Generative Artificial Intelligence as a Catalyst for Effective Cancer Treatments
Bibliographic record
Abstract
Current standard of cancer care is supported by a continuously expanding set of therapeutic options becoming available to cancer patients such as KRAS inhibitors, third-generation tyrosine kinase inhibitors (e.g., osimertinib), new immunotherapy drugs (e.g., durvalumab), as well as adaptive and combination therapies involving chemotherapy, radiotherapy, immunotherapy and targeted therapy. Despite these advances, therapeutic resistance persists as an inevitable challenge to cancer cure. While combination therapy is a plausible strategy to thwart therapeutic resistance, further research is needed on rationalizing drug combinations. On the other hand, adaptive therapy is emerging as a sound strategy to counter the evolving nature of cancer and thwart or delay the onset of therapeutic resistance. Indeed, cancer is a nonlinear time-varying dynamical system whose treatment can be viewed as a problem of steering the disease to a desired end-state of cure or stable management based on monitored treatment response. The success of this strategy depends on accurate and reliable estimations/predictions of disease state and tumor growth dynamics. Therein lies the potential of generative artificial intelligence (GenAI) and its underlying large language models (LLMs) to leverage the accumulating big clinical, radiomic and molecular data about cancer patients and their treatments to power its learning and predictive capabilities towards assisting treatment decision-making. This perspective starts with examples of current therapeutic advances and a succinct overview of persisting challenges to the development of effective cancer treatments, followed by a broad survey of artificial intelligence (AI) applications in oncology and their varying degrees of clinical readiness. Given this context, cancer treatment is framed as a problem of controlling a nonlinear time-varying dynamical system, where data-driven GenAI learning and predictive capabilities would be instrumental in resolving the challenges of disease monitoring and controllability. The perspective shares insights about potential pathways to GenAI-assisted improvement of cancer treatments and discusses key challenges to its deployment in real-world clinical settings, including data curation, clinical validation, LLM hallucinations and ethical concerns. Ultimately, advancing noninvasive tracking of treatment response dynamics and curating corresponding big LLM training datasets are essential to the potential of GenAI as a catalyst for effective cancer treatments.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.006 |
| Scholarly communication | 0.004 | 0.003 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.006 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".