Multimodal Deep Learning for First-Line Immunotherapy Response in Gastric Cancer

Objective:

To identify patients with gastric cancer who are most likely to benefit from first-line immunotherapy using multimodal deep learning.

Key Findings:

• The CRP model achieved an AUC of 0.97 for tumor response prediction in the training cohort.
• Patients predicted as partial responders had significantly longer progression-free and overall survival compared to non-partial responders.
• The CRP model outperformed unimodal models in both training and validation cohorts.

Interpretation:

The CRP model demonstrates robust predictive performance and generalizability across different cohorts, indicating its potential utility in clinical settings for predicting immunotherapy responses.

Limitations:

• The study is retrospective and may be subject to biases inherent in such designs.
• External validation cohorts were limited in size.

Conclusion:

Multimodal deep learning approaches like the CRP model can enhance the prediction of treatment outcomes in gastric cancer patients undergoing immunotherapy.