Abstract B011: Pan-cancer immunotherapy response prediction using the CURE AI large clinicogenomic foundation model
Bibliographic record
Abstract
Abstract We developed a machine learning platform that learns and measures the individualized benefit of a therapeutic intervention over standard of care, which we named CURE AI (Clinical trials Uncovering Real Efficacy Artificial Intelligence). CURE AI is a large clinicogenomic foundation model (LCGM) trained on clinical and multi-omics data from hundreds of thousands of oncology patients using a proprietary deep learning architecture and training schema. CURE AI calculates a counterfactual outcome for every patient, resulting in a continuous benefit prediction ranking of patients based on the most important and predictive clinical and genomic factors, which can be utilized to inform biomarker and target discovery. We have previously demonstrated that CURE AI, trained on lung cancer clinical trials (CURE Lung Cancer), could predict immunotherapy response on additional held-out lung cancer clinical trials. In this study, we evaluated how well CURE AI could perform cross-cancer immunotherapy response predictions. We applied CURE Lung Cancer to a clear cell renal cancer clinical trial (ccRCC) and predicted immunotherapy clinical response in ccRCC, which is the first cross-cancer utilization of an LCGM. We further show that CURE AI, finetuned on ccRCC immunotherapy data (CURE Renal Cancer) predicts immunotherapy response in lung cancer. In a pan-cancer TCGA analysis, CURE AI, trained on renal or lung data, successfully stratified pan-cancer patients by immunotherapy response (poor: low grade glioma, prostate, pancreas and high: melanoma, hepatocellular, leukemia and lymphoma) enabling informed indication expansion decisions. CURE AI-refined eligibility criteria allows for >50% trial size reduction while accelerating the time to trial significance by 6-18 months with calculated potential cost savings in the order of > $100M for an average phase 3 clinical trial. CURE AI has significant potential to lead to advancements in refining clinical trial eligibility and informing indication expansion decisions. Citation Format: Vitalay Fomin, Amit Weiss, Neil T. Pfister. Pan-cancer immunotherapy response prediction using the CURE AI large clinicogenomic foundation model [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Artificial Intelligence and Machine Learning; 2025 Jul 10-12; Montreal, QC, Canada. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(13_Suppl):Abstract nr B011.
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 | 0.002 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 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".