Development and validation of AI-assisted transcriptomic signatures to personalize adjuvant chemotherapy in patients with resectable pancreatic ductal adenocarcinoma.
Bibliographic record
Abstract
4015 Background: Adjuvant chemotherapies for PDAC include modified FOLFIRINOX (mFFX) or gemcitabine-based regimen in fit patients and gemcitabine or 5FU single agents in other patients. While more effective, mFFX is associated with a greater toxicity than other options. Moreover, therapeutic decisions still rely mainly on the patient's performance status rather than tailored to tumor-based criteria. Our study aims to personalize treatments by developing transcriptomic signatures specific to commonly used drugs for pancreatic cancer. Methods: We analyzed the response to drugs (5-fluorouracil, oxaliplatin, and irinotecan) in three types of preclinical models (primary cell cultures, tumoroids, and patient-derived xenografts in immunodeficient mice). We then associated the detected sensitivities to drugs with transcriptomic data from each model. We also incorporated the previously developed gemcitabine signature. Finally, we used a machine learning method, the "Least Absolute Shrinkage and Selection Operator-random forest," to improve the signatures, integrating the tumor microenvironment master regulators. The learning cohort were GemPred for gemcitabine (1) and COMPASS (2) for mFFX. The resulting transcriptomic predictive tool was called Pancreas-View. We validated these signatures in the PRODIGE-24/CCTG PA6 trial cohort comprising 343 patients (3). Results: The signatures may allow to identify responsive patients to specific drugs and showed a significant improvement in their cancer-specific survival (CSS) and disease-free survival (DFS) when they received a matched therapy (mFFX or gemcitabine). Additionally, a positive association was observed between the number of drugs for which tumors predict to be sensitive and patient’s survival when appropriately treated. Patients who received “appropriate” drugs (n = 164; 47.8%) displayed a longer DFS : 50.1 months (stratified HR: 0.31; 95% CI, 0.21-0.44; p < 0.001) in the mFFX arm, and 33.7 months (stratified HR: 0.40; 95% CI, 0.17-0.59; p < 0.001) in the gemcitabine arm, respectively. Conversely, patients that received a treatment not matched with the signature prediction (n = 86; 25.1%) and those predicted to be resistant to all drugs (n = 93; 27.1%) had the poorest DFS results (10.6 and 10.8 months, respectively). Conclusions: By integrating preclinical models and machine learning, we developed a comprehensive predictive tool based on the transcriptome that may help to identify tumors sensitivity to mFFX components and gemcitabine. Crucially, these transcriptomic signatures can also lead to reduce toxicity by avoiding the unnecessary administration of drugs predicted as ineffective for a given tumor. Nicolle R, et al. Ann Oncol 2021;32:250-260. Aung K, et al Clin Cancer Res 2018;24:1344–1354. Conroy T, et al. N Engl J Med 2018; 379:2395-2406.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".