Orakl raises $34M to scale AI platform predicting cancer drug responses

Why it matters: Most experimental cancer drugs fail during clinical development, and Norrsken claims Orakl’s technology can predict clinical trial outcomes with 90% accuracy. The company, spun out of French cancer centre Gustave Roussy in 2023, aims to improve drug-development decisions by testing treatments on laboratory-grown models of patients’ tumours before committing to expensive clinical trials.

In an April preprint, which has not been peer-reviewed, researchers from Orakl and Gustave Roussy reported that the platform’s predictions of median progression-free survival differed from published trial results by an average of 0.82 months.

Zoom in: Orakl grows three-dimensional models of patients’ tumours, known as patient-derived tumour organoids, and tests their responses to drugs, combining the findings with clinical histories and molecular data. The company markets its prediction technology as O-Predict and has identified SCOPE v2 as the AI model behind that offering.

The April SCOPE preprint evaluates a model that was developed using 54 treatment lines from 52 patients and a biobank of 81 organoid lines. Researchers evaluated its predictions against 32 treatment arms from 23 published colorectal and pancreatic cancer trials, correctly identifying the better-performing treatment in seven of eight comparisons.

Follow the money: The financing brings Orakl’s total funding to $50 million (€45 million), following rounds of €3 million in 2023 and €11 million in 2024. New investors include InfraVia, Serena, Bpifrance’s Patient Autonome Fund, teampact.ventures and Plug and Play.

In July, Orakl secured rights to evaluate four undisclosed Sanofi oncology assets, offering a practical test of whether its predictions can guide drug-development decisions.

Reality check: The reported 90% accuracy has important limitations. The preprint primarily validates predictions retrospectively against published trial outcomes, with conventional chemotherapy accounting for 88% of the treatment arms. Whether the platform can reliably predict outcomes for new cancer drugs across different patient populations remains to be demonstrated through independent, prospective studies.

What’s next: Orakl plans to increase experimental data generation fivefold, expand into lung and gastric cancers, and establish teams in the U.S. and U.K. It aims to apply its predictions to more than ten drug programmes and develop the ability to identify and validate additional drug targets that could be pursued alongside existing ones, using patient-derived data and laboratory screening.

Orakl did not respond to a request for comment by the time of publication.