| AI in Clinical Medicine, ISSN 0000-0000 online, Open Access |
| Article copyright, the authors; Journal compilation copyright, AI Clin Med and Elmer Press Inc |
| Journal website https://aicm.elmerpub.com |
Review
Volume 1, 2025, e5
Generative Artificial Intelligence as a Catalyst for Effective Cancer Treatments
Table
| Capabilities of LLMs | Potential support for effective cancer treatments |
|---|---|
| GenAI: generative artificial intelligence; LLMs: large language models. | |
| Foundational models for rapid learning from big multimodal data | LLMs can translate big multimodal data about cancer and its treatments into learned models (e.g., treatment response models) that can be used as decision-support tools for clinicians and tumor boards [137]. |
| High-dimensional reasoning capacity | LLMs have the potential to make timely recommendations of patient-tailored, optimized treatment schedules based on a comprehensive consideration of molecular, clinical radiomic and demographic variables in light of a dynamic space of available therapies [137, 138]. |
| Open to continuous learning through retraining and fine-tuning | Trained LLMs can be integrated to support adaptive therapy by providing real-time disease state estimation and treatment planning based on treatment response monitoring [39]. GenAI would enable a more accurate tracking of disease state dynamics, enhancing as a result the benefits of adaptive therapy, which include: 1) lower treatment toxicity [127, 128, 133]; 2) resistance management [121, 123, 126]; 3) treatment personalization [39]. |
| LLMs can be periodically retrained or fine-tuned to account for the dataset shift [80], hence maintaining their optimal utility to the treatment of the target patient population. | |