|Home | About | Journals | Submit | Contact Us | Français|
To evaluate if left ventricular outflow tract /aortic valve (LVOT/AO) diameter ratio measured by cardiac magnetic resonance (CMR) imaging is an accurate marker for LVOT obstruction in patients with hypertrophic cardiomyopathy (HCM) compared to Doppler echocardiography.
92 patients with hypertrophic cardiomyopathy were divided into 3 groups based on their resting echocardiographic LVOT pressure gradient (PG): <30mmHg at rest (non-obstructive, n=31), <30 mmHg at rest, >30mmHg after provocation (latent, n=29) and >30mmHg at rest (obstructive, n=32).The end-systolic dimension of the LVOT on 3-chamber steady state free precession (SSFP) CMR was divided by the end diastolic aortic valve diameter to calculate the LVOT/AO diameter ratio.
There were significant differences in the LVOT/AO diameter ratio among the 3 subgroups (non-obstructive 0.60±0.13, latent 0.41±0.16, obstructive 0.24±0.09, p<0.001). There was a strong linear inverse correlation between the LVOT/AO diameter ratio and the log of the LVOT pressure gradient (r=−0.84, p<0.001). For detection of a resting gradient >30mmHg, the LVOT/AO diameter ratio the area under the ROC curve was 0.91 (95% CI 0.85-0.97). For detection of a resting and/or provoked gradient >30mmHg, the LVOT/AO diameter ratio area under the ROC curve was 0.90 (95% CI 0.84-0.96).
The LVOT/AO diameter ratio is an accurate, reproducible, noninvasive and easy to use CMR marker to assess LVOT pressure gradients in patients with HCM.
Hypertrophic cardiomyopathy (HCM) is the most common inherited cardiomyopathy, with frequent features of asymmetric left ventricular hypertrophy, systolic anterior motion of the mitral valve and left ventricular outflow tract obstruction. The severity of LVOT obstruction determines therapy and is an, independent predictor of death and heart failure (1). Gradients of 30 mmHg or higher are thought to reflect true mechanical impedance to outflow and may have prognostic significance (1). In HCM patients, gradients are most commonly evaluated by Doppler echocardiography (2). More recently cardiac magnetic resonance (CMR) has emerged as a useful diagnostic tool in the evaluation of HCM patients and has been shown to provide information incremental to that provided by echocardiography (3-18). However, despite exquisite morphologic and functional cardiac assessment, CMR may not reliably assess outflow hemodynamics. Phase contrast magnetic resonance imaging has been used to assess aortic valve stenosis (19). However routine clinical phase contrast CMR sequences may not accurately quantify flow for severely stenotic valves due to limitations in echo time (TE) and temporal resolution (20, 21). In HCM patients, the flow in the LVOT may be extremely turbulent, resulting in significant CMR artifacts. CMR planimetry-derived LVOT area > 3.7 cm2 appears to rule out significant LVOT obstruction in HCM patients (22). However, this approach requires several special orthogonal views of the LVOT, which are not routinely included in the CMR exam and prolong scan time.
We hypothesized that the effective anatomic outflow tract, expressed as the ratio of LVOT to aortic valve diameter determines the extent of hemodynamic obstruction and may therefore predict severity of LVOT obstruction. We tested this hypothesis in 105 consecutive HCM patients using echocardiographic-Doppler gradients as the reference standard.
Institutional review board approval was obtained for this study. We enrolled 105 consecutive patients with confirmed HCM, who underwent CMR and Doppler echocardiography between January 2007 to May 2009. Diagnosis of HCM was based on genetic confirmation of a pathogenic mutation or on the conventional criteria of LV hypertrophy not originating from other causes (≥15 mm or ≥13 mm in documented familial disease).
Exclusion criteria included a left ventricular wall thickness <13mm, Doppler echocardiography not performed within 24 hours of the CMR exam, inadequate image quality or incomplete CMR or echocardiogram. Thirteen patients were excluded from the study because they did not meet the study criteria. Ninety two patients met the inclusion criteria for the study (Table 1).
ECG gated steady state free precession segmented cine images were acquired at 1.5 tesla (Avanto, Siemens, Germany). Acquisition parameters were: TR 4.7ms, TE 1.5ms, matrix 256×256, temporal resolution 35-40 ms, FOV 36×36cm, 30 reconstructed phases of the cardiac cycle. The LVOT was demonstrated on a “3-chamber view” (Figure 1). The most severe narrowing of the LVOT during systole was identified (distance between the anterior leaflet of the mitral valve and the interventricular septum) and was divided by the end diastolic diameter at the level of the aortic valve yielding the LVOT/AO diameter ratio (Figures 1 and and2).2). We measured the aortic valve diameter at end diastole because the aortic valve was best seen during this phase in the cardiac cycle. Two CMR readers (8 years and 2 years of CMR experience) blinded to the Doppler echocardiography results independently measured the LVOT/AO diameter ratio.
Echocardiography studies were performed with the patients in partial left lateral decubitus position with the use of a 2.5-MHz transducer. LVOT maximum velocity was measured at rest in the 3-chamber view using continuous-wave Doppler. Careful transducer angulation was performed in multiple projections to minimize angle between the jet and the incident beam. Gain and filter settings were adjusted to obtain the signal with the highest frequency, maximal peak velocity and the optimal signal to noise ratio. Special attention was made to distinguish the LVOT signal from that of mitral regurgitation by direct visualization of color Doppler on two-dimensional image and by recognizing that mitral regurgitation jet is characterized by an earlier onset and higher peak velocity than the outflow tract signal. Peak gradient across the LVOT and aortic valve were measured by Doppler echocardiography during resting period (baseline) and after exercise stress (peak provoked gradient) with electrocardiogram and blood pressure monitoring. Exercise stress echocardiography was performed using a treadmill protocol selected according to patient characteristics (Bruce or modified Bruce) and LVOT and aortic valve gradient was measured at peak stress. The LVOT PG was calculated using the Bernoulli equation: PG = 4×(V 21 - V 22),where V1 and V2 are proximal and LVOT velocities, respectively (23).
Patients were divided into 3 groups based on their echocardiographic LVOT PG: <30mmHg at rest (non-obstructive, n=31), <30 mmHg at rest and >30mmHg after provocation (latent, n=29) and >30mmHg at rest (obstructive, n=32).
Statistical tests were performed using SPSS (Cupertino, CA, USA). Data are presented as mean ± one standard deviation. Continuous variables were compared using the Student’s t-test with Bonferoni correction when appropriate. Correlations between continuous variables were tested using the Pearson correlation coefficient. Receiver operating characteristic (ROC) curves were used to define the accuracy of the LVOT/AO diameter ratio.
Adequate quality CMR and echocardiography data were obtained in 92 patients. Baseline characteristics of the patient population are given in Table 1.
The LVOT/AO diameter ratio was significantly different among the 3 LVOT pressure gradient subgroups (non-obstructive 0.60±0.13, latent 0.41±0.16, obstructive 0.24±0.09, Bonferoni adjusted p value < 0.017, Figure 3a). There was an inverse relationship between LVOT/AO diameter ratio and log of the resting LVOT pressure gradient (r=−0.84, p<0.001, Figure 3b).
For detection of a resting gradient >30mmHg the area under the ROC curve was 0.91 (95% CI 0.85-0.97, figure 4a); a cutoff LVOT/AO diameter ratio of 0.33 yielded a good balanced sensitivity (91%, 95% CI: 75 - 98) and specificity (80%, 95% CI: 67- 89). The cutoff LVOT/AO diameter ratio of 0.45 yielded a sensitivity of 97% (95% CI: 84 - 100) and specificity of 60% (95% CI: 47- 72) for detection of a resting gradient >30mmHg.
For detection of a resting and/or provoked gradient >30mmHg using the LVOT/AO diameter ratio the area under the ROC curve was 0.90 (95% CI 0.84-0.96, figure 4b). The cutoff LVOT/AO diameter ratio of 0.45 yielded a sensitivity of 82% (95% CI: 0.73- 0.91) and specificity of 86 % (95% CI: 0.78 – 0.93) for detection of a resting and/or provoked gradient >30mmHg.
There was good interobserver agreement (concordance correlation coefficient 0.89) for the LVOT/AO diameter ratio measurement.
Our data demonstrate that a simple, easy to measure, morphologic CMR-derived parameter, the LVOT/AO diameter ratio, is a reproducible, accurate marker of LVOT obstruction in HCM patients when compared to resting Doppler echocardiography. There was a strong inverse relationship between LVOT/AO diameter ratio and log of the resting LVOT pressure gradient (r=−0.84). For detection of a resting gradient >30mmHg, the LVOT/AO diameter ratio the area under the ROC curve was 0.91 (95% CI 0.85-0.97). For detection of a resting and/or provoked gradient >30mmHg, the LVOT/AO diameter ratio area under the ROC curve was 0.90 (95% CI 0.84-0.96). The cutoff LVOT/AO diameter ratio of 0.45 yielded a sensitivity of 97% (95% CI: 84 - 100) to detect >30 mmHg resting gradients in the LVOT. To our knowledge this is the first description of such a method to predict outflow tract hemodynamics in HCM by CMR.
Outflow tract obstruction is a common and clinically important feature of HCM (24-26). Delineation of the severity of obstruction has several clinical implications including prediction of outcomes such as heart failure and death (1). It also determines the composition and extent of medical therapy and whether the patient will be referred for interventional options such as myectomy or septal ablation (27).
CMR has been used extensively in the evaluation of HCM for many reasons. CMR provides very high resolution cine-images of cardiac anatomy. It allows reliable mapping of the patterns of ventricular hypertrophy. More recently, CMR-derived delayed enhancement has been shown to predict risk for cardiac arrhythmias (4, 28). An important limitation of CMR has been its relative inability to evaluate outflow tract hemodynamics. Various CMR techniques have been attempted to overcome this limitation, such as phase contrast imaging. Standard two-dimensional phase contrast sequences may not reliably capture the peak flow in the LVOT, because of the complex three-dimensional structure of the LVOT, significant through-plane motion during the cardiac cycle and often high turbulence within the LVOT during systole (19, 20). Novel time resolved three-dimensional phase contrast CMR techniques may overcome some of these problems, although their current temporal resolution for a given scan time (usually 60-80ms) needs to be improved to capture the relatively short-lived peak flow jet in the LVOT (29, 30). Furthermore, the post-processing time for three-dimensional phase contrast CMR data sets is currently quite labor intensive and currently not used in clinical routine.
Morphologic indices such as the LVOT cross sectional area have been similarly used to gain knowledge about the severity of LVOT obstruction (22). However, this approach requires several special orthogonal views of the LVOT,that are not routinely included in the CMR exam and prolong the exam time. With this method, the smallest LVOT area obtained in these orthogonal slices can be measured during systole, which sometimes can be challenging. Similar to our study, there was an inverse nonlinear relationship between the LVOT area derived by CMR planimetry and pressure gradient in Doppler echocardiography in a study of 37 HCM patients (22).
We decided to use the LVOT/AO diameter ratio measurement to estimate LVOT obstruction as this ratio would provide an anatomic analogy to the continuity equation. (23, 31). In our study, the LVOT/AO diameter ratio enabled identification of the presence or absence of a resting LVOT gradient >30mmHg with good accuracy. Also the area under the ROC curve for detection of a resting and/or provoked LVOT gradient >30mmHg using the LVOT/AO diameter ratio was quite high (0.90).
Due to the known day-to-day variation in PG measurements and variable preloading conditions it would have been ideal to perform both Doppler and CMR as consecutive tests (32-34). Yet, the correlation between PG and LVOT/AO ratio obtained within 24 hours remained strongly significant. In our study, the LVOT pressure gradient was only assessed noninvasively. Since invasive hemodynamic evaluation using the Gorlin formula is not routinely performed, we used noninvasive echocardiographic assessment as our reference standard. Echocardiography shows a very strong correlation to simultaneous invasive measurements (34).
In conclusion, LVOT/AO diameter ratio is an accurate, reproducible and noninvasive CMR marker to assess LVOT pressure gradients in patients with HCM. LVOT/AO ratio less than 0.45 was a suitable cutoff value to estimate clinically relevant LVOT pressure gradients in patients with HCM.
Funding: NIH Grant # HL 098046