We investigated the reproducibility of 3D-TSE and conventional 2D-DIR-TSE imaging in coronary vessel wall imaging in KD. To our knowledge, this is the first study to compare the 3D-TSE and 2D-DIR-TSE images in the coronary arteries. Qualitative and quantitative evaluations suggested reproducibility between the 3D-TSE and 2D-DIR-TSE images.
Despite slower flow velocity and larger influence of turbulence in the aneurysmal region than those in the normal region, the 3D-TSE images were equal to or better than the 2D-DIR-TSE images in the visual grade of the lumen boundary. High-resolution, 3D-TSE imaging with variable flip angle trains (e.g., VISTA, SPACE, and CUBE) is advantageous for vessel wall imaging at various sites, considering its high scan efficiency, enabled by the use of long echo trains and its intrinsic black blood effect [15, 20–23]. 3D-TSE imaging reportedly exerts a strong suppressive effect on blood flow signals in the coronary arteries. We used 3D-TSE imaging to select axial slice orientation. We observed suppression in coronary artery vessel wall imaging when the flow direction was close to parallel to the slice encoding direction. However, there was no suppression when the direction was close to perpendicular [11]. Aneurysms are often located in the proximal portion of the coronary artery [24]. By axially collecting the proximal portion of the coronary artery, the blood flow direction and the slice encoding direction become close to parallel, such that the blood flow signal was effectively suppressed.
In normal regions, the 3D-TSE images were equal to or less than the 2D-DIR-TSE images in the visual grade of the outer wall boundary. 3D-TSE imaging suggestively has a longer shot duration and scan time than 2D-DIR-TSE imaging, which increases the effects of heart action and respiration movement. In follow-up examination of KD, it is necessary not only to detect high-intensity thromboses but also to depict detailed anatomical structures, such as aneurysm regression. Considering the small anatomical size of coronary arteries and the movement during heart action and respiration, coronary vessel wall imaging remains more challenging than imaging of other vascular beds [25]. In a previous study, autopsy studies in subjects who did not die of coronary artery disease have revealed that the coronary vessel wall is typically 0.4–0.8 mm thick [26]. A previous report on the impact of spatial resolution on the accuracy of vessel WA measurement in simulations and phantom studies mentioned that a resolution ≤4 pixels across the wall leads to an overestimation of over 20% [27]. Moreover, coronary wall imaging reportedly requires a spatial resolution of 0.5 mm to 1 mm [28, 29]. In the present study, the acquired spatial resolution of 1.17×1.25×1.40 mm was further reconstructed to 0.55×0.55×0.70 mm. Nonetheless, it is necessary to further improve the spatial resolution with technological advancements in future.
3D-TSE and 2D-DIR-TSE images had high ICCs and no bias, suggesting the images were reproducible. A previous study investigated the reproducibility of the internal diameter between 3D-TSE images and MRA [11]. Larger the diameter of the aneurysm, narrower was the internal diameter of the 3D-TSE image than that of the MRA image. This may be attributed to laminar flow proximal to the vessel wall and increased turbulent flow, caused by changes in the shape of the vessel lumen. In contrast, there was no bias in the internal diameter between 3D-TSE images and MRA in normal regions, thereby generating the expected range. The signal intensity of conventional 2D-DIR-TSE imaging relies on the inflow of fresh blood into the imaging slice between selective re-inversion and imaging. Therefore, aneurysmal regions are susceptible to flow direction and velocity. 2D-DIR-TSE images have a strong directivity in the acquisition section and flow direction, similar to 3D-TSE images.
The inter- and intra-rater reproducibility in the area measurement suggested that 3D-TSE images may be suitable for follow-up examination of the coronary vessel wall. 3D-TSE imaging also allows for an easy visualization of accurate cross-sections, with less partial volume effects following post-processing.
While this study generated significant findings, there were several limitations. First, we did not compare MR coronary vessel wall imaging with IVUS or OCT [11, 30], a standard approach for coronary artery wall assessments. However, IVUS is an invasive technique with non-negligible risks and is difficult to perform routinely. Second, the voxel sizes of 3D-TSE and 2D-DIR-TSE imaging did not match. However, there occurs a trade-off between high-resolution imaging and increased scan time, which can increase patient discomfort.
In conclusion, 3D-TSE imaging was reproducible with conventional 2D-DIR-TSE imaging for coronary vessel wall assessment of KD. The aneurysmal regions necessitate careful WA assessment, owing to the residual lumen signals of 3D-TSE and 2D-DIR-TSE images. 3D-TSE imaging may be suitable for follow-up examinations of KD because of its wide FOV, accurate cross-sectional images, and the reproducibility of area measurements.