Department of Radiology, The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Dongming Road, Zhengzhou, Henan, 450008, China.
Department of Radiology and Intervention, The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Dongming Road, Zhengzhou, Henan, 450008, China.
Clin Radiol. 2022 Aug;77(8):e560-e567. doi: 10.1016/j.crad.2022.04.005. Epub 2022 May 18.
To explore the predictive value of the radiomics feature-based nomogram for predicting telomerase reverse transcriptase (TERT) promoter mutation status and prognosis of lower-grade gliomas (LGGs) non-invasively.
One hundred and seventy-six LGG patients (123 in the training cohort and 53 in the validation cohort) were enrolled retrospectively. A total of 851 radiomics features were extracted from contrast-enhanced magnetic resonance imaging (MRI) images. The radiomics features were selected using the least absolute shrinkage and selection operator (LASSO) method and a rad-score was calculated. Multivariate logistic regression analysis was used to build a radiomics signature based on rad-score, participant's age, and gender, and a radiomics nomogram was used to represent this signature. The performance of the signature was evaluated by receiver operating characteristic (ROC) curve analysis, and the patient prognosis was stratified based on the TERT promoter mutation status and the radiomics signature.
Seven robust radiomics features were selected by LASSO and the radiomics signature showed good performance for predicting the TERT promoter mutation status, with an area under the curve (AUC) of 0.900 (0.832-0.946) and 0.873 (0.753-0.948) in the training and validation datasets. With a median overall survival time of 28.5 months, the radiomics signature stratified the LGG patients into two risk groups with significantly different prognosis (log-rank = 47.531, p<0.001).
The radiomics feature-based nomogram is a promising approach for predicting the TERT promoter mutation status preoperatively and evaluating the prognosis of lower-grade glioma patients non-invasively.
探索基于放射组学特征的列线图预测端粒酶逆转录酶(TERT)启动子突变状态和低级别胶质瘤(LGG)患者预后的价值。
回顾性纳入 176 例 LGG 患者(训练队列 123 例,验证队列 53 例)。从增强磁共振成像(MRI)图像中提取了 851 个放射组学特征。使用最小绝对值收缩和选择算子(LASSO)方法选择放射组学特征,并计算放射评分。使用多变量逻辑回归分析,基于放射评分、患者年龄和性别构建放射组学特征模型,并使用放射组学列线图表示该模型。通过受试者工作特征(ROC)曲线分析评估模型的性能,并根据 TERT 启动子突变状态和放射组学特征对患者进行预后分层。
LASSO 筛选出 7 个稳健的放射组学特征,该放射组学特征在预测 TERT 启动子突变状态方面表现良好,训练集和验证集的曲线下面积(AUC)分别为 0.900(0.832-0.946)和 0.873(0.753-0.948)。中位总生存期为 28.5 个月,放射组学特征将 LGG 患者分为具有显著不同预后的两个风险组(对数秩检验=47.531,p<0.001)。
基于放射组学特征的列线图是一种有前途的方法,可用于术前预测 TERT 启动子突变状态,并无创评估低级别胶质瘤患者的预后。