Related Experiment Video
Updated: Sep 19, 2025
Multiplexed Immunofluorescence Analysis and Quantification of Intratumoral PD-1+ Tim-3+ CD8+ T Cells
Published on: February 8, 2018
Deep learning and inflammatory markers predict early response to immunotherapy in unresectable NSCLC: A multicenter
1Department of Thoracic Surgery, Northern Jiangsu People's Hospital Affiliated to Yangzhou University, Yangzhou, China; Institute of Translational Medicine, Medical College, Yangzhou University, Yangzhou, China.
A new deep learning model combined with the systemic immune-inflammatory-nutritional index (SIINI) can predict early responses to immune checkpoint inhibitors (ICIs) in non-small cell lung cancer (NSCLC). This approach offers a noninvasive method to personalize treatment and improve survival outcomes for patients.
Area of Science:
- Oncology
- Radiology
- Artificial Intelligence
Background:
- Immune checkpoint inhibitors (ICIs) show variable efficacy in unresectable non-small cell lung cancer (NSCLC).
- There is a critical need for noninvasive biomarkers to predict early therapeutic responses and enhance survival in NSCLC patients undergoing ICI treatment.
Purpose of the Study:
- To develop and validate a deep learning model integrated with the systemic immune-inflammatory-nutritional index (SIINI) for early prediction of ICI response in unresectable NSCLC.
- To assess the model's predictive performance and interpretability using CT-based radiomic features and clinical data.
Main Methods:
- A retrospective multicenter study involving 265 patients with unresectable NSCLC treated with ICIs.
- Development of a deep learning model (DenseNet121) incorporating radiomic features from chest CT scans and SIINI scores.
- Validation of the model on internal (30%) and external datasets, with performance evaluated using Area Under the Curve (AUC).
- Utilized Gradient-weighted Class Activation Mapping (Grad-CAM) for model interpretability.
Main Results:
- The combined deep learning and SIINI model demonstrated strong predictive performance.
- Achieved an AUC of 0.865 (95% CI: 0.7709-0.9595) in the internal validation cohort.
- Achieved an AUC of 0.823 (95% CI: 0.6627-0.9827) in the external validation cohort.
- Grad-CAM analysis identified key CT regions influencing predictions, enhancing clinical interpretability.
Conclusions:
- Integrating deep learning with radiomic features and the SIINI biomarker shows significant potential for predicting ICI response in unresectable NSCLC.
- This approach supports personalized ICI therapy by offering early prediction of treatment efficacy.
- Future research should focus on multi-omics integration, larger multicenter validation, and increased sample sizes to improve predictive accuracy and clinical translation.
