Medicine
23 June 2026
How DINOv2 helps Orakl Oncology in cancer drug discovery
N
New-ZZZ desk
AI at Meta Blog · 1 month ago
Orakl Oncology, a spin-off of the Gustave Roussy Institute, uses Meta's DINOv2 model to accelerate cancer drug research and development. Instead of traditional analysis, the team applies this model to process vast amounts of organoid images (mini-organs grown in a lab) to more accurately predict patient response to treatment. Using DINOv2 allows them to transition from qualitative image descriptions to precise quantitative data, significantly increasing the efficiency of the process.
Why it matters
- —It demonstrates the practical application of advanced computer vision models (DINOv2) in the critically important field of medicine.
- —It shows how AI allows the transition from manual, qualitative data analysis to precise, quantitative predictions.
- —It accelerates the drug development process, which is critically important for patients with oncological diseases.
Key facts
- Orakl Oncology uses organoids to simulate the reaction of cancer cells to drugs.
- Meta's DINOv2 model proved more effective for organoid image analysis than specialized models.
- The use of DINOv2 increased the accuracy of predicting patient response by 26.8% compared to other methods.
- Thanks to the AI platform, the company significantly reduced development time, achieving results that other biotech companies spent years on.
The full text is in the original source. Here we provide a brief summary and key facts.
