What are the current hot topics in the study of critical points during the process of cell mutation and development into tumors? How to conduct in-depth research on the future use of AI?

Tumorigenesis is not a linear accumulation of mutations but rather a dynamic process punctuated by critical transition points—windows during which normal cells cross irreversible thresholds toward dysplasia, carcinoma in situ, and ultimately invasive malignancy. Understanding the molecular and ecological drivers of these “tipping points” is fundamental to early interception and prevention. I would like to ask the cancer biology and computational oncology community: Which specific transition stages are currently receiving the most intense investigation—for example, the switch from field cancerization to clonal expansion, immune evasion checkpoints, metabolic reprogramming gates, or senescence escape mechanisms? What experimental models (organoids, lineage tracing, in vivo barcoding) are proving most informative? More importantly, how should AI be deployed to go beyond conventional omics pattern recognition? Are we moving toward AI systems capable of inferring causal trajectories from multi-modal longitudinal data, predicting individual-specific transition risks, or identifying hidden attractor states in epigenetic landscapes? I am also keen to hear critical perspectives on current limitations—what can AI not yet do in this domain, and where are the gaps in training data or biological theory? Insights from systems biology, single-cell genomics, and AI-driven drug discovery are all warmly welcomed. Thank you.