Machine learning for prediction of immunotherapeutic outcome in non-small-cell lung cancer based on circulating cytokine signatures

The CIRI model is highly accurate and reproducible in determining the patients with NSCLC who would benefit from anti-PD-1/PD-L1 therapy with prolonged OS and may aid in clinical decision-making before and/or at the early stage of treatment.