2024-04-23 06:45:18
Although one in eight women will develop breast cancer in her lifetime, this disease is well treated and can be cured in nine out of ten cases. But “I am an angry woman because despite the effectiveness of breast cancer screening and treatment, 33 women die from this disease every day in France and almost 1,800 worldwide,” explains co-founder Hakima Berdouz , with two other scientists, from Dóchas Valley AI, which has just been created in La Rochelle after a year of maturation within the Inria start-up Studio incubator.
To reassure
At the next VivaTech show, the very young start-up will present the first version of a smartphone application that will facilitate the early detection of faint signs of breast abnormalities, precursors to cancer. “Studies have shown that a significant number of patients who report breast cancer have previously had very mild and invisible symptoms. This application will therefore be intended to raise awareness and guide women, thanks to conversational AI, to the first clinical self-examination,” explains Hakima Berdouz. The application will be officially launched during the next “Pink October” breast cancer awareness campaign.
The company is already thinking bigger and its smartphone application must be the first brick to develop “ultra-personalized preventive medicine”. The second will be the “mammography system of the future”. Unlike current devices, the “Mammope” in the X-ray preparation will only use a multimodal imaging system (ultrasound, infrared thermography, computer vision) and will use artificial intelligence.
Hakima Berdouz, who initiated the project, does not come from health, but from the nuclear world. This engineer-researcher, a specialist in mathematical risk engineering, worked for almost 25 years at the CEA to predict the problems that existed within nuclear power plants, especially by relying on AI. It is these skills and technologies that she is now applying in health matters. In 2020, the CEA is asking its researchers to propose ways to combat the Covid pandemic, and more broadly in favor of health.
Collaboration with MIT
Hakima Berdouz first applied her work to a rare autoimmune disease, systemic lupus. The project brings together the AP-HP, the Sorbonne Nouvelle University and the National Center for Autoimmune Diseases for two years. “I checked it could save people. It was a revelation,” she recalls. Since 2022 she has been working on kidney cancer which has a genetic component, unlike systemic lupus. She runs her algorithm on an American database.
“We managed to predict that an aggressive type of cancer would occur at two years. And this from expressing about ten genes when previous methods required the analysis of hundreds,” she explains. The researcher then turns into an entrepreneur by focusing on breast cancer but choosing a different strategy. While most startups bet on AI to detect this disease using existing imaging systems, Hope Valley AI is betting on a new device.
Leaving patients
“Our approach is a breakthrough because we started from the best of science and also from user feedback. They often forget about screening and mammography because they underestimate its value and fear an unpleasant and anxious examination, with technology that is more ionizing and likely to cause significant radio-induced breast cancer, even if they are in the minority, argues Hakima Berdouz .
The device is developed in partnership with Aurys Industrie and the Doliam group. The first 24-month clinical study will be launched at Tours University Hospital.
Hope Valley AI, so far backed by business angels, will have to fund its development. She hopes to complete the first fundraising of 2 million euros this year. The start-up benefits from an exclusive license for all health-related applications of the six patents filed by the CEA and of which Hakima Berdouz is the lead author. Hope Valley AI will also look towards the United States and should sign an R&D collaboration agreement with MIT by the end of the year.
#Mammography #reinvented