Four-dimensional MR Evaluation of Ascending Aortic Systolic Flow Patterns

Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights. ORIGINAL RESEARCH n CARDIAC IM AGING Radiology: Volume 255: Number 1—April 2010 n radiology.rsna.org 53 Bicuspid Aortic Valve: Four-dimensional MR Evaluation of Ascending Aortic Systolic Flow Patterns 1 Michael D. Hope , MD Thomas A. Hope , MD Alison K. Meadows , MD , PhD Karen G. Ordovas , MD Thomas H. Urbania , MD Marcus T. Alley , PhD Charles B. Higgins , MD Purpose: To use time-resolved three-dimensional phase-contrast magnetic resonance (MR) imaging, also called four- dimensional fl ow MR imaging, to evaluate systolic blood fl ow patterns in the ascending aorta that may predispose patients with a bicuspid aortic valve (BAV) to aneurysm. Materials and Methods: The HIPAA-compliant protocol received institutional review board approval, and informed consent was obtained. Four- dimensional fl ow MR imaging was used to assess blood fl ow in the thoracic aorta of 53 individuals: 20 patients with a BAV, 25 patients with a tricuspid aortic valve (TAV), and eight healthy volunteers. The Fisher exact test was used to evaluate the signifi cance of fl ow pattern differences. Results: Nested helical fl ow was seen at peak systole in the ascend- ing aorta of 15 of 20 patients with a BAV but in none of the healthy volunteers or patients with a TAV. This fl ow pattern was seen both in patients with a BAV with a di- lated ascending aorta ( n = 6) and in those with a normal ascending aorta ( n = 9), was seen in the absence of aortic stenosis ( n = 5), and was associated with eccentric sys- tolic fl ow jets in all cases. Fusion of right and left leafl ets gave rise to right-handed helical fl ow and right-anterior fl ow jets ( n = 11), whereas right and noncoronary fusion gave rise to left-handed helical fl ow with left-posterior fl ow jets ( n = 4). Conclusion: Four-dimensional fl ow MR imaging showed abnormal heli- cal systolic fl ow in the ascending aorta of patients with a BAV, including those without aneurysm or aortic stenosis. Identifi cation and characterization of eccentric fl ow jets in these patients may help identify those at risk for develop- ment of ascending aortic aneurysm. q RSNA, 2010 1 From the Department of Radiology, University of California San Francisco, 505 Parnassus Ave, Box 0628, San Francisco, CA 94143-0628 (M.D.H., T.A.H., A.K.M., K.G.O., T.H.U., C.B.H.); and Department of Radiology, Stanford University School of Medicine, Stanford, Calif (M.T.A.). Received August 4, 2009; revision requested September 17; revision received September 30; accepted October 15; fi nal version accepted October 28. Supported by a 2008 RSNA Research Resident Grant. Address correspondence to M.D.H. (e-mail: michael.hope@radiology.ucsf.edu ). q RSNA, 2010 54 radiology.rsna.org n Radiology: Volume 255: Number 1—April 2010 CARDIAC IMAGING: Bicuspid Aortic Valve: Systolic Aortic Flow Patterns Hope et al tocol was approved by the institutional review board of the University of Cali- fornia San Francisco, and informed consent was obtained from all partici- pants. Time-resolved 3D phase-contrast MR imaging, or 4D fl ow MR imaging, was used to assess thoracic aortic blood fl ow in 53 individuals: 20 patients with a BAV, 25 patients with a TAV (see Table 1 for patient characteristics), and eight healthy volunteers (age, 30.8 years 6 5.2; one woman, seven men). The 4D fl ow MR imaging technique used has been previously validated ( 19,20 ). Imaging was performed on a 1.5-T system (Signa CV/i, GE Health- care) ( G max , 40 mT/m; rise time, 268 mi- croseconds) by using a radiofrequency- spoiled gradient-echo pulse sequence (repetition time msec/echo time msec, 4.6–5/1.7–2; fl ip angle, 15°; velocity encoding, 160–200 cm/sec; fractional fi eld of view, 300 3 270 mm; slab thick- ness, 78 mm; matrix, 256 3 192 3 30; spatial resolution, 1.17 3 1.56 3 2.60 mm; temporal resolution, 74–77 millseconds) and an oblique-sagittal slab encompass- ing the thoracic aorta. Imaging was per- formed with an eight-channel cardiac coil, respiratory compensation, and ret- rospective electrocardiographic gating. Parallel imaging with an acceleration factor of two was used. A total of 735 heartbeats were required for data ac- quisition, resulting in imaging times of tients with a BAV ( 10 ). Furthermore, investigators ( 11,12 ) have shown that the specifi c segments of dilated aorta vary with the type of aortic valve leafl et fusion. This is diffi cult to explain with a theory of intrinsic wall abnormality alone and raises the possibility that ab- normal fl ow patterns arising from dif- ferent aortic valve morphologies may account for variations in segmental aor- tic dilation in patients with a BAV. Time-resolved three-dimensional (3D) phase-contrast magnetic resonance (MR) imaging (also know as four-dimensional [4D] fl ow MR imaging) is well suited for evaluation of multidirectional blood fl ow velocity data in the thoracic aorta. Characterization of normal aortic fl ow patterns with the 4D technique has been shown to agree with that with two-dimensional and 3D phase-contrast techniques ( 13–18 ). Additionally, 4D fl ow MR imaging allows characterization of abnormal secondary blood fl ow pat- terns that are not well visualized with other techniques. For example, 4D fl ow MR imaging has been used to uncover abnormal helical- and vortical-type fl ow in aneurysmal ascending thoracic aortas ( 17,18 ). In this study, we used 4D fl ow MR imaging to evaluate systolic fl ow in the ascending aorta of patients with a BAV with the goal of uncovering abnor- mal fl ow patterns that may predispose this patient population to aneurysms. Materials and Methods M.T.A. receives research funding from GE Healthcare (Milwaukee, Wis). The Health Insurance Portability and Accountability Act–compliant pro- A bicuspid aortic valve (BAV) is the most common congenital heart defect and may account for more morbidity and mortality than all other congenital cardiac malformations com- bined ( 1 ). Patients with a BAV frequently have dilated ascending thoracic aortas compared with healthy subjects with a tricuspid aortic valve (TAV), even when matched for the degree of aortic stenosis and regurgitation ( 2,3 ). Some investiga- tors ( 3,4 ) have used this data to discount the role that altered hemodynamics play in ascending aortic dilation, arguing that because dilation is out of proportion to the degree of stenosis and regurgitation, it must be secondary to nonhemodynam- ic factors in patients with a BAV. Patho- logic evidence has been used to support an alternative theory for aortic dilation, in which it is postulated that a genetic or developmental abnormality in the proxi- mal aortic tissue leads to weakness of the aortic wall with a BAV ( 5–7 ). There are some hemodynamic and anatomic observations, however, that are at odds with the theory that an in- trinsic wall abnormality explains aortic dilation. Echocardiographic evaluation has shown signifi cantly higher peak aortic velocities and skewing of peak ve- locities toward the anterolateral aspect of the ascending aorta in patients with a BAV compared with matched healthy subjects with a TAV ( 8,9 ). The region of peripherally elevated peak velocities correlates with the typical location of asymmetric aortic dilation seen in pa- Implications for Patient Care Evaluation of eccentric systolic n fl ow jets in the ascending aorta in patients with a BAV can be achieved in a single free- breathing acquisition. Identifi cation and characteriza- n tion of eccentric fl ow jets in the ascending aorta in patients with a BAV may help identify those at risk for development of an ascending aortic aneurysm. Advances in Knowledge Time-resolved three-dimensional n phase-contrast MR imaging, or four-dimensional fl ow MR imag- ing, allows visualization of eccen- tric systolic fl ow jets in the ascending thoracic aorta that have not been well characterized with other techniques. Four-dimensional fl ow MR imag- n ing shows markedly abnormal helical systolic fl ow in the ascending aorta of patients with a bicuspid aortic valve (BAV), including those without aneu- rysm or aortic stenosis. Published online 10.1148/radiol.09091437 Radiology 2010; 255:53–61 Abbreviations: BAV = bicuspid aortic valve 4D = four-dimensional 3D = three-dimensional TAV = tricuspid aortic valve Author contributions: Guarantors of integrity of entire study, M.D.H., C.B.H.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of fi nal version of submitted manu- script, all authors; literature research, M.D.H., A.K.M., C.B.H.; clinical studies, all authors; statistical analysis, M.D.H.; and manuscript editing, M.D.H., T.A.H., A.K.M., T.H.U., C.B.H. See Materials and Methods for pertinent disclosures. Radiology: Volume 255: Number 1—April 2010 n radiology.rsna.org 55 CARDIAC IMAGING: Bicuspid Aortic Valve: Systolic Aortic Flow Patterns Hope et al Table 1 Patient Population Characteristics Characteristic Patients with a BAV ( n = 20) Patients with a TAV ( n = 25) No. of women 8 10 Age (y) * 23.5 6 11.3 29.4 6 19.7 Aortic coarctation 16 14 Previously repaired 15 13 Tetralogy of Fallot 1 7 Ascending aorta maximum diameter † Normal 13 21 Dilated 7 ‡ 4 Aortic insuffi ciency Mild 5 2 Moderate or severe 1 … Aortic stenosis Mild 5 … Moderate or severe 5 … Note.—Unless otherwise specifi ed, data are numbers of patients. * Data are means 6 standard deviations. † Normal is , 4 cm or , 2.2 cm/m 2 , dilated is � 4 cm or � 2.2 cm/m 2 . ‡ One patient had a maximum aortic diameter . 5 cm. Figure 1 Figure 1: Normal systolic fl ow in a patient with a TAV and normal thoracic aorta dimensions. (a) Four-dimensional fl ow MR imaging data in an oblique-sagittal orientation with 3D stream- lines (color-coded for velocity, see key in b ) during peak systole. Left: from right side of thoracic aorta. Right: from left side of thoracic aorta. Note smooth trajectory and absence of substantial secondary fl ow features. (b) Close-up of area in white box in a . (c) Vector analysis at sinotubular junction during peak systole. Left: cross-sectional depiction. Right: sagittal depiciton. Note relatively central velocity profi le. Ant = anterior. 8–15 minutes (mean, 11 minutes). Four- dimensional fl ow MR imaging was per- formed after standard cardiac MR imag- ing in all patients. Clinical indications for MR evaluation were aortic coarctation ( n = 30), Tetralogy of Fallot ( n = 8), and ascending aortic aneurysm ( n = 7). Prior to visualization, data were corrected for Maxwell phase effects, encoding errors due to the gradient fi eld distortions, and effects from eddy currents ( 21–23 ). Corrected velocity data were im- ported into 3D visualization software (EnSight; CEI, Apex, NC), which enabled the dynamic visualization of complex 4D datasets by providing a variety of data manipulation tools, including two- dimensional velocity vector fi elds mapped onto planes of interest, 3D streamlines, and particle traces. Streamlines are imagi- nary lines that are aligned with the local velocity vector fi eld at a given moment in time and provide a 3D perspective of 56 radiology.rsna.org n Radiology: Volume 255: Number 1—April 2010 CARDIAC IMAGING: Bicuspid Aortic Valve: Systolic Aortic Flow Patterns Hope et al able and determination of the fusion pattern was challenging; two of these cases were excluded from our study be- cause the fusion pattern could not be reliably determined. Degree of aortic regurgitation and stenosis was assessed by using echocardiography and was cat- egorized as nonexistent or trace, mild, or moderate or severe. Dimensions of the thoracic aorta were measured by using MR images in orthogonal planes at the levels of the sinuses of Valsalva and sinotubular junction; dimensions of the middle ascending aorta, at the level of the right pulmonary artery; and dimensions of the aortic arch, immedi- ately proximal to the innominate artery. On the basis of their maximum ascend- ing aorta diameter, participants were placed into one of two groups: normal ( , 4 cm or , 2.2 cm/m 2 ) or dilated ( � 4 cm or � 2.2 cm/m 2 ) ( 27–29 ). Nested helical systolic fl ow was de- fi ned as greater than 180° curvature of the majority of high velocity peak sys- tolic streamlines around slower central helical fl ow in the ascending thoracic aorta and was characterized as either right- or left-handed. Vector analysis was performed at peak systole by us- ing an orthogonal plane at the level of the sinotubular junction, and eccentric fl ow jets were defi ned as predominantly peripheral high-velocity vectors and/or clustering of high-velocity vectors away from midline into one of four quad- rants: right-anterior, right-posterior, left- anterior, and left-posterior. Flow char- acterization was performed by two authors (M.D.H. and T.A.H., with 7 and 6 years experience with 4D fl ow MR imaging, respectively). A Fisher exact test was used to evaluate the statistical signifi cance of fl ow pattern differences. Results Patients with Normal Ascending Aorta Diameter Four-dimensional fl ow MR imaging evalu- ation of the ascending thoracic aorta during peak systole revealed no rel- evant secondary fl ow features in any of the healthy volunteers ( n = 8) or patients with a TAV and normal aortic Figure 2 Figure 2: Images in a patient with a BAV and a focal ascending aortic aneurysm. (a) MR angiographic and (b) T1-weighted spin-echo MR images show focal aneurysm of proximal ascending aorta (up to 4.8 cm). Incidental note is made of mild proximal descending aorta narrowing at site of coarctation repair. Streamline analysis from (c) right and (d) left sides of thoracic aorta shows dramatic systolic right-handed helical fl ow in aortic root. instantaneous blood fl ow. Particle traces integrate fl ow over time and, thus, in- corporate the temporal evolution of velocities ( 24,25 ). Data Collection and Analysis Methods The number of aortic valve leafl ets and the type of leafl et fusion in patients with a BAV were determined with transthoracic echocardiography and/or steady-state free-precession cine MR imaging (Fiesta; GE Healthcare) ( 26 ). One author (C.B.H., with over 25 years experience in cardiac imaging) evaluated four cases for which only limited visu- alization of the aortic valve was avail- Radiology: Volume 255: Number 1—April 2010 n radiology.rsna.org 57 CARDIAC IMAGING: Bicuspid Aortic Valve: Systolic Aortic Flow Patterns Hope et al aortic valve disease. Five had no aortic stenosis. All 11 had fusion of the right and left aortic valve leafl ets. The four patients with left-handed nested systolic helical fl ow and left- posterior fl ow jets had fusion of the right and noncoronary leafl ets ( Fig 5 ). Their mean age was 12.3 years 6 4.7. All had some degree of aortic valve dis- ease: Three had moderate or severe aortic stenosis, one had mild aortic stenosis, and two had concomitant mild aortic insuffi ciency. Of the patients with a BAV with normal aortic dimensions ( n = 13), four had normal fl ow and nine had nested helical systolic fl ow. Five of these nine patients with nested helical fl ow were younger than 20 years (ie, 6, 14, 14, 17, and 19 years), and the remaining four were at least 30 years old (ie, 30, 39, 43, and 45 years). The patients with a BAV with mod- erate or severe aortic stenosis ( n = 5) had some aliasing of signal at the vena contracta, but this did not preclude ac- curate assessment of the overall fl ow pattern. Discussion Abnormal systolic helical fl ow is seen in the ascending thoracic aorta of patients with a BAV. Nested helical fl ow was demonstrated at peak systole in 75% of patients with a BAV but in none of the healthy volunteers or patients with a TAV. This abnormal helical fl ow was always associated with an eccentric sys- tolic fl ow jet in the proximal ascending aorta. Similar helical fl ow has been de- scribed in ascending aortic aneurysms associated with a BAV ( 18 ). We have replicated this fl ow pattern in six patients with dilated ascending aortas. However, we have also demonstrated the fl ow pat- tern in nine patients with normal aortic dimensions, suggesting that the pattern is not secondary to a dilated aorta and may be implicated in the pathogenesis of aneurysm formation. Of the nine pa- tients with normal aortic dimensions and helical fl ow, four were at least 30 years old, which provides an argument against the hypothesis that the abnormal helical Table 2 Ascending Thoracic Aorta Systolic Flow in Patients with a BAV Nested Helical Flow Characteristic Normal Flow ( n = 5) Right-handed ( n = 11) Left-handed ( n = 4) Age (y) * 21.2 6 9.4 28.6 6 11.0 12.3 6 4.7 Ascending aorta maximum diameter Normal ( n = 13) 4 7 2 Dilated ( n = 7) 1 4 2 Aortic insuffi ciency Mild ( n = 5) 1 2 2 Moderate or severe ( n = 1) … 1 … Aortic stenosis Mild ( n = 5) … 4 1 Moderate or severe ( n = 5) … 2 3 Aortic leafl et fusion Right-left ( n = 15) 4 11 … Right-noncoronary ( n = 5) 1 … 4 Note.—Unless otherwise specifi ed, data are numbers of patients. * Data are means 6 standard deviations. dimensions ( n = 21). High-velocity sys- tolic streamlines spanned the ascend- ing aorta with minimal deviance from the direction of bulk fl ow in all of these participants ( Fig 1 ). Additionally, all of these participants exhibited some de- gree of normal skewing of bulk systolic fl ow to the right side of the ascending aorta and slight right-handed twisting of relatively slow peripheral streamlines along the left side of the ascending aorta, which became more pronounced in late systole. Patients with a Dilated Ascending Aorta Ten of 11 patients with dilated ascending thoracic aorta had abnormal second- ary fl ow features. These abnormalities could be subdivided into two distinct categories that corresponded to the presence of a TAV or BAV. Patients with a TAV with dilated ascending aortas ( n = 4) had an abnormal systolic fl ow pattern in which streamlines wrapped back toward the aortic valve (a “vertical vortex”). Six of seven patients with a BAV with dilated ascending aortas had marked right-handed nested helical sys- tolic fl ow ( Fig 2 ). Patients with a BAV The nested helical fl ow demonstrated at peak systole in patients with a BAV and a dilated ascending aorta was also seen in nine of 13 patients with a BAV and normal aortic dimensions. Overall, this fl ow pattern was present in 15 of the 20 patients with a BAV but in none of the healthy volunteers or patients with a TAV. The association between nested helical systolic fl ow and pres- ence of BAV was shown to be signifi cant ( P , .001). No association was found be- tween aortic coarctation repair ( n = 28, 15 of whom had a BAV) or gothic aortic arch morphology ( n = 8, four of whom had a BAV) and abnormal systolic fl ow. Patients with a BAV exhibited normal or right-handed or left-handed nested helical systolic fl ow in the ascending thoracic aorta ( Table 2 ). Those with normal systolic streamlines ( n = 5) had systolic fl ow jets that appeared to be nearly normal ( Fig 3 )