LAUSR

The advancement in immunotherapy has opened a new epoch in cancer therapy (Wang et al., 2021). However, only subsets of patients have yielded considerable benefits from immunotherapy (Sharma et al., 2021). The unsatisfactory efficacy may primarily arise from the restricted understanding of resistance mechanisms, lack of robust biomarkers, and unavoided immune-related adverse events. Thereby, new assessment approaches are imperative to optimize immunotherapy and improve clinical outcomes. With advancements in high-throughput sequencing techniques and computational biology, numerous predictive tools combining clinical and molecular traits displayed excellent performance to robustly address the classification of tumor types, prediction of prognosis, and immunotherapy response (Wang et al., 2022a; Liu Z. et al., 2022). This Research Topic assembled eight original research articles, which applied artificial intelligence to develop novel targets, drugs, and immunotherapy predictive markers, from large-scale data, and provided excellent paradigms on how to exploit artificial intelligence to generate reliable predictive tools (Figure 1). Despite immune checkpoint inhibitors (ICIs) exhibiting durable responses and prolonged survival across multiple tumors, lower response rates, higher resistance rates, and costs hinder clinical utilization (You et al., 2020), which prompts researchers to develop novel biomarkers predicting the efficacy of immune checkpoint treatment (ICT). Currently, numerous biomarkers including microsatellite instability (MSI) (André et al., 2020), programmed death-ligand 1 (PD-L1) expression (Lin et al., 2018), and tumor mutational load (TMB) (Samstein et al., 2019) have been proposed to identify susceptibility to ICT. Nevertheless, previous studies reported no significant correlation between these biomarkers and ICT response, and some are even contradictory (Wu et al., 2019; Bai et al., 2020). Complex mechanisms of anti-tumor OPEN ACCESS