The findings of this study demonstrate that pediatric MFS patients have significantly lower WSS
mean in the pAAo and pDAo than age-matched healthy controls. These regions in the aorta correspond to the locations where aortic dissection and aortic rupture often originate [
2,
4,
5]. The subgroup of MFS patients with already dilated pDAo showed more prominent local vortex flow patterns and markedly lower WSS
mean in the dilated pDAo region, when compared to non-dilated pDAo MFS patients and healthy controls. Furthermore, an inverse relationship between regional aortic diameter and WSS
mean was observed in the entire study group.
Aortic dilatation in MFS patients may start in early childhood, warranting close clinical follow up of the diameters in the entire thoracic aorta from the time of diagnosis. The average aortic size and the heterogeneous involvement of the pDAo regarding dilatation found in this study are in concordance with previous studies [
27,
28]. Although MFS is known to be prone to ascending aortic complications, a substantial part of the aortic dissections in MFS occur in the pDAo before reaching surgical cut-off levels. European studies reported the pDAo to be the site of the first aortic complications in 16–18% of the MFS patients [
2,
29]. This proportion will increase after elective surgery of the AAo [
4,
5]. Changes in WSS in the aorta caused by changes in flow may be the key to better understand aortic disease development in MFS.
Aortic flow patterns and wall shear stress
Mechanisms leading to progressive aortic dilatation and elongation are not fully understood. Recently, aortic flow and derived hemodynamic factors from CMR, including WSS, have been associated with focal evidence of molecular and architectural tissue dysfunction in patients with bicuspid aortic valve and aortopathy [
15]. This study confirms that children and adolescent MFS patients already have disturbed aortic flow patterns in the thoracic aorta. The regions of prominent flow alterations were the aortic root and pDAo, with a significantly higher prevalence and strength of vortex flow patterns found in the pDAo in MFS compared to healthy subjects. Although it is known that small vortex flow patterns in the pDAo sometimes occur in normal aortas with an aortic spindle (i.e. a normal variation in the aortic contour not to be confused with an aneurysm) [
30], the high prevalence and extensive nature found in these young MFS patients in this study is remarkable and consistent with observations by two previous CMR studies in MFS, with prevalence rates ranging from 54 to 70% [
8,
31]. Local aortic geometry seems to be related to the formation of these flow phenomena as the MFS subjects with most dilated pDAo segments (Z-score ≥ 2.0) showed the most extensive focal vortex flow patterns in this study.
Several studies previously demonstrated that visibly pathologic flow patterns could lead to measurable alterations in WSS [
32,
33]. Interestingly, the observed vortex flow patterns in the pDAo in MFS patients in this study corresponded to the reduced WSS
mean observed in the same segment. A similar relationship has been described in a longitudinal follow-up study of MFS patients after root replacement where vortex flow disturbances within the pDAo segment together with reduced local WSS measures developed [
34]. In addition to the lower WSS
mean in the pDAo segment, considerable differences in WSS
mean were observed in the pAAo outer segment of the MFS patients compared to healthy subjects. However, no obvious flow pattern differences in the AAo were detected between MFS patients and the controls, except for local helix formation originating from the aortic root into the proximal AAo in four MFS patients. The significant difference in AAo size between the MFS and healthy controls could have contributed to this reduced regional WSS
mean.
Few studies have investigated aortic WSS in MFS disease cohorts so far [
9,
10] and comparable investigations on aortic WSS were conducted in similar patients groups with dilative aortopathy of unknown origin (i.e. non-bicuspid valve related diseases) [
32,
35]. As in MFS patients, these patients have in common that the AAo is more or less dilated and these studies revealed similar aortic WSS results. In general, lower systolic WSS values were found along the dilated thoracic aorta in both these patient groups, as provided by 2D planar WSS quantification methods [
9,
32,
35]. Similar to our findings, the specific locations of reduced systolic WSS were found at the outer curvature of the pAAo in most of these studies [
9,
35]. Contrary to our results, one of the WSS studies in pediatric MFS patients showed inhomogeneous but relatively higher peak systolic WSS values at focal points on the vessel lumen circumference in the AAo compared to a healthy control group [
10]. Although this is inconsistent to our findings, the higher WSS found in that study might be explained by the fact that the MFS patients were younger and had similar or even significantly smaller indexed aortic dimensions compared to the older healthy control group. Indeed, both factors (i.e. younger age and smaller vessel dimensions) have been proven to be associated with relatively higher WSS measures [
36,
37]. The pDAo WSS in the MFS cohorts from these 4D flow CMR studies was equal or lower in MFS, but statistically similar compared to the WSS of healthy controls [
9,
10]. It should be noted that the aortic WSS measurements in these comparative studies were all based on 2D planar WSS quantification methods, which means that WSS measurements are derived from single 2D cross-sectional planes along the thoracic aorta.
In this study, we used a novel approach providing 3D WSS along the entire aortic wall that was recently developed [
25,
26]. The major advantage of this 3D WSS approach is that the entire aorta is included rather than sampling only a fraction of the vessel wall, thus reducing the chance of missing important regional variations in WSS. Moreover, it allows regional comparison to control references and no manual placement of 2D cross-sectional planes is needed, which is more subject to observer variability. In addition, the estimation of 3D WSS over a certain regional vessel surface enables to study relationships between regional hemodynamics and geometric measures, for instance aortic diameter or diameter changes between vessel segments.
Aortic WSS and diameter
The findings of our study revealed a close interaction between altered aortic diameter and regional WSS
mean in MFS patients and healthy subjects. First, an inverse relationship between WSS
mean and aortic diameter was observed, which is congruent with existing literature [
35,
36,
38]. The most obvious correlation was found for the pDAo segment, where both diameter and diameter change between the aortic isthmus and the largest pDAo dimension were inversely proportional to WSS
mean. Furthermore, subgroup analysis of MFS patients with a dilated pDAo showed the lowest WSS
mean in the corresponding region. Second, the dilation of the aortic root and STJ in MFS patients may have resulted in reduced WSS
mean values in the pAAo segment and might explain the WSS
mean asymmetry between the inner and outer pAAo segments. Such WSS asymmetry was also observed in the pAAo segment in another 4D flow CMR study on MFS patients [
9]. Moreover, an
in-vitro computational fluid dynamic (CFD) investigation about the effect of morphological changes in the aortic root on flow hemodynamics supports these findings. That CFD study, which was conducted with MFS-specific aortic root models of increasing size, showed that progressive root dilatation resulted in a different orientation of the blood stream jet in the aortic root and pAAo [
39]. It is conceivable that a change in flow displacement from the sinus side towards the commissure side has resulted in the altered distribution of WSS
mean that was observed in the pAAo in the MFS patients of this study.
Clinical perspectives
Current management strategies for preventive aortic replacement in MFS are based on aortic diameters or rate of dilatation. However, this strategy remains imperfect and has led to a search for additional markers of disease severity and predictors of adverse vascular events, such as hemodynamic parameters. The results of this study showed already hemodynamic differences in young MFS patients compared to age-appropriate healthy subjects and determined close relationships between regional aortic geometry, flow patterns and 3D WSS over certain vessel regions. This relationship was most evident for the pDAo. Whether these hemodynamic parameters (i.e. flow and WSS) may add in the prediction of progressive dilatation, adverse events or optimization of management strategies in future in addition to the current diameter measurements is unknown and requires further longitudinal follow-up. In this regard and in line with our findings, a recent longitudinal 4D flow follow-up study on MFS patients after preventive valve-sparing aortic root replacement reported the occurrence of helical and vortex flow pattern formation along the inner curvature of the pDAo in 50% of the patients [
34]. Additionally, in that report, a single case was described with newly derived regional vortex formation in the pDAo and altered WSS in that region prior to the development of an intramural hematoma with subsequent aortic dissection type B in the pDAo. This occurred 5 years after root replacement and the changes in flow and WSS parameters are noteworthy and suggest that these factors may play a role in the onset of the adverse event [
34]. Together with our results, these observations show the clinical potential and future perspectives of comprehensive aortic flow hemodynamics assessments and monitoring of longitudinal changes by 4D flow imaging for optimization of preventive decision making. Based on these and our data, one might speculate that at least the MFS patients with pDAo Z-score ≥ 2.0, abnormal vortex flow patterns and decreased local WSS
mean, are at higher risk for adverse cardiovascular events. Additional longitudinal studies are warranted to further elucidate the diagnostic value of inter-individual differences in MFS hemodynamics and the applicability of 4D flow parameters in prediction models.
Study limitations
The current study is subject to several limitations. In this observational study there might be some selection bias. Although MFS patients underwent CMR as a regular part of their lifelong clinical follow-up according to the MFS imaging guidelines, we cannot completely exclude the possibility that some of them received MR imaging based on previously abnormal findings on echocardiographic imaging. These results therefore may not be generalizable. However, it is known that the MFS patients individually can show a large heterogeneity in the involvement of heart and cardiovascular problems, e.g. the onset and the progression of disease over time.
To account for patient age and body size for aortic dimensions comparison in children, aortic Z-scores were calculated using EchoIMS. While EchoIMS provides ultrasound-derived normative data, there is currently no large pediatric CMR-specific database and this method of normalization is the best available alternative, is used in clinical practice and has been applied in previous CMR studies [
22,
23].
Finally, potential sources of error, including the limited spatial resolution of 4D flow CMR, could have resulted in underestimation of 3D WSS
mean measurements. However, as both patient and healthy control groups underwent MR imaging at the same scanners with similar scan parameters, relative differences in WSS
mean between the two groups remain useful. Furthermore, 3D segmentation of the aorta based on the time-averaged systolic portion of the cardiac cycle is a potential source of error in this 4D flow study. However, averaging WSS over five systolic phases of the cardiac cycle with our algorithm, and using a well-defined segmentation protocol have been proven to produce low variability [
26].