Assessment of the right ventricle with cardiovascular magnetic resonance at 7 Tesla

Function, size and morphology of the right cardiac ventricle (RV) are known to be strong influencing factors on morbidity and mortality in various cardiac diseases, for instance congenital heart disease, pulmonary hypertension, myocardial infarction or arrythmogenic right ventricular cardiomyopathy (ARVC) [1‐4]. However, their evaluation by using non-invasive imaging techniques is often a challenge, mainly attributable to the asymmetric and highly variable shape of the RV, the predominantly longitudinal systolic shortening, the small myocardial wall thickness and the location behind the sternum. Cardiovascular magnetic resonance (CMR), mainly using steady-state free-precession (SSFP) cine imaging at a field strength of 1.5Tesla (T), has evolved as the gold standard for the assessment of RV dimensions and function due to its ability to image in any plane, its excellent blood-tissue contrast, its capability to depict even small wall motion abnormalities and its proven reproducibility [4‐7]. Nevertheless, expectations to characterize the myocardial tissue of the RV comparable to the LV including the differentiation of fibrosis, fat or edema, remained widely unaccomplished yet. As an example, the current task force criteria to assess ARVC include CMR to assess RV size and function, but emphasize that the evidence of fibro-fatty replacement within the RV wall is only obtained by endomyocardial biopsy [4]. Even the accuracy of CMR to quantify RV mass, which is known to be an important predictor of cardiovascular events, is uncertain due to the challenge of resolving the thin (2-5mm) RV free wall properly using protocols (typical voxel size of cine imaging 1.8x1.8x6mm³) common in today’s clinical [8‐10]. To extend the information that is extractable from CMR, technological improvements that increase the spatial and temporal resolution as well as signal-to-noise ratio (SNR) are therefore desired. As field strength positively correlates with SNR, CMR at ultrahigh fields (7T) offers the potential to depict even microscopic structures and to facilitate targeted tissue characterization [11, 12]. However, imaging at 7T comes with technical challenges, like increased B₀ heterogeneities, non-uniform B₁ distribution and increased radio frequency (RF) power deposition. Recently, the technical feasibility of cardiac cine imaging at 7T using fast gradient echo (FGRE) techniques has been demonstrated, and data demonstrating the ability for accurate LV chamber quantification at 7T were reported [13‐16]. Moreover, dedicated transmit and receive coils as well as cardiac trigger techniques have been developed to meet the demands of CMR at 7T [17‐21].

Aim of the present study was to extend the application of cine imaging at 7T to the assessment of RV size and function and to compare the results with the gold standard at 1.5T. This work is regarded as the first step toward a comprehensive assessment of RV function, size and morphology using CMR at 7T.

The present results demonstrated that FGRE cine imaging of the RV at 7T using a dedicated 16-channel transceiver RF coil and acoustic cardiac gating provided good quality cine images of the RV to assess its size and function. The findings confirm preliminary experiences with the LV at 7T [14] and underline that cine imaging at 7T is feasible.

We know from studies of the LV with FGRE cine imaging at 1.5T and 7T that the step from 1.5T towards 7T improved the image quality of FGRE cine images, in particular in long axis acquisitions parallel to the blood flow [14]. The blood/myocardium contrast is known to be reduced with low blood flow, attributable to the inflow-dependent blood pool signal intensity in FGRE sequences [22]. The ultrahigh field strength has the potential to minimize this source of error due to different field strength dependent T₁ prolongation of blood and myocardium. Translated to the RV, this observation gains importance due to sometimes low blood flow between the RV trabeculae. In the present study of the RV, images in transverse orientation acquired with SSFP at 1.5T received the highest mean image score compared to the acquisitions at 7T. Notwithstanding, it is notable that our statistical analysis revealed that both field strengths did not differ significantly in overall image quality in systole in general, as well as in diastole if the enhanced spatial resolution (voxel size (1.3x1.3x4)mm³) is regarded. This observation underlines the feasibility to obtain high-quality cine images of standard planes at 7T and forms the basement for accurate RV chamber quantification, which is predominantly done in transverse slices [6]. In contrast, the sagittal orientated slices exhibited significant differences in image quality between SSFP at 1.5T and FGRE at 7T, with SSFP at 1.5T being more often “excellent”. The observation that the image quality at 7T varied dependent on slice orientation can be attributed to the complex B₁ field pattern with various constructive and destructive interferences throughout the FOV. Depending on the volunteer and slice orientation, a specific field pattern is revealed. From a clinical perspective, sagittal views are currently mainly used to assess RV wall motion and morphology complementary to transverse slices. This comprehensive assessment is important for example in the diagnosis of ARVC, where regional akinesia, dyskinesia or asynchrony are relevant indicators [4]. Therefore, further technical improvements are necessary to enable patient and slice dependent transmit field shaping to provide stable image quality throughout a CMR examination at 7T.

It is a recognized limitation of this study that no SNR and blood/myocardium contrast-to-noise ratio (CNR) values are reported. A previous pilot study on cardiac chamber quantification at 1.5 T and 7.0 T showed that 2D cine FGRE imaging at 7.0 T provided SNR and blood/myocardium CNR that are superior to FGRE acquisitions at 1.5T using the same slice thickness [14]. Careful SNR and CNR validation demonstrated that SNR and CNR derived from 2D cine FGRE at 7T was found to be competitive with that derived from SSFP acquisitions at 1.5 T using standard clinical routine imaging parameter [14].

Regarding image contrast as assessed by visual evaluation, the present results underline that the combination of FGRE with ultrahigh field reaches a grade of contrast that is comparable to SSFP at 1.5T, similar to the results reported for the LV [14]. For the RV, this was true both for transverse and sagittal slice orientation and independent from the voxel size. Opposed to that, the acquisition of homogeneous images is still a challenge in ultrahigh field CMR. Both in sagittal and transverse direction, the results at 7T were inferior compared to 1.5T and a high proportion of 7T images were evident with poor homogeneity for both voxel sizes. Further improvements in subject specific and slice specific B₀- and B₁-shimming are therefore essential to fully exploit the potential of 7T [23, 24].

Regarding the quantification of cardiac size and function, it has been shown for 1.5T and 3T that RV volumes are overestimated and mass underestimated when using SSFP instead of FGRE [22, 10] - even though the absolute difference is relatively small with a mean RVEDV difference of about 4ml [22]. This difference is explained by the stronger contrast between myocardium and blood on the one side, and between myocardium and epicardial fat on the other side obtained with SSFP. This leads to an outward shift of the endocardial contour and an inward shift of the epicardial contour compared to FGRE. Furthermore, areas with slow flow, as within the apex, are often mistaken for myocardium when using FGRE. It is notable from the present results that the difference in RV quantification between SSFP and FGRE no longer existed when moving from 1.5T to 7T. This tendency matches the experiences we made for the LV at 7T [14]. The bias between the results obtained for FGRE imaging at 7T with voxel size (1.2x1.2x6)mm³and SSFP at 1.5T (RVEDV 1.2±5.9ml, RVESV 1.7±2.5ml) was within the range of the inter-observer variability reported for SSFP examinations of the RV [22]. Hence, close agreement can be postulated. This finding underlines the accurate endocardial border delineation achievable by FGRE at 7T. Another interesting aspect of the present results was that FGRE at 7T by tendency provided larger volumes than SSFP 1.5T when we decreased the voxel size from (1.2x1.2x6)mm³ to (1.3x1.3x4)mm³. We explain this trend mainly by the observation that the higher spatial resolution led to a more detailed differentiation of the RV free wall that caused an outward shift of the manually contoured endocardial borders (Figure 5). This finding on the one hand has to be kept in mind when trying to translate normal values from 1.5T to 7T. On the other hand, it underscores the potential of CMR at 7T to have a more detailed look at the RV free wall. As it can be anticipated from Figure 3, some inter-subject variability of trabecular structure and compact layer thickness may exist. For instance, the measurement of the compact layer of the free RV wall as well of the remaining trabecular layer may provide new insights into the microstructure of the RV myocardium. Furthermore, the accurate quantification of RV mass based on properly spatially resolved cine images would strengthen its role as a cardiovascular risk marker. After this first evidence of the feasibility of RV cine imaging at 7T, such interesting aspects need further investigation in future studies with sample sizes large enough to provide adequate power.

In conclusion, the present study demonstrated that cine imaging of the RV is feasible at 7T using FGRE techniques. The achieved image quality was comparable to SSFP at 1.5T and allowed accurate myocardial border delineation for RV chamber quantification, which provided RV volumes, mass and function with close agreement to SSFP at 1.5T. This work is regarded as fundament and stimulator for further efforts to explore whether enhanced spatial resolution together with ultrahigh field strength can provide more detailed insights into the structure of the RV.