CT-based radiomics for predicting the treatment response to PD-1/PD-L1 inhibitors combined with chemotherapy in unresectable gastric cancer - PubMed
7 hours ago
- #radiomics
- #immunotherapy
- #gastric cancer
- CT-based radiomics model developed to predict immunotherapy response in unresectable gastric cancer.
- Study included 368 patients from two centers, divided into training, internal validation, and external validation cohorts.
- Radiomics signature with 15 features showed good predictive performance (AUC = 0.868 in training, 0.816 in internal validation, 0.793 in external validation).
- Logistic regression model performed best with AUC values of 0.886, 0.831, and 0.826 in respective cohorts.
- Nomogram integrating Radscore and clinical factors demonstrated satisfactory calibration and clinical utility.
- Radscore correlated with immune cell infiltration (activated CD4+ memory T cells, regulatory T cells, CD8+ T cells).
- Study provides a non-invasive, biologically-informed tool for personalized immunotherapy treatment strategies.