Shi Zhaoqi, Cai Wenli, Feng Xu, Cai Jingwei, Liang Yuelong, Xu Junjie, Zhen Junhao, Liang Xiao
Department of General Surgery, Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang, China.
Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts.
Acad Radiol. 2022 Feb;29(2):213-218. doi: 10.1016/j.acra.2021.04.019. Epub 2021 Jun 25.
Rationale and Objectives To evaluate the effectiveness of radiomics analysis based on Gd-EOB-DTPA enhanced hepatic MRI for functional liver reserve assessment in HCC patients. Materials and Methods Radiomics features were extracted from Gd-EOB-DTPA enhanced MRI images in 60 HCC patients. Boruta algorithm was performed to select features associated with indocyanine green retention rate at 15 min (ICG R15). Prediction and classification model were built by performing Random Forest regression analysis. Pearson correlation analysis and AUC of ROC were used to assess performance of the two models. Results A total of 165 radiomics features were extracted. Six radiomics features were selected to build the prediction model. A Predicted value of ICG R15 for each patient was calculated by the prediction model. Pearson correlation analysis revealed that predicted values were significantly associated with actual values of ICG R15 (R value = 0.90, p < 0.001). Nine radiomics features were selected to build the classification model. AUC of ROC revealed favorable performance of the classification model for identifying patients with ICG R15 <10% (AUC: 0.906, 95%CI: 0.900-0.913), <15% (AUC: 0.954, 95%CI: 0.950-0.958), and <20% (AUC: 0.996, 95%CI: 0.995-0.996). Conclusion Radiomics analysis of Gd-EOB-DTPA enhanced hepatic MRI can be used for assessment of functional liver reserve in HCC patients.
评估基于钆塞酸二钠（Gd-EOB-DTPA）增强肝脏MRI的影像组学分析在肝癌（HCC）患者功能性肝储备评估中的有效性。材料与方法：从60例HCC患者的Gd-EOB-DTPA增强MRI图像中提取影像组学特征。采用Boruta算法选择与15分钟吲哚菁绿滞留率（ICG R15）相关的特征。通过随机森林回归分析建立预测和分类模型。采用Pearson相关性分析和ROC曲线下面积（AUC）评估两种模型的性能。结果：共提取165个影像组学特征。选择6个影像组学特征建立预测模型。通过该预测模型计算每位患者的ICG R15预测值。Pearson相关性分析显示，预测值与ICG R15实际值显著相关（R值 = 0.90，p < 0.001）。选择9个影像组学特征建立分类模型。ROC曲线下面积显示，该分类模型在识别ICG R15<10%（AUC：0.906，95%CI：0.900 - 0.913）、<15%（AUC：0.954，95%CI：0.950 - 0.958）和<20%（AUC：0.996，95%CI：0.995 - 0.996）的患者方面具有良好性能。结论：Gd-EOB-DTPA增强肝脏MRI的影像组学分析可用于评估HCC患者的功能性肝储备。
