Non-invasive prediction of prognostic immune subtypes in lung adenocarcinoma using PET/CT-based radiomics - PubMed
3 days ago
- #Radiomics
- #Immune Subtypes
- #Lung Adenocarcinoma
- Study addresses clinical challenges in LUAD treatment and prognosis by developing a non-invasive PET/CT radiomics biomarker to predict immune subtypes.
- Three immune subtypes were identified via consensus clustering: Inflammatory (ClusterA), Activated (ClusterB), and Evasive (ClusterC), with significant prognostic differences.
- A radiomics machine learning classifier achieved high accuracy (AUC = 0.84) in predicting these subtypes from PET/CT data.
- External validation in an independent cohort showed immunohistochemical patterns consistent with predicted subtypes and stratified survival in TKI- and ICI-treated patients.
- The approach integrates genomics and radiomics to enable non-invasive patient stratification and guide personalized therapy in LUAD.