基于多序列3.0T MRI的纹理分析预测恶性软组织肉瘤术后复发的价值

Value of Texture Analysis Based on Multi-Sequence 3.0T MRI in Predicting Postoperative Recurrence of Malignant Soft Tissue Sarcoma

目的探讨基于多序列3.0 T MRI的纹理分析对软组织肉瘤(STS)术后复发的预测价值。方法回顾性分析具有完整临床资料和随访结果的手术治疗STS患者MRI资料(常规MRI平扫、IVIM-DWI及MRI增强序列)。根据术后随访是否复发或转移,将数据资料分为复发组(10例)和非复发组(14例)。盲法条件下搜集肿块影像基本参数及临床资料、测量肿块实性成分的IVIM-DWI值、及在T_(2)WI、IVIM-DWI(b=1200 s/mm^(2))及增强T_(1)WI(CE-T_(1)WI)图像提取的纹理特征;根据数据分布类型,采用独立样本t检验或Mann-Whitney U检验、χ^(2)检验或Fisher’s精确检验进行组间对比分析,P<0.05为差异有统计学意义;然后通过ROC曲线分析,评估有统计学意义的参数之诊断效能,并采用Logistic回归分析影响STS患者复发的因素。结果复发组和非复发组临床特征及与IVIM-DWI的ADC值、D值、D^(*)值、f值差异无明显统计学意义(P>0.05);基于IVIM-DWI(b=1200 s/mm^(2))纹理分析显示DWIwavelet-LHL_glcm_InverseVariance、DWIwavelet-HHH_glszm_SizeZoneNonUniformityNormalized这两个参数的组间差异具有统计学意义(P<0.05),曲线下面积分别为0.829及0.793;Logistic回归分析进一步显示这两个因素是预测STS患者术后是否复发的独立因素,建立模型曲线下面积为0.907。结论基于IVIM-DWI序列提取的纹理特征在预测STS患者远期复发具有一定的意义,可为临床医师的治疗决策提供参考价值。

Objective To investigate the predictive value of texture analysis based on multi-sequence 3.0T MRI for postoperative recurrence of soft tissue sarcoma(STS).Methods This retrospective study investigated the MRI data of STS patients who had surgical treatment with complete information of clinical and follow-up visit.The multiple sequences included routine plain scan,intravoxel incoherent motion diffusion weighted imaging(IVIM-DWI)and contrast enhanced MRI.The patients divided into two groups,10 cases of recurrent group and 14 cases of non-recurrent group,according to the follow-up visit focused on postoperative recurrence or metastasis.The researchers collected the basic radiological and clinical information,measured the parameters of IVIM-DWI for solid component of tumors,and extracted textures features on MRI images(T_2WI,IVIM-DWI with b=1200 s/mm^(2) and contrast enhanced T_1WI)in blinded condition.The Independent Sample t test,Mann-Whitney U test,Pearson Chi-Square test or Fisher's Exact test were chose to compare the two groups according to the tests of normality.The results with P>0.05 were significant statistically.Then the diagnostic value of significant parameters were demonstrated by receiver operating characteristic curves,and the important predictive indicators in influencing recurrence of STS were analyzed by Logistic regression analysis.Results The clinical information and IVIM-DWI parameters of two groups had no significant differences(P<0.05)between the two groups.The comparisons of two texture features,called DWIwavelet-LHL_glcm_InverseVariance and DWIwavelet-HHH_glszm_SizeZoneNonUniformityNormalized,based on IVIM-DWI with b=1200 s/mm^(2) reached statistically significant levels(P<0.05).The areas under the curve(AUCs)of them were 0.829 and 0.793,respectively.Finally,the Logistic regression analysis showed that the two texture features were the independent factors to predict the recurrences of STS patients postoperatively,and the regression model were built with the AUC was 0.907.Conclusion The texture analysis based on IVIM-DWI has impressed potency in predicting the recurrence of STS patients,and may produced fundamental basis of individual accurate treatment strategy for clinicians.