Integrative Multimodal Biomarker and AI-Driven Prognostic and Predictive Modeling in Stage IV Non-Small Cell Lung Cancer: a Comprehensive Review

Purpose of Review: This review examines the potential application of artificial intelligence (AI), machine learning (ML), and deep learning (DL) in advancing a prognostic and predictive model for advanced-stage IV non-small cell lung cancer. It covers a wide range of factors, including the Neutrophil to Lymphocyte Ratio (NLR), Platelets to Lymphocyte Ratio (PLR), Lymphocyte to Monocyte Ratio (LMR), Advanced Lung Cancer Inflammation Index (ALI), Eastern Cooperative Oncology Group (ECOG) performance status, LDH, Albumin, PD-L1 Combined Positive Score (CPS), Tumor Mutational Burden (TMB), circulating tumor DNA (ctDNA), long non-coding RNAs (lncRNAs), histone modifications, chromatin remodeling, and radiomics features. This multidimensional analysis seeks to integrate diverse data sources to enhance the accuracy of prognostic models, thereby aiding clinical decision-making and supporting treatment-specific predictive modelling where sufficient evidence exists. Recent Findings: Recent advancements have demonstrated the effectiveness of AI in processing multi-omics data, enabling the integration of biochemical, clinical, and radiological frameworks into predictive models. Different AI algorithms have improved prognostic accuracy, identified patient subgroups likely to benefit from targeted therapies, and predicted treatment outcomes. Nonetheless, AI-driven radiomics and Ct-DNA-based biomarkers have been particularly useful. However, a wide range of molecular biomarkers for prognostic and predictive modeling in stage IV NSCLC, collectively and incorporating AI techniques, is needed. Summary: AI techniques represent a paradigm shift in prognostic modeling for advanced NSCLC, offering unparalleled opportunities for precision oncology. By dissecting the intricacies of each parameter, the review provides profound insights into their individual and combined contributions to creating an effective prognostic model for stage IV lung cancer and highlights selected biomarkers with treatment-specific predictive value, particularly in the context of immunotherapy and targeted therapies. This article lays the groundwork for future research aimed at improving both prognostic and predictive accuracy to enhance patient outcomes in oncology.