Summary
Background
Recently, 1.5-Tesla cardiac magnetic resonance imaging (CMR) was reported to provide a reliable alternative to transthoracic echocardiography (TTE) for the quantification of aortic stenosis (AS) severity. Few data are available using higher magnetic field strength MRI systems in this context.
Aims
To evaluate the feasibility and reproducibility of the assessment of aortic valve area (AVA) using 3-Tesla CMR in routine clinical practice, and to assess concordance between TTE and CMR for the estimation of AS severity.
Methods
Ninety-one consecutive patients (60 men; mean age 74 ± 10 years) with known AS documented by TTE were included prospectively in the study.
Results
All patients underwent comprehensive TTE and CMR examination, including AVA estimation using the TTE continuity equation (0.81 ± 0.18 cm 2 ), direct CMR planimetry (CMRp) (0.90 ± 0.22 cm 2 ) and CMR using Hakki’s formula (CMRhk), a simplified Gorlin formula (0.70 ± 0.19 cm 2 ). Although significant agreement with TTE was found for CMRp ( r = 0.72) and CMRhk ( r = 0.66), CMRp slightly overestimated (bias = 0.11 ± 0.18 cm 2 ) and CMRhk slightly underestimated (bias = –0.11 ± 0.17 cm 2 ) AVA compared with TTE. Inter- and intraobserver reproducibilities of CMR measurements were excellent ( r = 0.72 and r = 0.74 for CMRp and r = 0.88 and r = 0.92 for peak aortic velocity, respectively).
Conclusion
3-Tesla CMR is a feasible, radiation-free, reproducible imaging modality for the estimation of severity of AS in routine practice, knowing that CMRp tends to overestimate AVA and CMRhk to underestimate AVA compared with TTE.
Résumé
Contexte
L’imagerie par résonance magnétique cardiaque à 1,5-Tesla a récemment été proposée comme alternative à l’échocardiographie transthoracique (ETT) dans l’évaluation de la sévérité de la sténose aortique (SA). Peu de données sont actuellement disponibles utilisant des IRM avec champs magnétiques plus intenses.
Objectifs
Évaluer la faisabilité et la reproductibilité de la mesure de la surface aortique en utilisant l’IRM à 3-Tesla en pratique clinique, et la concordance entre la surface aortique estimée grâce à l’IRM 3-Tesla et en ETT.
Méthodes
Quatre-vingt-11 patients consécutifs (60 hommes, d’âge moyen 74 ± 10 ans) porteurs d’une SA documentée en ETT ont été inclus prospectivement dans cette étude.
Résultats
Tous les patients ont bénéficié à la fois d’une ETT permettant l’évaluation de la surface aortique par l’équation de continuité (0,81 ± 0,18 cm 2 ), d’une IRM 3-Tesla permettant une planimétrie directe (IRMp) (0,90 ± 0,22 cm 2 ) et une estimation de la surface aortique par la formule de Hakki (IRMhk) (0,70 ± 0,19 cm 2 ). Malgré des corrélations significatives avec l’ETT pour IRMp ( r = 0,72) comme pour IRMhk ( r = 0,66), IRMp surestimait significativement (biais = 0,11 ± 0,18 cm 2 ) et CMRhk tendait à sous-estimer (biais = –0,11 ± 0,17 cm 2 ) la surface aortique par rapport à l’ETT. La reproductibilité inter-et intra-observateur étaient excellentes pour les mesures en IRM ( r = 0,72 et r = 0,74 for IRMp, r = 0,88 et r = 0,92 pour le pic de vélocité aortique, respectivement).
Conclusions
L’IRM à 3-Tesla est une modalité non irradiante, fiable et reproductible pour l’évaluation de la sévérité d’une sténose aortique en pratique clinique, en sachant que l’IRMp a tendance à surestimer la surface aortique et IRMhk à la sous-estimer par rapport à l’ETT.
Background
Aortic stenosis (AS) is the most common form of acquired valvular heart disease, and its prevalence is expected to increase in Western populations. Transthoracic echocardiography (TTE) classically confirms the diagnosis and assesses the severity of AS. Proper visualization of aortic valve anatomy, including calcification and opening, together with haemodynamic measurements are mandatory to assess the severity of AS . Nevertheless, the accuracy of a TTE examination may be limited in patients with poor acoustic windows (mostly patients who are obese or have chronic pulmonary disease). Moreover, recent studies have reported discrepancies between mean transvalvular gradient and aortic valve area (AVA) in some patients , underlining the need for a multimodal approach in difficult cases.
Other imaging modalities, such as multidetector computed tomography and cardiac magnetic resonance imaging (CMR) might be helpful to complement or confirm the information obtained by TTE . CMR provides unique morphological and functional information in this setting. Recently, CMR was reported to be a reliable alternative to TTE for the quantification of AS severity, using either direct planimetry of the aortic valve orifice or velocity-encoded CMR techniques. These previous studies were all performed using 1.5-Tesla CMR systems. Magnetic resonance imaging systems with higher magnetic field strengths have become widely available, and 3-Tesla has become the favoured field strength for brain magnetic resonance imaging in clinical practice. 3-Tesla magnetic resonance imaging, with the use of parallel imaging, allows an increased signal-to-noise ratio, with faster acquisition, leading to potentially superior image quality . Recent data showed the advantages of 3-Tesla CMR over 1.5-Tesla CMR in clinical cardiac applications, such as perfusion imaging or delayed enhancement . However, little is known about the feasibility, in a routine clinical setting, of 3-Tesla CMR recordings in AS.
Thus, the purpose of this prospective clinical study was to evaluate the feasibility and reproducibility of the assessment of AVA using 3-Tesla CMR in routine clinical practice, and to assess the concordance between TTE and CMR for the evaluation of AS severity.
Methods
Between January and December 2015, 102 consecutive patients with known AS documented by TTE were eligible for inclusion into the study. Patients with left ventricular dysfunction (ejection fraction ≤ 50%), aortic regurgitation or atrial fibrillation were not excluded. Exclusion criteria were standard contraindications to magnetic resonance imaging and poor TTE imaging quality. Institutional review board approval was obtained before conducting the study. The study was conducted in accordance with institutional policies, national legislation and the revised Helsinki declaration.
Echocardiography
All patients underwent a comprehensive Doppler echocardiographic study, using commercially available ultrasound systems. Peak aortic velocity was recorded using continuous-wave Doppler in several acoustic windows (apical five-chamber view, right parasternal, suprasternal and subcostal). The highest aortic velocity was used to calculate the aortic time-velocity integral and the mean Doppler gradient. Pressure gradients were calculated using the simplified Bernoulli equation. AVA was calculated using the continuity equation , and indexed for body surface area. Stroke volume was calculated by multiplying the area of the left ventricular outflow tract by the outflow tract time-velocity integral. For patients in sinus rhythm, three cardiac cycles were averaged for all measures; for patients in atrial fibrillation, five cardiac cycles were averaged. Patients with an ejection fraction ≥ 50% ( n = 66) were classified into four groups depending on left ventricular flow state (normal flow versus low flow) and pressure gradient level (low gradient versus high gradient). Low flow was defined as an indexed left ventricular stroke volume < 35 mL/m 2 ; low gradient was defined as a mean transaortic pressure gradient < 40 mmHg . This classification results in the following four flow-gradient patterns: normal flow/low gradient, normal flow/high gradient, low flow/high gradient and low flow/low gradient. Left ventricular dimensions were assessed from parasternal long-axis views by two-dimensional guided M-mode using the leading edge methodology at end-diastole and end-systole. Ejection fraction was calculated using Simpson’s biplane method. Left ventricular mass was estimated by the formula based on linear measurements, and indexed for body surface area. The maximal velocity of tricuspid regurgitation was estimated using continuous-wave Doppler.
CMR technique
All patients underwent a comprehensive TTE and CMR examination within 24 hours, in comparable haemodynamic states. Patients were imaged with a 3-Tesla Siemens Skyra scanner (Siemens, Erlangen, Germany), 18-channel body flex coils and 45 mT/m gradients. Assessment of cardiac function was performed with a cine steady-state free precession pulse sequence (TrueFISP sequence) with retrospective gating, in end-expiration breath-hold. The following projections were acquired: two-chamber, four-chamber, parallel contiguous short-axis view (to cover the entire left ventricle from the mitral plane to the apex); two orthogonal left ventricular outflow tract, through-plane aortic valve view (acquired at the tip of the cusps to obtain the smallest valvular area, and at the time of maximum opening of the valve). We classified the bicuspid aortic valve type according to Sievers classification . Typical parameters were: repetition time/echo time 2.8–1.23 ms; field of view 341 × 430; matrix 216 × 272; flip angle 47–38°, depending on the specific absorption rate limit; slice thickness 6 mm (4 mm for the aortic valve); parallel acquisition with iPAT GRAPPA (factor × 3); number of phases 25; and temporal resolution 48 ms. In the presence of artefacts on the steady-state free precession sequences, a fast low angle shot gradient echo sequence in a cross-sectional aortic valve plane was used. Typical parameters were: repetition time/echo time 2.8–2.6 ms; field of view 276 × 340; matrix 125 × 192; flip angle 12°; slice thickness 5 mm; parallel acquisition with iPAT GRAPPA (factor × 2); number of interpolated phases 30; and temporal resolution 48 ms. For the flow quantification, through-plane phase-contrast images were acquired with a retrospective gated sequence, in end-expiration breath-hold, performed at different levels distal to the aortic valve, perpendicular to the jet, to find the peak aortic velocity. Typical parameters were: repetition time/echo time 4.9–2.56 ms; field of view 273 × 400; matrix 96 × 176; flip angle 20°; slice thickness 6 mm; parallel acquisition with iPAT GRAPPA (factor × 3); number of interpolated phases 30; and temporal resolution 39 ms. The velocity encoding was set to a value close to the aliasing threshold. The CMR data were processed offline, as described below, by an independent observer blinded to the results of TTE. Data were analysed using dedicated software (Circle cvi 42® , version 5.1; Circle Cardiovascular Imaging, Calgary, AB, Canada). On the cine images, we measured left ventricular ejection fraction, end-diastolic and end-systolic volumes, mass (on the short-axis views) and AVA with anatomical planimetry (on the through-plane aortic valve view). AVA was measured by direct planimetry on the cine images (CMRp), and by using Hakki’s formula (CMRhk), which is a simplification of Gorlin formula (AVA = cardiac output [L/min]/√4 V max 2 [mmHg]), using the phase-contrast images to assess V max (peak aortic velocity).
Statistical analysis
Data for the study population and TTE and CMR measurements are presented as number (percentage) or mean ± standard deviation after testing for normal distribution (Kolmogorov–Smirnov test). CMR and TTE measurements were compared by Student’s t -test or the Wilcoxon rank test, as appropriate. Correlation and agreement between CMR and TTE measurements were assessed by Pearson’s correlation and the Bland–Altman comparison, respectively. Inter- and intraobserver variabilities were examined for CMR measurements of AVA (CMRp and peak aortic velocity). Measurements were performed in a group of 12 randomly selected patients by one observer, then repeated offline on two separate days by two independent observers, who were blinded to the other’s measurements and to the study time point. Data are presented as correlation coefficient ( r ) and intraclass correlation coefficient. All P values are the result of two-tailed tests. A value of P < 0.05 was considered significant.
Results
Study population
Eleven patients were not included because of standard contraindications to magnetic resonance imaging (severe claustrophobia, n = 4; cardiac pacemaker or implanted cardioverter-defibrillator, n = 7). Thus, the final study group consisted of 91 patients (60 men; mean age 74 ± 10 years). When considering excluded patients, feasibility was 89% for CMR and 100% for TTE. Baseline demographic and clinical characteristics of the 91 patients are displayed in Table 1 . Using CMR, the bicuspid aortic valve was described in 38 patients (42%), predominantly type 1LR ( n = 28; 74%), followed by type 1RN ( n = 7; 18%) and type 0 ( n = 3; 8%).
| Variable | Overall population ( n = 91) | Normal flow AS ( n = 62) | Low flow AS ( n = 17) | P |
|---|---|---|---|---|
| Age (years) | 74 ± 10 | 72 ± 9 | 79 ± 8 | 0.01 |
| Men | 60 (66) | 39 (63) | 12 (70) | 0.09 |
| Body surface area (kg/m 2 ) | 1.82 ± 0.22 | 1.81 ± 0.23 | 1.84 ± 0.18 | 0.82 |
| Coronary artery disease | 31 (34) | 22 (35) | 5 (29) | 0.85 |
| Diabetes mellitus | 13 (14) | 10 (16) | 2 (12) | 0.90 |
| Hypertension | 31 (34) | 23 (37) | 6 (35) | 0.72 |
| CMR left ventricular ejection fraction (%) | 62 ± 10 | 64 ± 9 | 62 ± 8 | 0.54 |
| CMR AVA planimetry (cm 2 ) | 0.90 ± 0.22 | 0.86 ± 0.17 | 0.97 ± 0.21 | 0.16 |
| CMR AVA (Hakki’s formula) (cm 2 ) | 0.70 ± 0.19 | 0.70 ± 0.17 | 0.63 ± 0.16 | 0.06 |
| Indexed left ventricular end-diastolic volume (mL/m 2 ) | 71 ± 21 | 70 ± 17 | 61 ± 14 | 0.007 |
| Indexed left ventricular end-systolic volume (mL/m 2 ) | 29 ± 17 | 26 ± 11 | 24 ± 10 | 0.52 |
| Indexed left ventricular mass (mg/m 2 ) | 96 ± 25 | 95 ± 24 | 87 ± 20 | 0.24 |
| Stroke volume index (mL/m 2 ) | 42 ± 9 | 44 ± 9 | 34 ± 6 | 0.0001 |
| Peak aortic velocity (m/s) | 4.0 ± 0.6 | 4.1 ± 0.6 | 3.7± 0.6 | 0.07 |
| Aortic annulus diameter (mm) | 22 ± 2 | 22 ± 2 | 22 ± 2 | 0.83 |
| Aortic sinus diameter (mm) | 33 ± 5 | 32 ± 5 | 33 ± 4 | 0.64 |
| Ascending aorta diameter (mm) | 37 ± 8 | 36 ± 6 | 36 ± 6 | 0.90 |
| Indexed left atrial volume (mL/m 2 ) | 51 ± 24 | 51 ± 23 | 49 ± 27 | 0.75 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
