基于生境分析对肝细胞癌Ki-67表达的预测

Prediction of Ki-67 Expression in Hepatocellular Carcinoma Based on Habitat Analysis

摘要: 目的：构建一种基于钆塞酸二钠(Gd-EOB-DTPA)增强MRI的生境分析模型，用于术前预测肝细胞癌术前的Ki-67表达。方法：对来自两个医疗机构的433例经病理确诊的HCC患者进行回顾性分析。根据术后免疫组织化学检测的Ki-67表达水平，将患者分为Ki-67高表达组(n = 320)和低表达组(n = 113)，并进一步按时间顺序划分为训练集(n = 349)和测试集(n = 84)。采用最小绝对收缩和选择算子(LASSO)回归分析筛选最优预测因子，建立基于磁共振成像的生境成像模型。使用受试者工作特征曲线下面积(AUC)、准确度、敏感度等评估性能。结果：一共提取了3290个特征并将肿瘤内部划分成2个区域，经特征降维后保留61个特征。在训练集和测试集中，生境分析模型的AUC值分别为0.836和0.814。结论：基于Gd-EOB-DTPA增强MRI的生境分析模型可作为术前无创预测Ki-67表达状态的有效方法。

Abstract: Objectives: Construction of a habitat analysis model based on Gd-EOB-DTPA-enhanced MRI for preoperative prediction of Ki-67 expression in hepatocellular carcinoma. Methods: A retrospective analysis was conducted on 433 patients with pathologically confirmed hepatocellular carcinoma (HCC) from two medical institutions. Based on the Ki-67 expression level detected by postoperative immunohistochemistry, patients were divided into a high Ki-67 expression group (n = 320) and a low expression group (n = 113), and were further chronologically divided into a training set (n = 349) and a test set (n = 84). The least absolute shrinkage and selection operator (LASSO) regression analysis was used to select optimal predictors for establishing a habitat imaging model based on magnetic resonance imaging. The performance of the model was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, and sensitivity. Results: A total of 3290 features were extracted, and the tumor was divided into two subregions. After feature dimensionality reduction, 61 features were retained. The habitat analysis model achieved AUC values of 0.836 and 0.814 in the training set and the test set, respectively. Conclusion: The habitat analysis model based on Gd-EOB-DTPA-enhanced MRI can serve as an effective method for noninvasive preoperative prediction of Ki-67 expression status.