利用机器学习鉴别胶质母细胞瘤标准化治疗后真假性进展的研究

Application of machine-learning to differentiate true progression from pseudoprogression in GBM patients after standard treatment

目的:探讨基于T1加权增强成像(T1CE)的机器学习模型在鉴别胶质母细胞瘤(GBM)患者标准化治疗后真假性进展的诊断效能。方法:回顾性分析了我院2014年5月至2017年2月间经手术病理证实的77例胶质母细胞瘤患者,所有患者均行标准化治疗。利用ITK-SNAP软件划取全部强化部分为感兴趣区(VOI),应用A. K.软件提取9675个特征。此外,采集的临床信息包括:性别、年龄、KPS评分、切除范围、神经功能缺损和平均放疗剂量。利用随机森林分类器(RF)建立分类模型,以鉴别GBM标准化治疗后的真假性进展。通过计算受试者工作特征(ROC)的曲线下面积(AUC)、敏感度、特异度和准确性评估模型的诊断效能。结果:真性进展51例,假性进展26例。真假性进展患者的临床基线特征均无显著统计学差异。基于影像组学特征的机器学习模型诊断效能相对较高,AUC值、准确性、敏感度和特异度分别为0. 79(95%CI:0. 63-0. 98)、72. 78%、78.36%和61. 33%。结论:基于T1CE增强图像特征建立的机器学习模型对GBM标准化治疗后真假性进展的鉴别效能相对较高,有助于临床医生尽早制定适当的治疗方案。

Objective: To explore the efficacy of the machine learning using radiomics features from conventional T1-weighted contrast enhanced imaging( T1CE) in differentiating pseudoprogression from true progression of glioblastoma multiforme( GBM) patients after standard treatment. Method: A total of 77 patients including 51 patients with true progression and 26 patients with pseudoprogression. The volume of interest( VOI) covering the whole tumor enhancement were manually drawn on the T1CE slice by slice and a total of 9675 features were extracted from the VOI using Analysis-Kinetics software. The clinical features were also recorded. Random forest( RF) algorithm and 5-fold cross validation were adopted to construct the model to differentiate pseudoprogression from true progression. The commonly used metric was used for assessing the efficiency. Results: The RF strategy with radiomics features produced the stable diagnostic efficiency,with an AUC,accuracy,sensitivity,and specificity of 0. 79( 95% confidence interval [CI]:0. 63-0. 98),72. 78%,78. 36% and 61. 33%,respectively. Adding the clinical features did not significantly change the conclusion. Conclusion: Machine learning using T1CE radiomics features may be a promising strategy to differentiate true progression from pseudoprogression in GBM patients after the standard treatment. It may help doctors plan appropriate treatment therapy as early as possible.