文献速递胸部PETCT

基于PET-CT的影像组学模型用于术前预测临床I期纯实性浸润性肺腺癌的新IASLC分级及预后

Development of a PET-CT Based Radiomics Model for Preoperative Prediction of the Novel IASLC Grading and Prognosis in Patients with Clinical Stage I Pure Solid Invasive Lung Adenocarcinoma.
PubMed 文献 · Acad Radiol · 2025-03-22

研究回顾性纳入418例临床I期纯实性浸润性肺腺癌患者,基于术前18F-FDG PET/CT提取影像组学特征构建列线图预测IASLC分级和RFS。影像组学模型在训练集和测试集AUC分别为0.838和0.768,整合列线图预测性能优于CT模型(p=0.001)和PET模型(p=0.028),且预测分级可显著分层RFS(p<0.001)。

要点

  • 基于PET/CT的影像组学列线图可术前预测IASLC分级,训练集AUC 0.838、测试集AUC 0.768。
  • 整合列线图预测性能优于CT模型(p=0.001)和PET模型(p=0.028)。
  • 列线图预测的IASLC分级可显著分层RFS,低级别与高级别5年生存率差异显著(p<0.001)。

核心结论

  • 基于术前PET/CT的影像组学列线图是预测临床I期纯实性浸润性肺腺癌IASLC分级和RFS的潜在生物标志物。

研究设计与样本

  • 回顾性纳入418例临床I期纯实性浸润性肺腺癌患者,均行术前18F-FDG PET/CT检查;按IASLC分级分为低级别组(I级和II级,n=315)和高级别组(III级,n=103),并按7:3随机分为训练集(n=292)和测试集(n=126);从CT和PET图像的全肿瘤区域提取影像组学特征,多因素分析确定IASLC分级和RFS的独立预测因素,联合Radscore、临床和影像特征构建列线图。

诊断标准与关键数值

  • 影像组学模型在训练集和测试集的AUC分别为0.838和0.768;多因素logistic回归显示较高的SUVmax和空洞存在是IASLC分级的独立危险因素;整合列线图在训练集预测性能优于CT模型(p=0.001)和PET模型(p=0.028);列线图预测的IASLC分级可显著分层RFS,低级别与高级别肺腺癌5年生存率差异显著(p<0.001)。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

RATIONALE AND OBJECTIVES: To develop and validate a fluorine-18-fludeoxyglucose (18F-FDG) PET/CT-based radiomics nomogram for preoperative prediction of the International Association for the Study of Lung Cancer (IASLC) grading and recurrence-free survival (RFS) in patients with clinical stage I pure-solid invasive lung adenocarcinoma (LADC). MATERIALS AND METHODS: 418 patients with clinical stage I pure-solid invasive LADC who underwent preoperative 18F-FDG PET/CT examination were retrospectively enrolled. All patients were separated into the low-grade group (grade I and II; n=315) and the high-grade group (grade III; n=103) according to the IASLC grading system, and the cohort was randomly divided into a training set (n=292) and a testing set (n=126) at a ratio of 7:3. Radiomics features were extracted from CT and PET images in regions of the entire tumor. Multivariate analysis identified the independent predictors for IASLC grading and RFS. The Radscore, along with clinical and radiological features were combined to establish a predictive nomogram.

RESULTS: The ultimate Radiomics model, achieving AUCs of 0.838 and 0.768 in the training and testing sets. The multivariate logistic regression showed that higher maximum standard uptake value (SUVmax), cavity presence are the independent risk factors for IASLC grading. The integrated nomogram showed superior prediction performance than CT model (p=0.001) and PET model (p=0.028) in the training set. Furthermore, both pathological grade and preoperatively predictive IASLC grade derived by nomogram significantly stratified patients for RFS, with 5-year survival rates showing marked differences between low-grade and high-grade LADC (p<0.001).

CONCLUSION: The preoperative PET/CT-based radiomics nomogram represents a potential biomarker for predicting IASLC grade and RFS in patients with clinical stage I pure-solid invasive LADC.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Lan J, Wang H, Xin E, et al. Development of a PET-CT Based Radiomics Model for Preoperative Prediction of the Novel IASLC Grading and Prognosis in Patients with Clinical Stage I Pure Solid Invasive Lung Adenocarcinoma. Acad Radiol. 2025. PMID: 40121117. doi:10.1016/j.acra.2025.02.017

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。