Answers are grounded in this paper and cite their sources.
Lung cancer remains the leading cause of cancer-related mortality worldwide despite major therapeutic advances, including immune checkpoint inhibitors (ICIs). Although a subset of patients with advanced non-small cell lung cancer (NSCLC) experiences durable long-term responses, currently available predictive biomarkers, including programmed death ligand-1 (PD-L1) expression, have limited ability to reliably distinguish responders from non-responders. Artificial intelligence (AI) and machine learning (ML) have emerged as powerful approaches for extracting clinically meaningful patterns from complex, high-dimensional datasets spanning pathology, liquid biopsy, imaging, and integrated multimodal data. This narrative review examines current applications of AI in NSCLC immunotherapy across these complementary data modalities and evaluates the evidence supporting their clinical translation. Particular attention is given to the methodological limitations of existing studies and to the emerging shift from single-modality biomarkers toward dynamic and multimodal prediction. Although AI-based approaches show potential to complement established biomarkers, limitations in validation, generalizability, and prospective evidence currently restrict their clinical implementation.
No takes yet. Share an insight, caveat, or question.
Nangoli et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: