Rare diseases: what applications and prospects for AI? – news

Our events

A patient suffering from a rare disease experiences an average diagnostic delay of four years. Faced with this unacceptable observation, participants in the second PharmaHealthTech round table have developed tools that use artificial intelligence to reduce this wandering and improve the journey of the patients affected.

95% of the 7,000 to 8,000 listed rare diseases currently have no suitable treatment. “As an industrialist, our goal is to provide innovative solutions to patients,” says Pierre-Olivier Boyer, CEO of PTC Therapeutics France, speaking at the fourth edition of PharmaHealthTech, an event co-organized by Pharmaceuticals and TechToMed. We need to speed up and artificial intelligence will contribute to this. »
If France is according to him “one of the best” to allow rapid access to innovations for patients thanks to early and compassionate access systems, it is fishing on sustainability, for lack of knowing how to price these treatments. “What is the right way to get the right remuneration? AI will help us to have the best possible evaluation of these innovations,” he hopes. It is still necessary to know to whom to prescribe them. The diagnostic wandering – four years on average in France – is a real scourge for patients suffering from a rare disease.

Raising awareness among health professionals

First lever: “develop the culture of doubt” among primary care physicians faced with a complex clinical picture. Moderator of the round table, Jérôme Leleu, CEO of SimforHealth, took the opportunity to announce the official launch today of the digital simulation platform RareSim, the objective of which is precisely to raise awareness among healthcare professionals about rare diseases. Initiated in response to a call for projects launched by Coalition Next, RareSim has benefited from the institutional support of Ipsen, Novartis, Pfizer and Takeda.
Also with a view to reducing this diagnostic error, Sanofi has chosen to join forces with the French start-up Medical Intelligence Service (MIS). “AccelRare® is intended to be a digital companion to help local physicians consider a hypothesis of a rare disease, by providing them with a reliable pre-diagnosis”, explains Etienne Van Der Elst, digital product manager – Digital Innovation at Sanofi. Based on MIS’ MedVir™ medical decision support device, “which has already proven its reliability”, the new tool focuses on 270 rare diseases for which treatment or treatment already exists. appropriate support. In addition to a pre-diagnosis in the event of suspicion of a rare disease, the software must also provide the address of the nearest expert centre.
AI is not only used to design tools but also to make them known. For Virginie Druenne, who created the “RARE à l’enseignement” podcast (6,000 subscribers, more than 180 episodes available), the dissemination of information is essential. It exploits all the resources of search engine algorithms so that this content appears in a good place when certain keywords are entered. “Patients shouldn’t have to look for information, it should come directly to them,” she says.

Use of real life data

Certain tools not specifically developed for rare diseases could also prove very profitable in this field. This is the case, for example, of the two new software programs that the Quinten company intends to deploy, thanks to the €14 million collected during its very recent fundraising. The first, aimed at manufacturers, will be based on real-life data to model diseases. The second will be aimed at hospital pharmacists with the aim of securing prescriptions, and thus reducing iatrogenics.
Quinten’s ambition is above all to exploit the full potential of real-life data. According to its president and co-founder Alexandre Templier, the generation of evidence is now totally fragmented, with everyone working independently on their own data to design a specific algorithm for a given disease. “We must succeed in integrating these different models to reduce the time and cost of producing this evidence”, he pleads.
However, it is in a very specific therapeutic area, kidney transplantation, that Cibiltech has developed its software for predicting the risk of transplant rejection. “Some patients concerned have a genetic disease,” recalls Stéphane Tholander, co-founder and CEO of the company. Thanks to the organization of care pathways, according to him, it is easier in France than in other countries to access highly structured data, “thanks to which we can produce high-level algorithms. The result of the work of an Inserm-AP-HP team, this solution “developed on 4,000 patients and since validated on more than 10,000 patients on an international scale”, is already deployed in many hospitals.

Julie Wierzbicki

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *