MRI Research Laboratory, Hospital for Special Surgery, New York, New York, USA.
Center for Imaging Research, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.
Med Phys. 2022 Oct;49(10):6538-6546. doi: 10.1002/mp.15925. Epub 2022 Aug 23.
Diffusion-weighted imaging (DWI) provides quantitative measurement of random water displacement in tissue as calculated by the apparent diffusion coefficient (ADC). While heavily utilized in stroke and oncology applications, DWI is a promising tool to map microstructural changes in musculoskeletal applications including evaluation of synovial reactions resulting from total hip arthroplasty (THA). One major challenge facing the application of DWI in THA is the significant artifacts related to the conventional echo-planar imaging (EPI) readout used. Multispectral imaging (MSI) techniques, including the multiacquisition with variable resonance image combination (MAVRIC), have been shown to effectively reduce metallic susceptibility artifacts around total joint replacements to render clinically useful images. Recently, a 2D periodically rotated overlapping parallel line with enhanced reconstruction (PROPELLER) FSE acquisition that incorporates a diffusion preparation pulse with 2D-MAVRIC has been developed to mitigate both distortion and dropout artifacts. While there have been some preliminary assessments of DWI-MAVRIC, the repeatability of DWI-MAVRIC and the effects of key parameters, such as the number of spectral bins, are unknown.
To evaluate the quantitative accuracy of DWI-MAVRIC as compared to conventional diffusion sequences.
A diffusion phantom with different reference diffusivities (ADC = 113-1123 μm /s) was used. Scans were performed on two 1.5T MRI scanners. DWI-EPI and DWI-MAVRIC were acquired in both the axial and coronal planes. Three spatial offsets (0 cm, 10 cm left, and 10 cm right off iso-center) were used to evaluate effects of off-isocenter positioning. To assess intraday and interday repeatability, DWI-EPI and DWI-MAVRIC acquisitions were repeated on one scanner at same-day and 9-month intervals. To assess inter-scanner repeatability, DWI-EPI and DWI-MAVRIC acquisitions were compared between two scanners. ADC maps were generated with and without gradient nonlinearity correction (GNC). Linear regression, correlation, and error statistics were determined between calculated and reference ADC values. Bland-Altman plots were generated to evaluate intraday, interday, and interscanner repeatability.
DWI-MAVRIC had excellent correlation to reference values but at reduced linearity (r = 1.00, slope = 0.91-0.94) as compared to DWI-EPI (r = 1.00, slope = 0.99-1.01). A greater than 5% ADC bias was observed at the lowest ADC values, predominantly in the DWI-MAVRIC scans. ADC values did not vary with DWI-MAVRIC parameters. DWI-EPI acquisitions had intraday, interday, and interscanner repeatability of 3.18 μm /s, 19.2 μm /s, and 20.2 μm /s, respectively. DWI-MAVRIC acquisitions had inferior intraday, interday, and interscanner repeatability of 13.3 μm /s, 44.7 μm /s, 110 μm /s, respectively. Lower ADC errors were found at isocenter, as compared to the left and right positions. GNC reduced the absolute error by 0.31% ± 0.89%, 3.6% ± 1.4%, 0.65% ± 2.4% for the center, left, and right positions, respectively.
DWI-MAVRIC provides good linearity with respect to reference values and good intra- and interday repeatability.
弥散加权成像（DWI）通过表观扩散系数（ADC）提供组织中随机水分子位移的定量测量。虽然在中风和肿瘤学应用中得到了广泛应用，但 DWI 是一种很有前途的工具，可以在肌肉骨骼应用中绘制微观结构变化，包括评估全髋关节置换术（THA）引起的滑膜反应。DWI 在 THA 中的应用面临的一个主要挑战是与传统的回波平面成像（EPI）读出相关的显著伪影。多光谱成像（MSI）技术，包括多采集可变共振图像组合（MAVRIC），已被证明可以有效地减少全关节置换周围的金属敏感性伪影，从而提供临床有用的图像。最近，开发了一种 2D 周期性旋转重叠并行线增强重建（PROPELLER）FSE 采集，该采集结合了扩散准备脉冲和 2D-MAVRIC，以减轻失真和失物伪影。虽然已经对 DWI-MAVRIC 进行了一些初步评估，但 DWI-MAVRIC 的重复性以及关键参数（如光谱 bin 的数量）的影响尚不清楚。
评估 DWI-MAVRIC 与传统扩散序列的定量准确性。
使用具有不同参考扩散系数（ADC = 113-1123 μm/s）的扩散体模。在两台 1.5T MRI 扫描仪上进行扫描。在轴向和冠状平面上采集 DWI-EPI 和 DWI-MAVRIC。使用三个空间偏移量（0cm、10cm 左侧和 10cm 右侧离等中心）来评估离等中心定位的影响。为了评估日内和日间重复性，在同一台扫描仪上以相同的日期和 9 个月的间隔重复进行 DWI-EPI 和 DWI-MAVRIC 采集。为了评估扫描仪间的重复性，比较了两台扫描仪之间的 DWI-EPI 和 DWI-MAVRIC 采集。生成了具有和不具有梯度非线性校正（GNC）的 ADC 图。确定了计算值与参考 ADC 值之间的线性回归、相关性和误差统计数据。生成 Bland-Altman 图以评估日内、日间和扫描仪间的重复性。
与 DWI-EPI 相比，DWI-MAVRIC 与参考值具有极好的相关性，但线性度降低（r = 1.00，斜率 = 0.91-0.94）（r = 1.00，斜率 = 0.99-1.01）。在最低 ADC 值处观察到大于 5%的 ADC 偏差，主要在 DWI-MAVRIC 扫描中。ADC 值不受 DWI-MAVRIC 参数的影响。DWI-EPI 采集的日内、日间和扫描仪间重复性分别为 3.18 μm/s、19.2 μm/s 和 20.2 μm/s。DWI-MAVRIC 采集的日内、日间和扫描仪间重复性分别为 13.3 μm/s、44.7 μm/s 和 110 μm/s，重复性较差。与左侧和右侧位置相比，在等中心处发现的 ADC 误差较小。GNC 将绝对误差分别降低了 0.31%±0.89%、3.6%±1.4%、0.65%±2.4%，用于中心、左侧和右侧位置。
DWI-MAVRIC 与参考值具有良好的线性关系，具有良好的日内和日间重复性。
