PET/CT Radiomics and Clinical Features for Classifying Transcriptome-Defined Immune Subgroups in Non-Small Cell Lung Cancer: A Retrospective Study
Journal of Visualized Experiments · 2026-11-10 · (237)
DOI: 10.3791/74396 ↗At a glance
Journal article in Journal of Visualized Experiments, published online on 2026-11-10. Written by Liqun Huang and 4 co-authors.
Abstract (original language)
Tumor immune phenotypes are clinically relevant in non-small cell lung cancer (NSCLC), but tissue-based assessment is invasive and vulnerable to sampling bias. This retrospective observational study evaluated whether positron emission tomography/computed tomography (PET/CT) radiomic features combined with clinical variables could classify transcriptome-defined immune subgroups. Data from 184 patients were analyzed between January 2021 and December 2024 and used a computer-generated stratified 7:3 split to form a modeling cohort (n = 129) and an internal validation cohort (n = 55). Consensus clustering of transcriptomic profiles defined two immune subgroups. After reproducibility filtering, minimum redundancy maximum relevance analysis and least absolute shrinkage and selection operator (LASSO) logistic regression retained six radiomic features. The final combined model included the radiomic score, smoking history, pathological type, and pleural adhesion. The combined model achieved an area under the receiver operating characteristic curve (AUC) of 0.841 (95% confidence interval [CI], 0.772-0.909) in the modeling cohort and 0.822 (95% CI = 0.712-0.932) in the validation cohort. Compared with the radiomics-only model, the AUC gain was significant in the modeling cohort (p = 0.014) but not in the validation cohort (p = 0.590). At the modeling-derived threshold of 0.502, sensitivity and specificity were 76.5% and 82.0% in the modeling cohort and 75.9% and 69.2% in the validation cohort. Calibration intercepts and slopes were approximately 0 and 1.000 in the modeling cohort and -0.229 and 1.310 in the validation cohort. These findings support an internally evaluated association between routine PET/CT phenotypes and transcriptome-defined immune subgroups, but external validation and recalibration are required before clinical use.
Cite
APA
Huang, L., Dai, Y., Su, Y., Wang, J., & Xie, L. (2026). PET/CT Radiomics and Clinical Features for Classifying Transcriptome-Defined Immune Subgroups in Non-Small Cell Lung Cancer: A Retrospective Study. Journal of Visualized Experiments. https://doi.org/10.3791/74396
Vancouver
Huang L, Dai Y, Su Y, Wang J, Xie L. PET/CT Radiomics and Clinical Features for Classifying Transcriptome-Defined Immune Subgroups in Non-Small Cell Lung Cancer: A Retrospective Study. Journal of Visualized Experiments. 2026. doi:10.3791/74396
BibTeX
@article{Huang2026,
author = {Huang, Liqun and Dai, Yangbin and Su, Yingrui and Wang, Jianlin and Xie, Long},
title = {PET/CT Radiomics and Clinical Features for Classifying Transcriptome-Defined Immune Subgroups in Non-Small Cell Lung Cancer: A Retrospective Study},
journal = {Journal of Visualized Experiments},
year = {2026},
number = {237},
doi = {10.3791/74396}
}
EUVIMED