Quantitative normal values of helical flow, flow jets and wall shear stress of healthy volunteers in the ascending aorta

Objectives 4D flow MRI enables quantitative assessment of helical flow. We sought to generate normal values and elucidate changes of helical flow (duration, volume, length, velocities and rotational direction) and flow jet (displacement, flow angle) as well as wall shear stress (WSS). Methods We assessed the temporal helical existence (THEX), maximum helical volume (HVmax), accumulated helical volume (HVacc), accumulated helical volume length (HVLacc), maximum forward velocity (maxVfor), maximum circumferential velocity (maxVcirc), rotational direction (RD) and maximum wall shear stress (WSS) as reported elsewhere using the software tool Bloodline in 86 healthy volunteers (46 females, mean age 41 ± 13 years). Results WSS decreased by 42.1% and maxVfor by 55.7% across age. There was no link between age and gender regarding the other parameters. Conclusion This study provides age-dependent normal values regarding WSS and maxVfor and age- and gender-independent normal values regarding THEX, HVmax, HVacc, HVLacc, RD and maxVcirc. Key Points • 4D flow provides numerous new parameters; therefore, normal values are mandatory. • Wall shear stress decreases over age. • Maximum helical forward velocity decreases over age.


Introduction
The aortic diameter is not a good predictor of complications of aortic dilatation [1,2]. The reason for using it is the law of Laplace, which links the expansion rate with the diameter of a vessel, but the biological genesis of aortic dilatation and dissection is more complex [3]. Other sources underlined that aortic flow abnormalities can contribute to the progression of dilatation [4,5], e.g. in the setting of bicuspid aortic valve (BAV) [6][7][8].
4D flow allows the visualization and measurements of flow parameters like helical flow or analysis of the flow jet and may provide deeper insights into cardiovascular pathologies [9]. 4D flow might deliver new imaging biomarkers for surveillance of patients with aortic pathologies and may help to identify patients at risk for complication of aortic dilatation such as rupture and dissection, and finally to adapt patient management. This study is an extension of previous works, which have shown that absolute quantification of helical flow patterns, e.g. of helical volumes and helical duration, can help to differentiate between physiological and pathological flow patterns [10].
However, those new techniques are still neither an established part of the evaluation of patients with aortic disease, nor part of the decision-making process. Nevertheless, before it comes to analysis of pathologies, it is necessary to identify the normal range of these parameters. Thus, the aim of this study was to utilize 4D flow in the aorta of healthy individuals to generate normal values and elucidate changes of helical flow and flow jet as well as wall shear stress (WSS).

Data analysis
Vessel segmentation, blood flow visualization and pre-processing All processing steps were carried out using the software Bloodline (University of Magdeburg, Germany) [13][14][15][16]. The ascending aorta (ascA) was defined as the volume of the aorta between the aortic valve and the origin of the brachiocephalic trunk and the thoracic aorta as the volume between the aortic valve and the diaphragm.
Bloodline enables the analysis of defined regions, e.g. the ascA, and provides the volume as well as the length of the segmented vessel and visualizes blood flow using timeresolved pathlines [17]. We corrected for phase wraps, eddy currents and background noise as described previously [18,19].

Measurements and flow quantifications
The automatic identification of complex flow patterns utilizes a recently described relative pressure calculation [10,20]. Those parameters were evaluated recently [10]. Additionally, bloodline enables a fully automated, adaptive extraction of timeresolved flow jet information [16] and uses filtering techniques to extract the pathlines within the regions of helical flow to visualize them separately from the pathlines with non-helical flow as described by Köhler et al [17].
The parameter temporal helical existence (TH Ex ) was introduced earlier [10]. This parameter describes the duration of helical flow throughout the cardiac cycle. This parameter was Fig. 1 3D visualization of the evolution of a helix in the ascending aorta during the entire cardiac cycle of a healthy individual (left column) and a patient with BAV (female, 39 years old, right column). (a) At early systole, the helix occupies only the very proximal part of the ascending aorta in die healthy volunteer and a big part in the patient. (b) At midsystole, it has grown in volume and length and reached its maximumthe maximum helical volume (HV max ). (c) At end-systole at 300 ms, most of the helix vanished. (d) The summation of all volumes that contributed to the helix during the entire cardiac cycle is the accumulated helical volume (HV acc ), and its length is the accumulated helical volume length (HVL acc ) > given as an absolute value in milliseconds (ms) and as a relative value in percentage (%) of the cardiac cycle. The size and length of a helix were assessed with different parameters, introduced earlier [10]: 1. The maximum helical volume (HV max ) as an absolute value in millilitres (ml) and the relative value in percent (%) of the volume of the thoracic aorta and of the ascA at a definite timepoint of the cardiac cycle ( Fig. 1 a-c). 2. Furthermore, the accumulated helical volume (HV acc ) was defined as a summation of all volumes that contributed to the helix during the cardiac cycle. This parameter can be described as a roadmap of the helix. It is given as an absolute value in milliliters (ml) or relatively in percent (%) based on the volume of the ascA (Fig. 1 d). 3. According to the HV acc , we assessed the accumulated length of a helix with the accumulated helical volume length (HVL acc ). This parameter describes the length of the roadmap of a helix. It is given as an absolute value in millimetres (mm) or relatively in percent (%) based on the length of the ascA (Fig. 1 d).
Helical flow velocities were assessed as follows: 1. The maximum forward velocity (maxV for ) of helices in metres per second (m/s) describes the maximum helix velocity in downstream direction (Fig. 2 a).

2.
The maximum circumferential velocity (maxV circ ) of helices in m/s describes the maximum cross-sectional helix velocity in circumferential direction (Fig. 2 b). 3. General forward flow velocity of helical and non-helical flows as a "2D parameter" in the ascA in m/s.
The flow jet was defined as suggested by Köhler et al: Aortic flow can be assumed as laminar. It follows the vessel course and the highest velocities-the flow jet-is usually located in the vessel's centre; therefore, the flow jet represents the area with the highest velocities within the parabolic flow profile [16]. Furthermore, we evaluated the angle between the vessel's centreline and the flow jet-the flow jet angle (FJA) [5] in the ascA.
Additionally, we assessed the normalized flow displacement (FD) in peak systole in the ascA as a relative, cross-sectional parameter. It is the relative distance between the fastest pixel in peak systole and the centreline normalized to the vessel radius, where 0% means that the fastest pixel lies on the vessel's centreline and 100% means that the fastest pixel was located at the vessel wall [22].
Furthermore, we assessed the maximum wall shear stress (WSS max ) within the ascA as an absolute value in pascal (Pa) as reported previously [23] (Fig. 3). The maximum circumferential velocity (maxV circ ) in m/s describes the maximum cross-sectional helix velocity in circumferential direction within the axial measuring plane (blue arrow)

Statistical analysis
All results were given as their median values and quartiles (1 st /3 rd ). Statistical analysis was performed using the statistical software package SAS 9.4 (SAS Institute Inc.). Binary outcome variables were investigated in logistic regression analyses with age and gender as potential regressors. For the metrical outcome variables, the dependence of the target variables on age and gender or the interaction of both was checked in type II regression models using the procedure SAS/TRANSREG that automatically includes a Box-Cox transformation of the target variables to obtain approximately normal distributions of the residuals. As there were no significant dependencies on the interaction of age and gender, only the p-values of the tests for the main effects of age and gender are reported here. If the Box-Cox transformed target variable did not show significant dependencies on age in the corresponding two-sided linear model tests at an error level of 5%, then reference ranges for the transformed variables were derived as (mean ± 1.96 · standard deviation) where mean and standard deviation were obtained from the total sample or separately by gender depending on the significance status for the influence of gender in the above regression model. In case of age dependencies, agedependent reference ranges were obtained as (predicted value ± 1.96 × SQRT (residual variance)), where the regression parameters were taken from the above regression model with two regressors if the influence of gender was significant, too, or from a reduced regression model with only age as regressor otherwise. Finally, the mean and the limits of these reference ranges were reverse transformed with the inverse Box-Cox transformation yielding estimates for the median and the limits of the reference range of the original variables. Variables with a range including negative values were shifted before the above-described procedure with a corresponding back shift of the reference limits and the centre at the end. The obtained reference ranges should cover approximately 95% of the healthy population.

Volunteer characteristics
Eighty-six volunteers were included (46 females, mean age 41 ± 13 years). Mean body mass index was 24.1 ± 5.2 kg/m 2 with no significant differences between the age groups. The mean resting heart rate during the examination was 69 ± 16 beats per minute, the mean cardiac output was 4.4 ± 1.2 l/min.
We found a good positive correlation between age and aortic volume (R = 0.629, p = 0.002) and aortic diameter (R = 0.637, p = 0.009), indicating a steady growth of ascA with increasing age (Fig. 4 and Table 1). The correlation coefficient R between the aortic volume and gender was 0.406 (p = 0.003), revealing a bigger aorta in men.
The median length of the thoracic aorta was 390mm (260/ 420). There was no significant correlation between the aortic

Presence of helices in the ascending aorta
We detected at least one helix in all 86 participants, only one within the ascA in 70/86 (81%) subjects. The helix started right downstream from the aortic valve and evolved through the ascA towards the aortic arch. Helices in the early systole were small (Fig. 1a), while they reach their maximum volume and length around mid-systole (Fig. 1b). There was no strong link between TH Ex and age (R = 0.366) or gender (R = 0.259) ( Table 2).
Maximum (HV max ) and accumulated (HV acc ) helical volumes The median absolute HV max was 8.4 ml (4.8/13.4), which was 6.0% (4.9/12.3) of the volume of the thoracic aorta and 14.5% (9.1/28.2) of the volume of the ascA. We found a weak correlation between the measurements and age (R = 0.301; p = 0.106) and gender (R = 0.308; p = 0.048) in both vessel sections. The median absolute HV acc was 9.3 ml (4.8/13.4), which was 7.6% (4.9/12.3) of the volume of the thoracic aorta and 19.4% (12.1/23.1) of the volume of the ascA. Although there was a mediocre correlation between the measurements and age, it was not statistically significant (R = 0.401; p = 0.101). Additionally, there was a weak correlation between the measurements and gender (R = 0.040) in both vessel sections (Table 3). Accumulated helical volume length (HVL acc ) of helices in the ascending aorta The median absolute HVL acc overall was 58 mm (31/61), which was 16.3% (11.6/21.5) of the length of the thoracic aorta and 56.6% (32.9/75.2) of the length of the ascA. There was a weak correlation between the measurements and age (R = 0.269; p = 0.022) and no correlation for gender (p = 0.096; p = 0.573) in both vessel sections (Table 3).

Rotational direction (RD) of helices in the ascending aorta
In 61 (71%) cases, helical flow was mainly right-handed, and in 25 cases, we found mainly left-handed helical flow. There was no significant correlation between RD and age (R = 0.228, p = 0.748 and R = 0.212, p = 0.840) or RD and gender (R = 0.241, p = 0.392 and R = 0.225, p = 0.572).

Wall shear stress (WSS)
The median maximum WSS in the ascA was 0.61Pa (0.52/ 1.01). There was a significant (R = − 0.514, p = 0.039) negative correlation between WSS and age, but no link between WSS and gender (R = 0.198, p = 0.320) (

Discussion
Numerous 4D flow parameters were already described [19,[24][25][26]. These parameters were mostly assessed visually only, which might limit the use in the clinical routine. We are convinced that quantitative parameters, which can be assessed in a highly standardized manner, might be one way to overcome this limitation. We utilized 4D flow to establish normal values regarding helical flow, the flow jet and WSS in the ascA in a large cohort. In line with other studies, we found that the ascA diameter increased with age [3,27]: Our calculations regarding the normal range of the ascending aortic diameter indicate an upper limit > 40 mm in volunteers > 51 years, which could be interpreted as an aortic ectasia [1,2]. Since we only included healthy subjects in this study, one possible conclusion-in line with other sources-could be that the aortic diameter alone is a poor parameter for the evaluation of aortic diseases [1,3,28].
Furthermore, we were able to detect at least one helix in the ascA in all participants. Additionally, we found a second helix in the aortic arch in 16 (19%) participants. Interestingly, the aortic arch helices started right at the spot where the ascending aortic helices ended. We state that this sharp separation of the helices is most likely artificial due to the used technique of pressure-based helix extraction. However, it shows that there seems to be a "transition-zone" in between those two helices, right at the transition from the ascA to the aortic arch. In other words: There are differences regarding pressure and flow characteristics between the ascA and the aortic arch, which is in accordance with the literature: Frydrychowicz et al found significant differences in the distribution of WSS between the ascA and the aortic arch. The authors stated that this is helpful for explaining why atherosclerotic lesions predominantly develop and progress mainly at the origins of the supra-aortic vessels [29]. Other sources underlined that the flow abnormalities in the setting of BAV might extend into the aortic arch (Fig. 1), which fits our observation [6].
We found out that in the ascA, helices mainly occur during systole and vanish during diastole. This is in line with the literature: Kilner et al [30] described "spiral" flow predominantly occurring during systole. Other studies investigated the evolution of helical flow by visual analyses and demonstrated that helical flow emerges mainly in peak systole [31].
Additionally, helices in the early systole were usually small, while they reached their maximum volume and length around mid-systole. We assume that some small helices were missed due to the overlay of pathlines with laminar and helical flow, while (semi)automatic analyses might enable the detection of even small helices.
We found a TH EX of 25.7% to be normal; previously, it has been shown that TH EX in BAV patients can be more than 2fold elevated [10], indicating that TH EX can be helpful discriminating normal and abnormal flow.
This current study extends previous attempts to utilize the technique of pressure-based helix extraction to measure volumes and lengths of helices. We found a HV acc of up to 39.6ml to be normal, whereas a recent study found a HV acc of up to 236ml in BAV patients [10].
The helical maximum flow velocities within a helix in healthy volunteers were generally lower than the general "2D peak flow velocities". We found a reduction of helical forward velocity (V for ) by 55.7% with age (18 to 63 years), while there was no link between age or gender and the circumferential velocity (V circ ), and we found a general decrease of peak flow velocity by 29.8% in the ascA. One possible explanation for this finding could be the ascA diameter increases causing a slower laminar and non-laminar flow in general. This is partially in line with the results of van Ooij et al [32]. They investigated the maximum velocity of bulk blood flow and demonstrated decreasing velocities with age. Although they did not investigate helical blood flow, their findings fit our findings, indicating decreasing forward velocities with age. Contrarily, we found no correlation between the helices' circumferential velocity (V circ ) and age or gender. To the best of our knowledge, there is no study that has systematically evaluated the impact of circumferential helical flow velocities, yet.
In line with Lorenz et al, we found both right-and lefthanded helical flows to be normal in healthy volunteers, suggesting that this parameter cannot distinguish between healthy and pathological flows [33].
Dyverfeldt et al stated that the normalized flow displacement (FD) is a suitable parameter for identifying and risk-stratifying patients who are likely to develop clinically significant aortic dilation, but their study did not include healthy volunteers [34]. Later, Sigovan et al compared patients with BAV and seven healthy volunteers and found the normalized flow displacement (FD) to be of importance for distinguishing between physiological and pathological flows [35]. We found a normalized FD of 0.02 to be normal; they found a mean normalized flow displacement of 0.12 in patients while Dux-Santoy et al found a normalized displacement of 0.05-0.08 in patients with BAV [6]. This suggests that the parameter "normalized displacement" can distinguish between physiological and pathological flows. As expected, we found that WSS was closely related to forward velocity: Both parameters demonstrated a negative correlation  [32,36]. Additionally, they elucidated WSS in patients with BAV and found a good correlation between abnormal velocities and WSS. They found WSS of 0.8 to be normal for healthy adults < 30 years; in our study, we found 0.88 Pa to be normal for this age group (the differences between those values occur due to slightly different scan parameters). They found significantly elevated WSS in BAV patients using matched WSS maps [32,36]; therefore, WSS could potentially be an important marker for aortic dilatation. It is known that WSS measurements depend strongly on the used scan parameters [26,37]. This indicates that the here provided normal values only apply when using the same parameters as reported.
One limitation of our study is that we did not differentiate between helices and vortices. A visual, qualitative but also a quantitative differentiation between both is often not possible. There is a smooth transition between both phenomena and definite cutoffs do not exist. Nevertheless, this differentiation could be of interest because there is evidence that vortical flow is a relevant factor, e.g. in aortic dissection, and should be addressed in future research [38].
In conclusion, we demonstrated in 86 healthy volunteers that strong correlations exist between age and the hemodynamic parameters helical forward velocity and WSS at peak systole. Interestingly, the "spatio-temporal helical parameters" like volume, length and temporal existence did not depend on gender and age. We elucidated that the parameters TH EX , the helical volumes and normalized displacement enable to discriminate between physiological and pathological flows. We provided normal ranges for all these flow parameters, which might be an important presupposition for the assessment of patients with aortic disease.

Declarations
Guarantor The scientific guarantor of this publication is Prof. Dr. Matthias Gutberlet.

Conflict of interest
The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.
Statistics and biometry One of the authors has significant statistical expertise.
Informed consent Written informed consent was waived by the Institutional Review Board.
Ethical approval Institutional Review Board approval was obtained.

Methodology
• prospective • diagnostic or prognostic study • performed at one institution Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.