PREDILIFE STRENGTHENS ITS ASSETS WITH THE ISSUANCE OF A PATENT FOR ITS AI METHOD

 PRESS RELEASE MAY 29, 2024 .

PREDILIFE STRENGTHENS ITS ASSETS WITH THE ISSUANCE OF A PATENT FOR ITS AI METHOD

Villejuif, France - May 29, 2024 at 6:30 p.m. CET - PREDILIFE (Euronext Growth: ALPRE FR0010169920), a specialist in artificial intelligence methods for predicting disease risks for personalized medicine, is pleased to announce the issuance of a European patent by the European Patent Office for its risk prediction method for breast cancer used in the MammoRisk^® test.

The breast cancer risk prediction method used in the MammoRisk^® test is the result of many years of research conducted by Predilife in collaboration with Sorbonne University and the Institut Telecom. The data used involved nearly 1,700,000 people in France and the United States. To protect this work, a patent application was filed and this patent has just been issued by the European Patent Office under number EP3164824.

This Artificial Intelligence method, and more specifically supervised learning, falls within the field of AI and Machine Learning. Very strong growth of over 30% is expected on these subjects in the coming years1.

The patent proposes an innovative method consisting in selecting the individuals closest in the multidimensional space of risk factors (age, personal history, parameters from imaging, genetic data ...) and then observing their outcome over a given period. For example, by taking the 1,000 closest neighboring profiles and observing 30 cancers 10 years later, the risk will be estimated at 3%. This method applies to both quantitative and qualitative data such as words. The patent integrates data such as ethnicity or breast density into risk factors with a method for quantifying this type of factor. This process can therefore be easily combined with generative AI methods using language models by providing quantitative answers to text-based questions such as: "what happens if I stop smoking or have too much tension?". The applications are therefore very promising.

This method was optimized on French data and then validated on 1,000,000 American women from the BCSC (Breast Cancer Surveillance Consortium). The predictive performances have proven to be excellent, as they avoid any overestimation or underestimation in subgroups unlike more classic methods (of parametric statistics), while maintaining great flexibility (compared to those of data mining).

This method has been validated for breast cancer and the patent claims have been accepted for all diseases.

In addition to its general applications, this patent strengthens the protection of the MammoRisk® software, in addition to copyright. This patent constitutes a valuable asset as this model has been selected by the European Union within the framework of the MyPeBS clinical trial which aims to modify the screening of the 100 million European women over 40 by measuring the risk of breast cancer of each woman.

 1. https://www.linkedin.com/pulse/artificial-intelligence-machine-learning-market-size-u7skf/

This asset, limited to this application, was recently valued by Sorgem Évaluation at €207 million.

“The granting of this patent gives us additional protection in predictive medicine. This proprietary AI method can be applied to all diseases and we will decline it on different databases. This work from French research centers has received international validation by its application on American data and its selection by the European Union" comments Stéphane Ragusa, Chairman and CEO of Predilife.

About PREDILIFE

PREDILIFE is a pioneer in the design and development of prediction tests to prevent major diseases. It uses artificial intelligence methods applied to clinical and genetic medical data in a secure legal framework. The prediction of these individual risks makes it possible to propose a personalized follow-up protocol and earlier identification of pathologies.

To learn more: https://www.predilife.com/en/