Multimodal model fuses CT deep learning with ProGRP and bone metastasis to predict 1-year survival in ES-SCLC immunotherapy

Combining mediastinal-window CT deep learning with ProGRP and bone metastasis yielded an AUC of 0.902 for predicting 1-year overall survival in extensive-stage small cell lung cancer patients receiving first-line immunotherapy, outperforming CT alone (AUC 0.813).

• ProGRP and bone metastasis were independent prognostic factors for 1-year OS (AUCs 0.705 and 0.666).
• Mediastinal-window CT deep learning (MWDL) achieved higher AUC (0.813) than lung-window CT model (0.756).
• The multimodal MWGB-Cox model’s predictive performance was validated in an independent internal cohort (n=25).

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.