Artificial Intelligence Empowering Oncology Precision Medicine: Current Status, Key Insights, and Future Perspectives

Objective: The aim of this study was to comprehensively assess the current status, hot topics, and future trends of artificial intelligence (AI) in precision oncology research worldwide. Methods: Based on the Web of Science Core Collection, this study used multiple tools to analyze the application of AI in precision oncology from multiple dimensions, including publication volume, authors, institutions, journals, and countries. In addition, text mining methods were used to quantify the size of tumor types and algorithms and their relationships, and in-depth analyses were performed. Results: To date, 709 relevant papers have been published in the field, covering 332 journals, 64 countries, and a total of 4675 researchers. The most commonly used algorithms are neural networks, random forests and support vector machines. The most studied tumor types include gastric, colorectal, breast, and lung cancers. Recent advances emphasize the integration of AI with imaging genomics, multi-omics, and other data types for multimodal data fusion and applications in biomarker screening and identification. Conclusions: The field is currently in a phase of steady development and has attracted extensive attention from several countries, institutions, and researchers worldwide. However, stable collaborative research clusters have not yet been established. The continuous development of algorithms, especially the application of neural networks, random forests and support vector machines in precision medicine for tumors, has demonstrated significant advantages. Looking forward, AI and AI-assisted biomarker studies incorporating multimodal analysis are expected to gain more research interest and attention and play an important role in advancing precision oncology.