Planimetry is the most direct, and thus should be the most stable approach to determine a valvular orifice area. It is not derived from other hemodynamic and geometric calculations and hence is widely independent from hemodynamic conditions. For mitral valves, planimetry showed the best correlation with anatomical mitral valve area as assessed on explanted valves [
9] and is therefore considered as the reference measurement of mitral valve area [
10]. Based on the successful application of CMR planimetry in aortic bioprostheses, we hypothesized that CMR planimetry of mitral bioprostheses would also provide accurate orifice areas, as long as the image quality allows for a clear-cut delineation of the prosthetic cusp borders.
Feasibility of CMR to image mitral bioprostheses
We have demonstrated the feasibility of CMR to image mitral bioprostheses, even in the presence of arrhythmias. Indeed, only 6 subjects presented in sinus rhythm, while 1 had frequent ventricular extrasystoles and 11 were experiencing atrial fibrillation. This incidence of arrhythmia is typical for patients previously suffering from mitral regurgitation or stenosis and having atrial dilatation [
11]. As SSFP cine movies collect data over several cardiac cycles, the image quality negatively correlates with the extent of arrhythmia and with increasing heart rate. Even in the present sample, differences in image quality were observed between patients in sinus rhythm and those with arrhythmia. Nevertheless, none of the examinations hat to be excluded due to non-diagnostic image quality. But, optimal heart rate control is recommended preceding the assessment of a mitral bioprosthesis by CMR.
A strong stent artifact appeared in one subject with a Perimount® prosthesis, whereas this phenomenon did not occur in the remaining subjects, who received other types of devices. This underlines that some prostheses are less suitable for CMR assessment and that the type of the valve makes an important difference concerning image quality. The differences probably depend on the shape and quantity of metal within the stent, and on its composition, which is a cobalt-chromium spring alloy in the Perimount®. Further trials including a greater variety of prosthetic types are warranted.
Artifacts caused by turbulent flow did not constitute a significant challenge in mitral prostheses, which can be attributed to the low transmitral blood flow velocity. Therefore, SSFP is the sequence of choice despite its general susceptibility for turbulent flow [
12,
13]. In rare cases with flow artifacts, less flow sensitive FGRE [
13] may be applied. However, mitral prosthetic orifice areas obtained by FGRE showed less agreement with TTE than those determined by SSFP, which is in accordance with the use of FGRE to evaluate aortic stenosis [
14].
Accuracy of CMR to quantify the orifice area of mitral bioprostheses
Direct comparison of the orifice areas obtained by CMR planimetry with a gold standard would be desirable to confirm our hypothesis that CMR of mitral bioprostheses provides accurate prosthetic orifice areas. However, a true gold standard does not exist. Even when evaluating native mitral stenosis it is recommended that calculated mitral valve area by any one method should not be used as the single measure of severity of stenosis [
15]. Orifice areas provided by the manufacturers are not a suitable gold standard, as they mostly comprise the geometric orifice area in-vitro, which is different from the orifice area effectively used in-vivo [
16]. Planimetry using echocardiography is complicated by difficulties in identifying the correct cross-section [
17,
18]. Three-dimensional imaging may improve echocardiography [
19,
20], but has not yet entered routine clinical practice. The accuracy and reproducibility of the continuity equation are hampered by the number of contributing measurements, which is even more problematic in arrhythmias [
18]. Furthermore, this algorithm is inappropriate in those with aortic or mitral regurgitation, as in our study subjects. The validity of pressure half time derived orifice areas, as traditionally applied in native mitral stenosis [
8], is unclear in prosthetic valves, because of the dependence on left ventricular and left atrial compliance, on left atrial pressure and heart rate [
1,
21‐
23]. Nevertheless, in the absence of a better standard, and bearing these limitations in mind, we selected this parameter as the most appropriate for our study.
CMR planimetry of mitral bioprosthetic orifice area was significantly correlated with data obtained by TTE. The agreement between both methods was similar to the successful application of CMR to assess native mitral valve stenosis by Djavidani et al. (
r = 0.98; mean difference 0.03 ± 0.09 cm
2) [
24] and even superior compared to another series by this group (r = 0.81; mean difference 0.13 ± 0.24 cm
2) [
4]. Compared to the application of computed tomography to assess native mitral valves in comparison to TTE-planimetry (r = 0.88; mean difference 0.20 ± 0.17 cm
2) [
25], the present results in mitral bioprostheses showed even closer agreement. In comparison to our results of CMR and TTE to assess aortic bioprostheses (r = 0.82; mean difference -0.02 ± 0.24 cm
2) [
3], the limits of agreement of CMR and TTE in mitral prosthetic assessment were smaller.
The present results demonstrated similar agreement between CMR- and TTE- derived orifice areas both in the presence or absence of arrhythmias. There was a tendency towards underestimation of orifice areas using CMR compared to TTE in examinations rated with "moderate" and "impaired" image quality, which may be explained by less accurate delineation of the border between blood and prosthetic cusps. However, the sample size is too small to draw firm conclusions regarding the influence of arrhythmias and image quality on prosthetic orifice assessment; thus, larger series are required.
CMR and TTE showed agreement that was independent of prosthetic orifice area. However, the study sample only contained 3 patients with an orifice area ≤1.7 cm2 and just 1 patient with an orifice area ≤1.5 cm2. Prosthetic dysfunction may be associated with increased turbulent flow causing significant flow artifacts, which may influence image interpretation. Therefore, larger clinical trials are required to extend the use of CMR to the assessment of dysfunctional mitral prostheses.
The inter- and intra-observer variabilities of CMR planimetry were low and within the range of accepted echocardiographic results that report 5% to 8% variability for experienced observers examining native aortic valves [
26]. Observer dependency was also comparable to data obtained using CMR in aortic bioprostheses with an intra- and interobserver variablity of 6.7% and 11.5%, respectively [
3].
Clinical impact of CMR to quantify the orifice area of mitral bioprostheses
TTE will remain the first choice for the assessment of mitral bioprostheses due to its ready availability, even bedside in the intensive care, cost effectiveness, non-invasiveness and proven accuracy. Bioprosthetic mitral valves constitute a very small number of patients undergoing mitral valve surgery, and of those an even smaller percentage will have unfavorable acoustic windows making Doppler interrogation unreliable. Nevertheless the present results demonstrated that for selected subjects, e.g. in case of insufficient acoustic windows, discordant echocardiographic results or within clinical research regarding patient-prosthesis mismatch [
27], CMR planimetry poses a reliable non-invasive method to quantify the prosthetic orifice area.