优化b值对肝脏体内不相干运动扩散加权成像参数值测量准确性和可重复性的影响

Influence of optimizing b-value on the accuracy and repeatability of the intravoxel incoherent motion diffusion-weighted imaging parameter measuremens of the liver

目的研究优化b值对肝脏体内不相干运动(intravoxel incoherent motion,IVIM)弥散加权成像(diffusion imaging,DWI)参数值测量准确性的影响,以期提高肝脏IVIM DWI数据采集的效率。方法 60位志愿者在3.0T磁共振(MRI)扫描仪上进行肝脏IVIM DWI检查,采用16个b值(0~800s/mm2)为对照组,其中20例用于评估IVIM参数值的可重复性。利用双指数模型测定肝脏的IVIM DWI的参数,包括灌注分数(perfusion fraction,PF),真性弥散系数D,假性弥散系数D*。实验组采用多个b值组合,b值数量从4~15个,对每一个亚组的定量参数值均进行计算,并计算整体参数评估误差(global parameter estimation error,GPER)。评估在两次扫描之间的扫描间误差(interscan estimation error,IEE),以评估肝脏IVIM参数值的可重复性。在最佳b值分布的亚组中,筛选GPER低于重复扫描误差且b值数量最少的亚组,确定最佳组合亚组。结果随着b值数目的降低,IVIM所有参数的GPER均增高,反映测量参数准确性的下降。对于任意组合,所得到的IVIM-DWI参数总体误差均在重复测量误差之内。利用优化后的4个b值组合(0、15、150、800s/mm2),D、PF、D*的误差分别为7.5%、24.9%、67.5%,均在重复测量误差的范围之内,D、PF、D*的重复测量误差分别为13.0%、35.8%、210.8%。结论与16个b值采集方案相比,优化后4个b值在不明显降低IVIM测量参数值准确性和重复测量误差的前提下,可以大幅度的降低采集时间。

Objective To study the influence of optimizing b-value on the accuracy and repeatability of the intravoxel incoher- ent motion （IVIM） diffusion-weighted imaging （DWI） parameter measurements of the liver, so as to improve the efficiency of IVIM DWI data acquisition of the liver. Methods Sixty volunteers （M/F： 35/25, mean age： 55 ＋ 13 y） underwent IVIM DWI at 3.0 T magnetic resonance imaging （MRI） scanner using 16 b-values （0 to 800 s/ram2）, of which 20 cases were used for the reproducibility assessment of IVIM parameters. A double exponential model was used to quantify IVIM DWI parameters of the liv- er, including perfusion fraction （PF）, true diffusion coefficient （D） and pseudo diffusion coefficient （D ＊ ）. The experimental group used a combination of multiple b values, with the number of b values ranging from 4 to 15, and the quantitative parameters of each subgroup and the global parameter estimation error （GPER） were calculated. Interscan estimation errors （IEE） were also evaluated between two scans to assess reproducibility of the IVIM parameters measurement in the liver. The optimal b-values com- bination subgroup was selected, in which the number of b-values were minimal while keeping parameters estimation error lower than interscan reproducibility error. Results As the number of b-values decreased, the global parameter estimation error of all IVIM parameters increased, reflecting decreased precision of IVIM metrics. For any b-values combinations, the overall errors of IVIM-DWI parameters were within the scope of the repeated measurement error. By using an optimal combination of 4 b-values （0, 15, 150 and 800 s/mm2 ）, the errors were 7.5, 24.9 and 67.5 % for D, PF and D＊ respectively, which were within the scope of the repeated measurement error, respectively, repeated measurement error were 13.0%, 35.8%, 210.8% for D, PF, D＊. Conclusion Compared to the 16 b-values acquisition scheme, the set of 4 optimized b-values could be used to reduce the acquisitiontime significantly, without significant decrease of the precision and reproducibility of IVIM parameters measurement.