中国中西医结合影像学杂志2025，Vol.23Issue(2)：168-173,6.DOI:10.3969/j.issn.1672-0512.2025.02.007
双能量CT多参数图像的影像组学模型列线图对头颈部鳞状细胞癌PD-L1表达水平的预测价值
Preoperative prediction of PD-L1 expression in head and neck squamous cell carcinoma using dual-energy CT radiomics nomogram
摘要
Abstract
Objective:To develop and validate a radiomics nomogram based on dual-energy CT(DECT)multiparametric features for preoperative prediction of programmed death-ligand 1(PD-L1)expression in head and neck squamous cell carcinoma(HNSCC).Methods:This study enrolled 173 pathologically confirmed HNSCC patients who underwent dual-phase DECT enhancement.Patients were randomly allocated to a training cohort(121 cases)and a validation cohort(52 cases)at a 7∶3 ratio.Immunohistochemical PD-L1 status defined negative and positive subgroups.Volumes of interest(VOIs)were segmented on polyenergetic images and co-registered to iodine density(ID)images,effective atomic number(Zeff)images,and virtual monoenergetic images(VMI).Radiomics features were extracted from all parametric maps,followed by dimensionality reduction.Support vector machine(SVM)classifiers were built for single-parameter and multiparametric fusion models.The ROC curve was used to evaluate the predictive performance of these models.The model with the highest performance was selected as the best radiomics model,and its results were converted into radiomics scores(Rad-score).Then,the Rad-score was incorporated into a nomogram with clinical risk factors identified via multivariate logistic regression.Results:The multiparametric fusion model achieved superior PD-L1 prediction performance with AUC of 0.918 and 0.859 for the training cohort and the validation cohort.Age<60 years old and oral cavity/oropharyngeal tumor location were independent clinical predictors.The combined nomogram integrating Rad-score with clinical factors demonstrated enhanced discrimination,AUC of 0.946 and 0.909 for the training cohort and the validation cohort.Conclusions:The DECT-derived radiomics nomogram provides a noninvasive tool for preoperative PD-L1 status assessment in HNSCC,potentially guiding immunotherapy strategies.关键词
头颈部鳞状细胞癌/程序性死亡配体1/体层摄影术,X线计算机/能量成像/影像组学Key words
Head and neck squamous cell carcinoma/Programmed death-ligand 1/Tomography,X-ray computed/Spectral imaging/Radiomics引用本文复制引用
王伟康,陈炜越,应海峰,夏水伟,纪建松..双能量CT多参数图像的影像组学模型列线图对头颈部鳞状细胞癌PD-L1表达水平的预测价值[J].中国中西医结合影像学杂志,2025,23(2):168-173,6.基金项目
浙江省医药卫生科技计划一般项目(2024KY568). （2024KY568）
