The value of&nbsp; <sup>18</sup> F-FDG PET/CT radiomic features for predicting <em>EGFR</em> gene mutation status in patients with lung adenocarcinoma

JIANG Xuemei; SHEN Jiaqi; PENG Xiaojuan; LI Peiqi; TAN Xiaofei; DANG Jun; YE Zhenyan; YOU Jinhui; CHENG Zhuzhong

doi:10.19732/j.cnki.2096-6210.2023.04.010

您当前的位置：

首页 >

文章列表页 >

The value of ¹⁸ F-FDG PET/CT radiomic features for predicting EGFR gene mutation status in patients with lung adenocarcinoma

更新时间：2025-12-15

- The value of ¹⁸ F-FDG PET/CT radiomic features for predicting EGFR gene mutation status in patients with lung adenocarcinoma
- Vol. 32, Issue 4, Pages: 373-380(2023)
- 作者机构：
  
  1. 四川省肿瘤医院 . 研究所，四川省癌症防治中心，电子科技大学医学院核医学科,四川,成都,610041
  2. 川北医学院医学影像学院,四川,南充,637000
- 作者简介：
- 基金信息：
- DOI：10.19732/j.cnki.2096-6210.2023.04.010
  CLC：
- Published Online：13 September 2023，
  
  Published：13 September 2023
- 稿件说明：
移动端阅览
江雪梅,谌佳琪,彭小娟,等. ¹⁸ F-FDG PET/CT影像组学特征对肺腺癌患者EGFR基因突变的预测价值[J]. 肿瘤影像学, 2023, 32(4): 373-380 https://doi.

org/10.19732/j.cnki.2096-6210.2023.04.010
江雪梅,谌佳琪,彭小娟,等. ¹⁸ F-FDG PET/CT影像组学特征对肺腺癌患者EGFR基因突变的预测价值[J]. 肿瘤影像学, 2023, 32(4): 373-380 https://doi. DOI： 10.19732/j.cnki.2096-6210.2023.04.010.

org/10.19732/j.cnki.2096-6210.2023.04.010 DOI：

摘要

目的：

分析治疗前

F-FDG正电子发射体层成像（positron emission tomography，PET）/计算机体层成像（computed tomography，CT）影像组学特征在预测肺腺癌患者表皮生长因子受体（epidermal growth factor receptor，

EGFR

）基因突变状态中作用。

方法：

回顾并分析2017年1月—2021年3月于四川省肿瘤医院就诊的212例（男性115例，女性97例，年龄32~82岁，平均61岁）肺腺癌患者治疗前

F-FDG PET/CT图像及

EGFR

表达资料。将患者按7∶3随机分为训练集（148例）及验证集（64例）。用LIFEx软件手动逐层勾画CT图像的感兴趣区，半自动勾画PET图像的感兴趣区，并提取图像特征参数，使用最小绝对收缩与选择算子（least absolute shrinkage and selection operator，LASSO）及10折交叉验证进行特征筛选并计算模型最佳的λ值，选出最佳的特征参数，使用logistic逐步回归建立模型。使用PET参数、CT参数及PET+CT参数建立3种模型，并结合临床特征即性别、吸烟史建立联合模型。用受试者工作特征（receiver operating characteristic，ROC）曲线分析模型预测

EGFR

突变的价值，获得曲线下面积（area under curve，AUC）、灵敏度、特异度、准确度，用DeLong检验比较上述模型的AUC。

结果：

212例肺腺癌患者中，EGFR基因野生型86例（38.68%），

EGFR

基因突变型126例（61.32%）。从PET参数、CT参数、PET+CT参数中分别筛选出3、2、5个参数构成回归模型，三者验证集的AUC分别为0.719、0.717、0.723，灵敏度分别为76.9%、82.1%、74.4%，特异度分别为68.0%、60.0%、72.0%，准确度分别为67.2%、71.9%、73.4%。加入临床特征后建立联合模型，三者的AUC分别为0.701、0.748、0.704，灵敏度分别为84.6%、71.8%、61.5%，特异度分别为52.0%、72.0%、76.0%，准确度分别为67.2%、73.4%、67.2%。PET、CT、PET/CT模型间的AUC差

异学无统计学意义（P＜0.05）。

结论：

PET、CT、PET/CT影像组学特征参数与肺腺癌

EGFR

基因表达状态相关，PET/CT预测模型的预测效能稍高。

Abstract

Objective:

To assess the value of pre-therapy

F-FDG PET/CT-based radiomic features in predicting epidermal growth factor receptor (

EGFR

) gene mutation status in patients with lung adenocarcinoma.

Methods:

A retrospective analysis was performed on

F-FDG PET/CT images and

EGFR

gene status of 212 patients (115 males and 97 females

aged range: 32-82

averageage: 61 years) with pulmonary adenocarcinoma Between January 2017 and March 2021. Patients were randomly divided into a training set (148 cases) and a validation set (64 cases) in a 7∶3 ratio. LIFEx software was used to delineate the volume of interest of CT and PET images and extract image features. The least absolute shrinkage and selection operator (LASSO) and ten-fold cross- validation were used to conduct feature screening and calculate the best λ value of the model. The best feature parameters were selected and logistics regression was used to establish the model. PET parameters

CT parameters and PET+CT parameters were used to establish three models

and combined with clinical characteristics

namely gender and smoking history

to establish combined models. The receiver operating characteristic (ROC) curve was generated and the area under curve (AUC)

sensitivity

specificity and accuracy were obtained. DeLong test was used to compare the AUC of the above models.

Results:

Among 212 patients with lung adenocarcinoma

86 patients (38.68%) were

EGFR

wild-type and 126 patients (61.32%) were

EGFR

mutant. There were 3

5 parameters were selected from PET parameters

CT parameters and PET+CT parameters respectively to form the regression model. The AUCs of the validation sets of the three models were 0.719

0.717%

0.723

with sensitivities of 76.9%

82.1%

74.4%

spe

cificities of 68.0%

60.0%

72.0%

accuracies of 67.2%

71.9%

73.4%

respectively. The combined models were established after adding clinical features

and the AUCs of the three were 0.701

0.748

0.704

with sensitivities of 84.6%

71.8%

61.5%

specificities of 52.0%

72.0%

76.0%

and accuracies of 67.2%

73.4%

67.2%

respectively. There was no significant difference in AUC between PET

CT and PET/CT models (

＜0.05).

Conclusion:

PET

CT and PET/CT radiomic features were correlated with

EGFR

gene expression status in lung adenocarcinoma

and PET/CT model had a slightly higher predictive efficiency.

关键词

Keywords

references

Views

439

下载量

CNKI被引量

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research progress in predicting EGFR mutation of NSCLC patients using machine learning based on medical imaging

The non-invasive prediction analysis of radiomic features from ¹⁸ F-FET PET/CT dual imaging modality for tumor grading in untreated adult gliomas

The effect of preprocessing methods for ¹⁸ F-FDG PET/CT images from different institutions in distinguishing lung adenocarcinoma EGFR mutation status

Research on a prediction model for superficial lymphoma based on multimodal ultrasound radiomics

Intelligent imaging for precision diagnosis and treatment of pancreatic tumors: current applications and future perspectives

Related Author

SUN Yuanxin

SHEN Leilei

YE Xiaodan

HUA Tao

ZHOU Weiyan

ZHOU Zhirui

HUANG Qi

LI Ming

Related Institution

Shanghai Institute of Medical Imaging

Department of Radiology, Zhongshan Hospital, Fudan University

Department of Nuclear Medicine/PET Center, Huashan Hospital, Fudan University

Department of Radiotherapy, Huashan Hospital, Fudan University

Shanghai Key Laboratory of Molecular Imaging, Shanghai University of Medicine and Health Sciences

AI问答

⁰

The value of 18 F-FDG PET/CT radiomic features for predicting EGFR gene mutation status in patients with lung adenocarcinoma

DOI：10.19732/j.cnki.2096-6210.2023.04.010

摘要

Abstract

关键词

Keywords

references

The value of ¹⁸ F-FDG PET/CT radiomic features for predicting EGFR gene mutation status in patients with lung adenocarcinoma