Author + information
- David Messika-Zeitoun, MD⁎ (, )
- Jean-Michel Serfaty, MD, PhD,
- Jean-Pierre Laissy, MD, PhD,
- Mohammed Berhili, RT,
- Eric Brochet, MD,
- Bernard Iung, MD and
- Alec Vahanian, MD
- ↵⁎Cardiovascular Department, AP-HP, Bichat Hospital, 46 rue Henri Huchard, 75018 Paris, France
To the Editor:Management of mitral stenosis (MS) relies on accurate assessment of the mitral valve orifice area (MVA). Planimetry measured by Doppler echocardiography is considered the reference method, but must be precisely performed at the tips of the leaflet on a well-oriented plane and therefore requires experienced operators.
Preliminary reports suggest that electrocardiogram (ECG)-gated multislice computed tomography (MSCT) may provide an accurate assessment of mitral valve morphology (1). The MSCT allows a three-dimensional acquisition of the entire heart throughout the cardiac cycle and multiple plane reconstructions. Therefore, it could theoretically provide, similar to two-dimensional echocardiography, a parasternal short-axis view of the mitral valve orifice at the tips of the leaflet. Moreover, acquisition of MSCT images depends on the protocol used, which can be preregistered in the MSCT system, and a step-by-step manual of operations for plane orientations can be proposed that may reduce the operators’ dependence on MSCT measurements.
The goals of the present study were to assess the feasibility, accuracy, and reproducibility of MVA measurements by MSCT using two-dimensional echocardiography as a reference.
We prospectively evaluated 29 patients (50 ± 13 years, 86% female) with a wide range of MS severity by two-dimensional echocardiography and MSCT (performed within 48 h). Exclusion criteria were: 1) nonfeasible planimetry; 2) atrial fibrillation (AF); 3) pacemaker; 4) abnormal renal function (creatinine >2 mg/dl); and 5) pregnancy. Informed consent was obtained from all patients.
Planimetry (MVAECHO) was performed in early diastole by one experienced operator, in the parasternal short-axis view, adjusting the probe for optimal mitral valve orifice.
The MSCT acquisition was performed within a single breath hold using a 16-detector-row Philips MX-8000 scanner (Philips, Eindhoven, the Netherlands) with commercially available cardiac reconstruction software. Contrast enhancement was achieved with 100 to 120 ml of nonionic contrast material. The computerized ECG tracing was recorded during acquisition for retrospective reconstruction. Measurements were performed in early diastole (75% RR) according to the following manual of operations. In four- and two-chamber views, the anteroposterior long axis was first positioned across the center of the mitral valve orifice to the apex of the left ventricle (Figs. 1Aand 1B). The other plane was then adjusted perpendicularly to the mitral orifice at the tip of the leaflets to obtain the parasternal short-axis view (Fig. 1C), and the MVA was measured (MVACT) (Fig. 1D).
The MSCT measurements were performed blinded to any clinical and echocardiographic data by three operators: two experienced operators for MSCT validation and intraobserver and interobserver variability and one nonexperienced operator (radiology technologist with no previous experience in valvular disease) for reproducibility of MSCT measurements. After brief training, the radiology technologist was asked to perform the measurements alone.
Comparisons of MVA obtained by MSCT and echocardiography were analyzed by using paired ttests, Pearson correlation, and Bland-Altman analysis. Intraobserver and interobserver variability was calculated as mean difference ± SD. Statistical significance was defined as p < 0.05.
Image quality was good in most patients and acceptable in 4 patients. In all patients MVA measurements could be performed. The time required to perform MSCT measurements was 3 ± 1 min.
The MVACT(1.51 ± 0.47 cm2[range 0.85 to 2.4]) did not differ from and correlated well with MVAECHO(1.44 ± 0.53 cm2[range 0.7 to 2.3], p = 0.15; r = 0.88, p < 0.0001) (Fig. 2A).The mean absolute difference between both methods was small (0.20 ± 0.17 cm2). Quality control plots showed no trend for underestimation or overestimation (Fig. 2B). Fifteen patients (52%) had severe MS (MVAECHO<1.5 cm2), and none were misclassified using MSCT. Intraobserver and interobserver variability, established in 15 patients randomly selected from the study population, were low (respectively, 0.14 ± 0.12 cm2and 0.17 ± 0.13 cm2).
The MSCT measurements, performed by the nonexperienced operator in 26 patients, did not differ from and correlated well with both the MSCT measurements performed by the first experienced operator (1.45 ± 0.55 cm2vs. 1.47 ± 0.49 cm2, p = 0.70; r = 0.89, p < 0.0001) and MVAECHO(1.38 ± 0.52, p = 0.25; r = 0.82, p < 0.0001). The mean differences between MSCT and echocardiographic measurements for nonexperienced and experienced operators were not different (respectively, 0.27 ± 0.18 cm2vs. 0.21 ± 0.17 cm2, p = 0.17). Also, the mean difference of the MSCT measurements between experienced and nonexperienced operators did not differ from the interobserver variability established between experienced operators (0.21 ± 0.13 vs. 0.18 ± 0.13 cm2, p = 0.76).
Echocardiographic evaluation of the MVA is the cornerstone of the assessment of MS severity. Planimetry, which provides an anatomic measurement of the mitral valve orifice, is considered the reference method but requires experienced operators. Measurements must be performed at the tips of the leaflets with a correct plan orientation. This requirement is the major limitation of the planimetry method.
The MSCT has rarely been used for the evaluation of valvular disease despite attractive technical characteristics; MSCT offers the possibility of a three-dimensional acquisition of the entire heart throughout the cardiac cycle and multiple plane reconstructions. In the present study, we show that MSCT can provide an accurate MVA evaluation. The MSCT measurements were highly reproducible and were performed in only a few minutes. Intraobserver and interobserver variability were low and comparable with that of other methods. Therefore, MSCT can be added to the repertoire of techniques available for MS severity assessment. It is worth noting that MSCT compares favorably with two other three-dimensional techniques, namely three-dimensional echocardiography and magnetic resonance imaging (2,3).
Even with experienced operators, planimetry of the mitral valve orifice by echocardiography may not be feasible in 5% to 10% of patients, mainly because of poor echocardiographic windows. For such patients, an easy nonechocardiographic method for MVA evaluation is desirable, and MSCT seems to be an attractive alternative.
In the present study, CT measurements performed by a radiology technologist with no previous experience in valvular disease were not different from echocardiographic measurements. Methodology of CT measurements can be standardized with a step-by-step manual of operation, and we show that with minimal training, an accurate MVA evaluation can be achieved. Because MSCT is widely used throughout the world, it may help teams less accustomed to evaluating patients with MS. This is of particular importance because MS is a relatively rare disease with which few physicians have a high level of expertise.
The main limitation of the present study was the exclusion of patients not in normal sinus rhythm. Second, only 8 patients with a calcified mitral valve were enrolled. The MSCT should be further evaluated in this subgroup, but our results suggest that calcification should not affect CT accuracy (yellow circles in Fig. 2). In the present study, intraobserver and interobserver variability of echocardiographic measurements have not been evaluated but are low and have been previously reported (4). Finally, because of intrinsic MSCT limitations (risk of irradiation and of iodine contrast), we do not recommend MSCT as a first-intention method for MS severity assessment.
The MSCT can provide accurate and reproducible planimetry of the mitral valve orifice in patients with MS. In addition, accurate MVA measurements can be achieved with minimal training, showing that planimetry of the mitral valve orifice by MSCT is not operator dependant. Thus, MSCT can be added to the repertoire of techniques available for MS severity assessment and can be considered as an alternative for MVA measurements in clinical practice, especially in patients with poor echocardiographic windows or for teams less accustomed to evaluating patients with MS.
Please note: Drs. Messika-Zeitoun and Serfaty have equally contributed to this article.
- American College of Cardiology Foundation
- Zamorano J.,
- Cordeiro P.,
- Sugeng L.,
- et al.
- Djavidani B.,
- Debl K.,
- Lenhart M.,
- et al.