IDEAL-IQ技术结合纹理分析对肾脏少脂肪血管平滑肌脂肪瘤的鉴别诊断价值

IDEAL-IQ sequence with texture analysis for differentiating renal angiomyolipomas with minimal fat from non angiomyolipomas

目的:探讨磁共振IDEAL-IQ技术结合纹理分析对肾脏少脂肪血管平滑肌脂肪瘤(AML)的诊断价值。方法:将42例经手术或临床确诊的肾脏肿瘤患者(共44个病灶)纳入研究,其中少脂肪AML患者8例,非AML患者34例。每例患者均采用2个序列进行MRI扫描来评估病灶及其内的脂肪成分:(1)双回波T_1WI(同/反相位);(2)IDEAL-IQ序列。对同/反相位图像采用阅片分析。对IDEAL-IQ序列图像,首先在脂肪分量(FF)图像上手动勾画ROI并融合形成全病灶ROI,然后使用软件进行纹理分析,取得全病灶的多项纹理分析参数,包括最小值、最大值、中位值、平均值、标准差、方差、第5、第15和第30百分位值、均匀性、偏度、峰度、范围和熵,比较在2个序列图像上2组病变特征的差异。5和第30百分位值、均匀性、偏度、峰度、范围和熵,比较在2个序列图像上2组病变特征的差异。结果:同反相位T_1WI显示的含脂病灶的组内占比在少脂AML组与非AML组之间的差异无统计学意义(P>0.05)。IDEAL-IQ序列获得的病灶的14项纹理参数中,最小值、最大值、中位值、平均值、标准差、方差、第5、第15和第30百分位值及均匀性这10项参数值在少脂AML组与非AML组间的差异有统计学意义(P<0.05),两组间偏度、峰度、均匀性和熵值的差异无统计学意义(P>0.05)。结论:磁共振IDEAL-IQ序列结合纹理分析能准确测量肾脏肿瘤病灶内的少量脂肪,对少脂肪肾脏血管平滑肌脂肪瘤的鉴别诊断具有重要的临床价值。

Objective:To assess the ability of MR IDEAL-IQ technique combined with texture analysis in differentiating angiomyolipomas with minimal fat(MF-AML)from non-angiomyolipomas(non-AML).Methods:42 patients with 44 renal tumors(36 non-AMLs and 8 MR-AMLs)were scanned with dual-echo T 1WI(in-phase and opposed-phase)and iterative decomposition of water and fat with echo asymmetry and least squares estimation-image quantification(IDEAL-IQ)sequence.The fat content on dual-echo T 1W images was analyzed by two radiologists subjectively.On fat fraction images obtained by IDEAL-IQ sequence,volume ROI of whole lesion was drawn and then texture ana-lysis was used to obtain several texture parameters including minimum,maximum,median,average,standard deviation(SD),variance,and 5,15 and 30 percentile(named as Per5,Per15 and Per30),uniformity,skewness,kurrosis,range and entropy.Results:There was statistically significant difference in ten texture parameters between MF-AMLs and non-AMLs,these parameters included minimum,maximum,median,average,SD,variance,Per5,Per15,Per30 and uniformity;while no significant difference was found in skewness,kurrosis,range and entropy.There was no significant difference of the ratio of the lesions with fat showed on dual-echo T 1WI between MF-AMLs and non-AMLs.Conclusion:MR IDEAL-IQ technique combined with texture analysis can effectively identify minimal fat in renal tumors,and is of great value in the differential diagnosis of renal angiomyolipomas with minimal fat.