The evolution of AI in oncology has progressed from SVMs and classical machine learning to deep learning and big data, now expanding into multi-omics, federated learning, organoid biobanks, and closed-loop AI validation with wet-lab integration.

China now leads in publication volume, surpassing the U.S. in 2022, but shows the lowest level of international collaboration among top countries, while the UK stands out for high cross-border collaboration.

Research hotspots center on AI-driven oncology, with common targets including HER2 and hormone receptors in breast cancer, EGFR in lung cancer, androgen receptor in prostate cancer, CDK4/6, and immune checkpoint inhibitors in liver cancer.

Industry leaders such as Roche, Pfizer, Novartis, AstraZeneca, and Merck are primary players, with major funding coming from the National Natural Science Foundation of China and the U.S. Department of Health and Human Services.

Despite growth, structural weaknesses persist: tumor complexity, data fragmentation and batch effects, limited clinical translation (roughly 5% progress beyond phase 1), and regulatory challenges due to black-box AI models.

The article provides a comprehensive bibliometric analysis of AI-driven anticancer drug design from 2011 to 2025, drawing on over 12,900 original articles and nearly 2,650 reviews after filtering.

Publication growth is surging globally, with an average annual growth rate above 48% since 2018, spurred by milestones like IBM Watson’s trial-matching and AlphaFold’s protein structure predictions.

Top authors include Alex Zhavoronkov, with highly cited researchers largely European and rooted in information science and medicine, signaling interdisciplinary collaboration.

There is a shift from traditional structure-based design toward predicting pharmacodynamic efficacy and exploring biotherapeutics such as vaccines, antibodies, and antibody-drug conjugates; immunotherapy protocols including neoantigen design and T/NK cell therapies are prominent.

Looking ahead, the field aims to expand the druggable space to include protein design and RNA interactions, pursue AI-driven nanoparticles, and reduce costs and development timelines through AI-enabled discovery, with several trials like CV8102 and ISM-series mentioned.

Privacy and regulatory challenges from data breaches underscore the need for clinically driven innovation, multi-omics integration, federated learning, and active validation platforms (CRISPR, organ chips) to link AI predictions to biology.

Harvard, the Chinese Academy of Sciences, and the University of California system rank as leading institutions by output, with two major collaboration networks forming around the US and China.