OwkinAI Drug Development Solutions

We unlock insights from multimodal datasets created and curated by world-class medical academic centers

In collaboration with researchers, we train AI models for optimizing clinical trials, identifying patient subgroups, and biomarker discovery. Because the promise of AI for drug development begins with the highest quality data. 

Get in touch

We help you meet clinical research challenges

Leveraging our partner datasets, we deploy state-of-the-art, interpretable AI to refine our understanding of diseases. We build machine learning models (70+) across a variety of therapeutic areas that support drug development efforts and market access with:

AI Drug development


Intended for translational researchers and pathologists, these machine learning models help identify novel, medically relevant, predictive and prognostic patient characteristics from histology and molecular connected data. In breakthrough instances, they help see the previously unseen.
AI Drug development

AI patient identification tools

Intended for clinical development or precision medicine teams, these models help identify patients with specific genetic alterations or biomarkers that are typically difficult and costly to test— permitting clinicians to deploy therapeutic pathways that are more effective and precise to each patient.
AI Drug development

Trial solutions

Intended for translational researchers, clinical development and commercial teams, these models create external control arms, and can be used across all stages of the clinical drug development process to enhance clinical trial design (RCTs) and evaluation, as well as to optimize product strategy.

Power precision medicine across the development pipeline

The Owkin difference

AI Drug development

Critical clinical context and unique data network

First in class collaborations with top research institutions and KOLs for actionable relevant data sets.
AI Drug development

Cutting edge technology to prioritize privacy

Federated Learning permits secure, confidential, and traceable training of AI models across multiple data centers.
data alliance network

Unique combinations of data types for the complete picture

Multimodal machine learning generates novel insights across heterogenous data and can reveal valuable HistoGenomic signals.
AI Drug development

The black box demystified for better understood outcomes

Owkin AI is interpretable, allowing us to understand underlying mechanisms.

Case Studies

Better Identification of High-Risk Patients with Breast Cancer in the Race AI Project

Breast cancer, when discovered early, has a favorable long-term prognosis. Nevertheless, the incidence of breast cancer is high, and approximately 10% patients relapse after initial treatment each year. Currently, no cure exists for these patients, as a result the disease goes into a “chronic” phase with long-term treatment and management.

Together with Gustave Roussy in Paris, France, we have developed an AI diagnostic tool to better identify these high risk patients. Should it be novel hormonal therapies, additional chemotherapy, or other innovative treatments, we are now more able to accelerate the clinical utility of certain treatments and give the patient the best possible care pathway.

“Thanks to this solution we now have a better understanding of the underlying mechanism of highly aggressive tumors and the treatment needs for these patients. Identifying very high risk patients earlier will enable us to adjust the therapeutic strategy for more favorable patient outcomes.”

Prediction of genomic defects from histology images: better prediction of PARP inhibitor benefits in patients with Ovarian Cancer

Ovarian cancer is the first cancer where DNA repair defects in tumors were identified as a target for a new generation of drugs that inhibit PARP activity. However, current methods to identify these defects, based on genomic sequencing, are costly and neither sensitive nor specific enough.

Together with Gustave Roussy in Paris, France, we are developing an Patient Identification Tool that uses routine histology images to recognize and understand tumors with these genomic defects, and subsequently identify patients who could benefit from this treatment in both 1st line and relapsed settings. Furthermore, we can use this novel AI approach across many cancers to better identify patients eligible for targeted therapies.

Image caption: HRD score for scale. Images of survival curves without Drug name use PARP inhibitor instead.

“Such a tool works using little tumor tissue, could easily be used in any center, and answers an urgent need for robust and clinically feasible predictive biomarkers of PARPi benefit in ovarian cancer and beyond.”

Our therapeutic focus areas

Owkin partners focus their research on specific therapeutic areas, pathways, or mutations. As such, we create targeted alliances designed to answer some of the most challenging research questions for a myriad of unmet medical needs.

Explore how we boost medical research with Innovative Collaboration.

AI Drug development


AI Drug development


AI Drug development


AI Drug development


AI Drug development

B Cell Neoplasm (Multiple Myeloma, Lymphoma)

AI Drug development


AI Drug development


Recent publications

Integrating deep learning CT-scan model, biological and clinical variables to predict severity of COVID-19 patients

Nature Communications 27th January 2021
View Publication

Using StyleGAN for Visual Interpretability of Deep Learning Models on Medical Images, NeurIPS 2020

Medical Imaging meets NeurIPS 10th December 2020
View Publication

Fully automatic segmentation of diffuse large B cell lymphoma lesions on 3D FDG-PET/CT for total metabolic tumour volume prediction using a convolutional neural network

European Journal of Nuclear Medicine and Molecular Imaging 15th October 2020
View Publication
View all publications