < 0.01; proximal arch R = −0.63, P < 0.01). In contrast, for the Dilated group, a negative correlation between age and distensibility was evident only at the mid-ascending aorta (R = −0.36, P = 0.01) and proximal arch (R = −0.54, P < 0.01); however, no significant correlation was observed at the aortic root (R = −0.11, P = 0.44). Lastly, among the Marfan group, strong negative correlations were seen between distensibility and age at the mid-ascending (R = −0.71, P < 0.01) and proximal arch (R = −0.80, P < 0.01) levels. Further, there was a moderate, negative correlation observed at the root (R = −0.52, P = 0.04) where the degree of dilation was maximal.

Scatterplots demonstrating the associations between aortic diameter and distensibility for all subgroups are depicted in Figure 7. In the Non-dilated subgroup, there was a moderate, negative correlation between diameter at the STJ, mid-ascending aorta, and proximal arch (STJ: R = −0.38, P = 0.02; mid-ascending: R = −0.57, P < 0.01; proximal arch: R = −0.55, P < 0.01). Conversely, among the Dilated and Marfan groups, a weak negative correlation between diameter and distensibility was only evident at the proximal arch (Dilated: R = −0.35, P = 0.01; Marfan: R = −0.49, P = 0.05), and no correlations were observed at the level of the STJ and mid-ascending aorta.

Ascending aortic metrics aortic growth (Dilation subgroup)

Scatterplots of aortic distensibility (baseline scan) vs. growth rate of the ascending aorta during surveillance measured by VDM-G for patients in the Dilation group are depicted in Figure 8. In patients with Dilated, baseline distensibility measured at the mid-ascending aorta demonstrated a moderate and positive correlation with the growth rate of the ascending aorta (R = 0.34, P = 0.02), suggesting that higher distensibility may be associated with higher aortic growth rate. Conversely, distensibility assessed at the aortic root showed no significant correlation with the growth rate of the aortic root (R = 0.20, P = 0.16), indicating a lack of association in this region of the aorta. No significant correlations were observed between motion metrics including axial and in-plane displacement and aortic growth. In a multivariate model adjusting for age and aortic diameter, distensibility of the mid-ascending aorta measured at baseline was found to be independently associated with growth rate of the ascending aorta (β = 0.19, 95% CI 0.007–0.374, P = 0.042).

Longitudinal assessment of aortic motion metrics and distensibility in patients with TAA

Longitudinal assessment of aortic motion metrics in Dilated patients is depicted in Table 3. Patients with Dilated revealed minor reductions in the magnitude of both total and axial displacement at follow-up, with the mean displacement decreasing from 5.27 to 5.11 mm and the mean axial displacement from 3.09 to 2.78 mm; though, differences did not reach statistical significance (P = 0.63 and P = 0.43, respectively). Distensibility at the root, mid-ascending aorta, and proximal arch levels all demonstrated minor decreases in mean values at follow-up; however, similarly these differences were also not statistically significant.

Discussion

In this study, we utilized VDM-D to demonstrate patient-specific differences of regional aortic stiffness and multi-directional ascending aortic motion across time and disease states. The key findings of our study are as follows: (i) 3D assessment of ascending aortic motion by VDM-D demonstrates both expected age-related stiffening and clear regional differences in ascending aortic stiffness between those with and without aortic dilation and between those with sporadic and heritable (i.e. Marfan syndrome) aetiologies. (ii) Total and axial aortic root motion was significantly decreased in the Dilated group compared with other groups. (iii) Aortic stiffness was significantly higher in the Dilated group compared with the Non-dilated group and exhibited a more diffuse pattern of stiffening compared with the Marfan group in which stiffening was localized to the root. (iv) The moderate-to-strong relationships between age and aortic stiffness observed in non-dilated segments were either significantly blunted or absent in dilated segments, implying that wall stiffening owing to the local aortic wall pathology in aortic aneurysm supersedes age-related stiffening. (v) There was an inverse relationship between aortic stiffness and aortic growth at follow-up imaging, suggesting that less stiff (i.e. more compliant) aortas may be at higher risk of significant growth.

Current techniques for measuring aortic stiffness have important limitations in their ability to comprehensively characterize ascending aortic biomechanics. Methods such as PWV provide a more global assessment of aortic stiffness across longer vascular segments but are not capable of capturing local variation or gradients in stiffness.^24,25 Methods for localized assessment of ascending aortic stiffness typically employ either cine MRI or echocardiography to measure cyclic changes in aortic diameter or area, although these techniques are subject to variability in user-defined analysis plane placement, as well as through-plane motion of the aorta over the cardiac cycle owing to downward systolic excursion as a result of LV contraction.^12,26,27 Although these 2D analysis methods have revealed much about the associations between aortic stiffness and a host of patient characteristics and cardiovascular risk factors, only a small number of these studies have focused on patients with ascending aortic dilation and even fewer have provided multi-level assessments of stiffness and/or longitudinal assessments of growth.^27,28 Finally, a large amount of our current understanding of aortic stiffness has been drawn from ex vivo tissue testing of surgical aortic specimens^29–31. Although ex vivo mechanical testing of the aortic tissue clearly provides the most accurate assessment of wall biomechanics, key limitations include the inaccessibility of tissue from patients with pre-surgical aortic size—in whom risk stratification assessments are most pertinent—and the inability to replicate the complex, in vivo motion and physiological conditions of the ascending aorta. These gaps underscore the need for clinically applicable advanced imaging techniques such as VDM which are capable of capturing 3D aortic growth, while integrating in vivo analysis of localized, patient-specific aortic biomechanics to advance understanding of the relationships between aortic wall remodelling and disease progression. Although multi-phase CTA is commonly used in clinical practice (e.g. transcatheter valve planning), it does come with a trade-off of additional radiation exposure. However, our data show that aortic root motion and ascending stiffness metrics do not significantly change over the mean 4.5-year duration of follow-up in our study, suggesting that if such dynamic metrics can be shown to predict risk of progression or other clinically meaningful outcomes, a higher-dose multi-phase CTA could be acquired only once, early in surveillance, and values could likely be reasonably assumed to remain relatively constant for a 5-year period or longer.

An important observation from our study is that aortic stiffness can vary substantially over the length of the ascending aorta, and such stiffness gradients may be important both for understanding disease phenotypes and for informing aetiology-specific measurement locations in future studies of ascending aortic stiffness. Our current understanding of regional aortic stiffness across the ascending aorta largely derives from ex vivo analyses of surgical specimens of aortic tissue,^29,30,32 which have similarly shown variable stiffness along the length of the ascending aorta. Specifically, a prior study that investigated aortic tissue biomechanics in a mixed population of surgical patients with aTAA showed similar stiffness trends to the ones we observed with VDM-D, specifically with elevated stiffness in the root compared with the mid- and distal ascending aorta, with root-specific stiffening most pronounced in heritable aortopathy.³³ However, a key advantage of VDM-D in assessing stiffness ascending aortic stiffness gradients is the ability to gather such multi-level stiffness assessments in vivo from non-invasive imaging, also in a manner that also to allow tracking of location-specific growth over follow-up by VDM-G. With larger data sets, this technique could be leveraged to create nomograms and reference values delineating typical degrees of stiffness based on age, sex, and across different aortic regions. This could be instrumental in understanding a patient’s unique disease state, progressing towards personalized management of thoracic aortic disease.

As expected, we observed significant correlations between aortic diameter and localized stiffness in Non-dilated and Dilated groups (i.e. larger diameter associated with higher stiffness).^34,35 However, these associations were only moderate-to-weak in strength, supporting the fact that that aortic size and the degree of underlying aortic wall remodelling are two separate processes, and thus, characterization of stiffness in both the maximally dilated and adjacent non-dilated (i.e. proximal arch) locations may provide a more complete understanding of an individual’s ascending aortic stiffness state. Also consistent with existing literature, we observed clear correlations between age and stiffness within the Non-dilated group, with stiffness increasing with age.^13,26,27 However, clear age-related stiffening trends were not as pronounced within the Dilated group at levels of maximal dilation (i.e. root and mid-ascending aorta). Overall, these observations suggest that the presence of dilation and accelerated local wall stiffening in these segments supersedes the effects of normal age-related aortic wall remodelling.

Lastly, we observed a weak but significant inverse correlation between mid-ascending aortic stiffness and growth rate (derived from VDM-G) over follow-up in the Dilated subgroup, a relationship that has been described before in heritable aortopathy, but not as clearly in sporadic disease.¹² This observation is contrary to the concept that stiffening of the ascending aorta is a universally deleterious process that promotes growth and complications, and suggests instead that collagen-mediated stiffening may in part be adaptive by increasing aortic wall strength and slowing growth, particularly in early phases of disease such as those patients in our study (i.e. the mean mid-ascending diameter of 44 mm in the Dilated group). Furthermore, the observation that faster growth was observed at smaller aortic diameter suggests a period of lower wall integrity earlier in disease. This finding may be relevant to the well-described phenomenon whereby Type A aortic dissection most commonly occurs in patients with pre-dissection sizes that are either normal or mildly dilated (i.e. aortic size paradox).³⁶

Limitations

This study has several limitations. First, by including only patients with the availability of high-quality, ECG-gated (multi-phase) CT angiograms with a minimum interval between the two scans of >18 months, there is potential for selection bias. Secondly, the single-centre nature of this study does not allow for external validation of our results and our findings may not hold in other institutions with different patient populations and imaging techniques, highlighting the need for future multi-centre studies. Thirdly, the small size of the Marfan group limits out ability to make statistically rigorous inferences in this group and to assess the relationship between stiffness and growth in this population; however, despite small numbers, we did observe expected trends indicating premature aortic stiffening in this group. Lastly, given that VDM-D is acquired from multi-phase CTA data reconstructed over a single heartbeat, we were not able to assess potential beat-to-beat variability in derived metrics, although we were reassured that aortic motion and stiffness metrics did not significantly change over relatively long-term follow-up.

Conclusion

Multi-level stiffness of the ascending aorta can be assessed via analysis of multi-phasic CTA data by VDM-D. Our results demonstrate that ascending aortic stiffening is not a homogeneous process and stiffening tends to increase with degree of local dilation. Expected age-related stiffening trends were noted in borderline or non-dilated aortic segments, although age-related stiffening was less apparent in dilated segments owing to superimposed disease-related wall abnormalities. The spatial heterogeneity of ascending aortic stiffening emphasizes that the location of stiffness measurement needs to be carefully considered in future studies on this topic. Lastly, mid-ascending stiffness showed an inverse relationship with local aortic growth during follow-up, highlighting the complexities of aortic wall remodelling and highlighting the fact that significant growth may occur in smaller and more compliant ascending aortas.

Conflict of interest: N.S.B. is entitled to royalties related to licensure of intellectual property for VDM technique to Imbio Inc. Otherwise, all authors declare freedom of investigation and no conflict of interest is related to the contents of this manuscript.

Consent

This retrospective study was performed as part of an Institutional Review Board approved study and informed consent was waived.

Funding

N.S.B. was supported by the National Institutes of Health (R01HL170059 and R44HL145953). H.J.P. was supported by the Joe D. Morris Professorship, the David Hamilton Fund, the Phil Jenkins Breakthrough Fund, and the Family of Harpreet and Sangeeta Ahluwalia Fund. C.A.F. was supported by the Edward B. Diethrich Professorship, the Bob and Ann Aikens Aortic Grants Program, and the Frankel Cardiovascular Center.

Data availability

Anonymized data are available upon reasonable request to corresponding author.

Lead author biography

Dr Nicasius S. Tjahjadi is a clinical research fellow at the Department of Cardiac Surgery, University of Michigan, Ann Arbor, MI, USA. He also holds a position as a PhD candidate at the Department of Vascular Surgery, Utrecht University Medical Center, The Netherlands. His work focuses on biomechanics of thoracic aortic disease.

References