Idibine Presents AI Clinical Trial Success Prediction Results at ESMO 2026

Validated 291 cancer drug development cases
27 of 28 in the "High-Confidence Group" advanced to subsequent clinical trials

Idibine, a company that predicts clinical trial success using AI, announced on the 20th that a study in which it used its proprietary artificial intelligence (AI) model to predict the likelihood of cancer drugs succeeding in clinical trials had been selected as an electronic poster for the 2026 European Society for Medical Oncology (ESMO) Congress. The conference will take place in Madrid, Spain, from October 23 to 27.

ESMO is considered one of the world's three leading oncology organizations, alongside the American Society of Clinical Oncology (ASCO) and the American Association for Cancer Research (AACR). This is the first time Idibine's research has been presented at ESMO. The abstract was selected in the field of "AI for Clinical Research and Drug Development."

The study is titled "Confidence-Stratified AI Prediction for Successful Oncology Drug Pipelines: A Deep Ensemble Prediction Interval Approach." It used "i.Dv 1.5," a model newly developed by Idibine. The model combines predictions from 100 AI models trained differently to calculate not only the probability of success but also how much the predictions fluctuate and how stable they are.

Based on the analysis results, cancer drug candidates are divided into three tiers: the "High-Confidence Group (Go)," with high predicted success probability and confidence; the "Watch Group (Watch)," which requires further review; and the "Reassessment Group (Pass)," which requires reconsideration due to high uncertainty.

After validating 291 cancer drug development cases across three cancer types, including non-small cell lung cancer, breast cancer and colorectal cancer, the research team found that of the 94 cases the AI model predicted to have a high probability of success, 84 (89.4%) actually advanced to the next clinical trial stage or received FDA approval. Among the 28 cases classified in the High-Confidence Group, 27 advanced to later-stage clinical trials or received approval, resulting in a precision rate of 96.4%.

Idibine expects these findings to be useful in helping pharmaceutical and biotech companies set development priorities, select drug candidates for technology transfer, and allocate research and development resources. Companies can weigh both the likelihood of success and the uncertainty of predictions when choosing which projects to pursue first, while investors can use the findings as a reference when assessing candidates' potential for clinical development and their risks.

Co-CEO Lee Siyoung said, "In drug development, how much a prediction can be trusted is just as important as the number representing the probability of success." Lee added, "This study provides practical evidence for developing differentiated strategies for each pipeline, moving beyond the binary approach of pass or fail."

This content was produced with the assistance of AI translation services.