Artificial intelligence derived virtual oncology

Tumors are dynamic multicellular ecosystems in which heterogeneity, evolution, and cross-scale interactions remain major barriers to precision oncology. Although multiple technological advances have improved tumor characterization, they remain largely static and descriptive, limiting their ability to capture state transitions, treatment-induced remodeling, and evolving intercellular communication. Artificial intelligence virtual cell (AIVC) provides a foundation for modeling cellular states, perturbation responses, and functional outputs from multimodal data. We extend this concept to artificial intelligence virtual oncology (AIVO), a framework designed to reconstruct multicellular communication, coordinated evolution, and system-level remodeling among tumor, immune, and stromal compartments. This review summarizes the conceptual basis, construction strategies, and emerging applications of this progression from AIVC to AIVO in tumor diagnosis, dynamic monitoring, treatment response prediction, and drug screening. We further discuss key challenges in data integration, interpretability, patient-specific calibration, and clinical validation. Grounded in AIVC-based modeling, we propose AIVO as an emerging framework for multimodal integration, dynamic modeling, and decision support in future oncology.