4D flow CMR analysis comparing patients with anatomically shaped aortic sinus prostheses, tube prostheses and healthy subjects introducing the wall shear stress gradient: a case control study

The established therapy for aneurysms of the aortic bulb and ascending aorta is the implantation of a tubular prosthesis while preserving the aortic valve (valve sparing aortic root replacement, VSARR). Conventional tube prostheses (TP) are known to interfere with the opening of the aortic valve, inducing flow diversions in the aorta and changes in the vessel’s anatomy like kinking, post-prosthetic dilatation, and increased transvalvular pressure differences [1‐4]. Consequently, anatomically pre-shaped sinus prostheses (SP) were developed to mimic the aortic sinus with the goal to preserve near physiological hemodynamic conditions [2, 5‐8]. Yet, the clinical significance of this innovation remains a subject of controversy [9].

Time-resolved 3-dimensional cardiovascular magnetic resonance (CMR) phase contrast imaging (4D Flow CMR) has provided unprecedented insights into hemodynamics of the aorta in vivo, offering visualization of macroscopic blood flow behavior as well as quantification of flow- and velocity-derived parameters [10]. There are several works confirming altered aortic hemodynamics after conventional VSARR [11‐14]. Of note, in a previous study using 4D Flow CMR, we could confirm almost physiological flow patterns in the aortic root of the SP compared to disturbed flow in the TP [15]. Nonetheless, distal to both prostheses types, significantly more secondary flow patterns than in healthy subjects were detected [8, 16]. These secondary flow patterns were only evaluated semi-quantitatively on a Likert scale; the study lacked a quantitative parameter to objectively quantify flow abnormalities. While secondary flow patterns are often used as markers for the presence of pathological hemodynamics or even for a specific pathology, e.g. pulmonary artery hypertension [17, 18], the effect of these secondary flow patterns on pathological changes of the vessel wall architecture and thus the long-term morbidity of patients exhibiting these flow patterns remain a subject of debate.

We hypothesize that an increase in flow disturbances results in pathological, elevated mechanical stress to the vessel wall. A parameter able to quantify these changes is the wall shear stress (WSS). WSS is a force by blood flow acting tangentially through mechano-transduction on the endothelial cells of the vessel wall. Altered values are linked to pathological vessel wall changes, including remodeling of the vessel wall, atherosclerotic plaque induction, and potential aneurysm growth [19‐22]. Recent studies evaluating WSS concentrate on averaged (WSS_avg) and maximum segmented WSS (WSS_seg) values thereby neglecting minimal values. As both, WSS maxima and minima do affect the vessel architecture [23, 24], a comprehensive parameter depicting the gradient between maximum and minimum WSS values potentially provides valuable information to assess hemodynamics and its effects on the vessel wall.

So far, there is only one study assessing WSS in patients who received prostheses with neo-sinus with a relative small number of examined patients [16]. As anatomically pre-shaped SP are specifically designed to re-establish near physiological flow conditions, a comprehensive analysis investigating the interdependencies of pathological flow patterns on WSS and other quantitative flow parameters in SP compared to conventional TP is of fundamental interest, potentially affecting post-procedural patient morbidity and surgical strategies in the long term.

Hence, the aim of this work was to compare quantitative parameters derived from 4D Flow CMR with focus on the newly introduced WSS gradient in patients after implantation of SP to patients who received conventional TP and to age-matched, healthy subjects. In a sub-analysis, the dependency of WSS on the expression of secondary flow patterns was investigated.

This study revealed that physiologically shaped SP seem to have a positive impact on aortic WSS compared to conventional TP. We introduced the WSS gradient as a new quantitative parameter to discriminate between different patient groups. Since the WSS gradient is strongly correlated not only with the expression but also with semi-quantitative grading of secondary flow patterns, it provides further evidence to link disturbed flow conditions to degenerative vessel wall changes.

WSS changes in patients were apparent not only at the level of the prosthesis but also in the course of the vessel, potentially inducing long-term pathological altering of the vessel architecture and increasing post-operative morbidity. Significant differences between both prostheses were found in and adjacent to the prostheses. Even if differences did not reach statistical significance, values in the SP were closer to the physiological values of healthy subjects than in the TP underlining the importance of aortic sinuses for hemodynamics. Comparing both prostheses to healthy subjects, significant differences in the AAo were only found for the TP but not the SP. Albeit comparison of different works is hampered by a lack of standardized protocols regarding the evaluation of WSS, the results in the aortic bulb and the AAo are in concordance with a previous study that assessed WSS in a prosthesis with neo-sinuses of a different vendor using a volumetric approach [16]. Arguably, Gaudino and colleagues also reported significant differences for WSS_avg and maximum WSS in the DAo, which was only partially reproducible in our study. Possible explanations can be found in manufacturer-specific characteristics of the prostheses, various WSS analysis protocols or the small number of patients in each study but remain speculative.

The WSS gradient is a newly introduced parameter, which has not been published before. The concept behind this parameter is to consider not only markedly increased WSS values but also strongly decreased values. While elevated WSS induces dilatation through increased activity of endothelial nitric oxide synthase and degradation of elastic fibers, reduced WSS is known to induce vessel constriction [23, 24]. Consequently, both, maxima and minima are of pathological relevance promoting endothelial dysfunction and resulting in opposing anatomical changes that have a stronger potential to impair hemodynamics than local WSS alterations [24]. However, WSS_avg,max. WSS_seg, and min WSS_seg may only partially portray these changes due to averaging and the exclusion of either peak or minimum values. Both shortcomings are overcome via the WSS gradient. However, one of the disadvantages of 4D flow CMR (e.g. compared to 2D flow CMR), due to the relatively low spatial and temporal resolution, and high signal-to-noise ratio, is an increased inaccuracy in the measurement of maximum and minimum values [25]. We have tried to reduce this drawback by not comparing the maximum and minimum values in individual voxels, but rather in eight regions orientated towards landmarks in the vessel’s geometry (Fig. 1), each grouped over several voxels. While the placement of the regions exhibits a potential source of error, good to excellent inter-reader agreement confirms the applicability of this approach.

Although there were marked differences between patients and healthy subjects when analyzing flow visually, established WSS parameters mostly exhibited insignificant tendencies between groups. However, the visually obvious differences in hemodynamics between groups could be quantified with the WSS gradient that was able to depict significant differences between both patient groups and healthy subjects with good inter-reader reproducibility. In both prostheses, the WSS gradient was increased compared to healthy subjects, especially in the dAAo and the AA. Still, measurements in the SP were closer to those of healthy subjects than in patients with TP. While established WSS parameters already indicated differences between the groups, in most cases these did not reach the significance level. This was only achieved by the WSS gradient, which seems to be more sensitive to differences of altered shear forces acting on the vessel wall than established parameters WSS_avg and max. WSS_seg. The elevated WSS gradient coincides with an increased number of secondary flow patterns already discussed in a previous paper [8]. Secondary flow patterns are a result of regional accelerations and decelerations of flow and velocity as visualized in the scalar maps in Fig. 2. It seems straightforward that these altered hemodynamic conditions directly influence mechanical forces acting on the vessel wall and it has been hypothesized, that secondary flow patterns are not only a consequence of altered aortic geometry, but also induce its development [28]. The correlations between WSS parameters and a previously introduced semi-quantitative grading system of secondary flow patterns [8] provide evidence that an increased number of secondary flow patterns can potentially lead to postoperative aneurysm growth and therefore affect postoperative morbidity. As the WSS gradient was the parameter demonstrating the highest correlation, it may be suitable to quantify and follow-up the effects of disturbed blood flow on the vessel wall. Theoretically, the WSS gradient should be better at identifying vortices characterized by antegrade and retrograde flow in one cross section at the same time in contrast to helical flow. Unfortunately, the study collective was too small to generate the statistical power for such an analysis. Future studies could evaluate the possibility of discrimination between secondary helices and vortices including gradients of other WSS parameters such as axial and circumferential WSS. In both flow patterns there are differently pronounced flow accelerations, which potentially have different effects on the WSS and thus on the vessel wall architecture.

The results for hemodynamic and geometrical quantitative parameters are in concordance with a growing body of literature evaluating patients who underwent VSARR [11‐14, 16, 29]. The significantly increased velocities at the level of the prosthesis most likely indicate the reduced compliance of both, the grafts’ material and the associated limited Windkessel function of the vessel [30]. This is also an explanatory approach for the significantly increased maximum flow in both prostheses during systole. Nonetheless, maximum velocities were lower in the SP, suggesting a limited reservoir function of the artificial sinus. Moreover, the SP allows for formation of vortices in the sinus [15]. Vortices, in general, have been discussed to preserve kinetic energy [17, 18] which may then be released later in the cardiac cycle, partly mimicking a Windkessel effect. Whether the increased maximum area of the aorta in patients is linked to post-prosthetic dilatation or the individual’s pre-existing vessel dilation remains hypothetical. Lower vessel area distal to the prostheses in SP compared to TP may be interpreted in favor of the SP, however, vessel areas in the descending aorta where slightly larger in patients with SP. Since patients were not examined before surgery, it is possible that any differences existed beforehand and were not related to the type of prosthesis.

Overall, these findings add to the ongoing discussion about the role of the aortic sinuses by offering quantitative proof of improved flow conditions in the anatomically pre-shaped SP. Although no differences in the clinical outcome have been reported so far [31], Guzzardi et al. and Bollache et al. have impressively demonstrated the thinning of elastic fibers at regions of altered WSS in the aorta [23, 24]. Hence, an effect on the architecture of the aortic vessel wall is highly probable, but may only become evident over the course of several decades. As VSARR is increasingly used in younger patients to maintain function of the native aortic valve [32, 33], this may promote the manifestation of late post-operative morbidity.

Despite the positive influence on WSS and hemodynamics as described above, several quantitative parameters still show significant differences between patients with SP and healthy subjects. Besides the limited compliance of graft material, the SP is primarily designed to imitate the geometry of the aortic bulb to allow for smooth closure of the aortic valve and perfusion of the coronaries [34, 35]. In the course of the prosthesis and the distal aorta, no measures are taken to improve aortic geometry, resulting in kinking of the prosthesis or at the distal anastomosis as well as post-prosthetic dilatation in both patient groups [8]. To our experience, the post-procedural geometry of the AA substantially affects flow conditions in the DAo [36]. As described in a previous paper, the aortic geometry of our patients, regardless of the prosthesis, differed considerably from healthy subjects [8]. The two patients added to the TP group in this paper also exhibited a more angulated “cubic” aortic anatomy compared to the “round” aortic arch of healthy subjects. This provides a possible explanation for the lack of differences in WSS and other quantitative parameters in the distal aorta. An approach to overcome this limitation is the implantation of so-called 90° prostheses, which attempt to emulate the natural shape of the AAo. Additional analyses, including these prostheses as well as developing methods to quantify the geometry of the aorta, may further help our understanding of post-procedural hemodynamics after VSARR and improve surgical strategies.

Physiologically shaped SP seem to have a positive impact on the mechanical forces affecting the aortic wall compared to straight TP in patients after VSARR. Nonetheless, significant differences in WSS, hemodynamic, and geometrical parameters are still present and call for further optimization of surgical techniques and prostheses characteristics. These may include more compliant material and individually tailored grafts, adaptive to the native aortic anatomy.

It should be noted that although established WSS parameters revealed tendentious differences between patients, these did not reach the significant level. We present the novel parameter “WSS gradient”, which seems to be more sensitive to changes in WSS than established WSS parameters. Moreover, the WSS gradient correlates with the occurrence and strength of secondary flow patterns. It therefore links secondary flow patterns to pathological, degenerative vessel wall changes potentially affecting post-procedural morbidity. However, while the effect of WSS on the vessel wall is well validated on a histopathological level and different pilot-studies link altered WSS to various pathologies, long-term follow-up studies evaluating the clinical impact of the WSS changes in different patient collectives are warranted.