%0 Journal Article
%T 基于常规MRI特征的脑胶质瘤IDH1基因型预测模型的建立与验证<br>Establishment and Validation of a Predictive Model for IDH1 Genotype in Gliomas Based on Conventional MRI Features
%A 李吉贤
%A 张华
%A 刘学军
%J Advances in Clinical Medicine
%P 787-796
%@ 2161-8720
%D 2024
%I Hans Publishing
%R 10.12677/acm.2024.1441090
%X 目的：基于常规MRI特征构建脑胶质瘤异柠檬酸脱氢酶1 (isocitrate dehydrogenase 1, IDH1)基因型的预测模型，并对模型进行验证。方法：回顾性分析254例脑胶质瘤患者的临床资料及术前MRI图像，采用随机分组的方法按约7:3比例分为训练集(n = 177)和验证集(n = 77)。收集其影像学特征，使用单因素分析、多因素逐步回归分析筛选特征后建立Logistic回归预测模型。采用受试者工作特征(receiver operating characteristic, ROC)曲线下面积(area under curve, AUC)检验模型的预测效能，并采用Hosmer-Lemeshow检验模型的拟合优度，同时绘制校准曲线及决策曲线。结果：单因素分析与多因素逐步回归分析显示T2-Flair错配、强化特点、皮层侵犯3个变量差异有统计学意义，采用3个MR特征构建多因素Logistic预测模型并进行验证，结果显示该模型在训练集和验证集的AUC分别为0.832和0.828，Hosmer-Lemeshow检验结果显示模型具有良好的拟合度(训练集<i>χ</i><sup>2</sup> = 4.568，p = 0.335，验证集<i>χ</i><sup>2</sup> = 2.744，p = 0.433)，校准曲线显示模型的校准度较好，决策曲线分析表明模型具有较高的净收益。结论：基于常规MRI特征构建的预测模型可有效预测脑胶质瘤IDH1基因型。<br />
Objective: To develop and validate a predictive model for the genotype of isocitrate dehydrogenase 1 (IDH1) in gliomas based on MRI features.<i> </i>Methods:<i> </i>The clinical and preoperative MRI data of 254 patients with glioma were analyzed retrospectively. All patients were randomly divided into a training set (n = 177) and a validation set (n = 77) according to a ratio of 7:3. Imaging features were collected and utilized to establish a Logistic regression predictive model, following feature selection through univariate analysis and multivariate stepwise regression analysis. The model’s predictive performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) and the Hosmer-Lemeshow test for goodness of fit, along with the generation of calibration and decision curves.<i> </i>Results:<i> </i>Univariate and multivariate stepwise regression analyses revealed statistically significant differences in three variables: T2-Flair mismatch sign, enhancement pattern, and cortical invasion. A multivariate Logistic predictive model was constructed using these three MRI features and subsequently validated. The results demonstrated that the model achieved AUCs of 0.832 and 0.828 in the training and validation sets, respectively. The Hosmer-Lemeshow test indicated a good model fit (training set <i>χ</i><sup>2</sup> = 4.568, p = 0.335; validation set <i>χ</i><sup>2</sup> = 2.744, p = 0.433). Calibration curves showed that the model was well-calibrated, and decision curve analysis indicated that the model had a high net benefit.<i> </i>Conclusion:<i> </i>The predictive model based on MRI features can effectively predict the IDH1 genotype in gliomas.
%K 磁共振成像，胶质瘤，异柠檬酸脱氢酶1，列线图<br>Magnetic Resonance Imaging
%K Glioma
%K Isocitrate Dehydrogenase-1
%K Nomogram
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=84368