By integrating high-resolution biological, clinical and medical imaging data, AI makes it possible to identify in detail the particularities of patients, and to provide them with personalized treatment.
Between the patient’s disease understood in its smallest particularities and the optimal medicine predicted among millions of possibilities, artificial intelligence (AI) is changing the face of medicine and pharmacy, to the benefit of the patient.
The Covid-19 pandemic has provided a striking example. In the first months of the epidemic, when the world was confined, two biotechs unknown to the public, Moderna and BioNTech, announced to everyone’s surprise the development of two remarkably effective vaccines… in barely one year, compared to more than ten usually! We are therefore witnessing an unprecedented synergy between biotechnologies, pharmaceutical sciences and AI, a discipline which reproduces via machines four important dimensions of human intelligence – perception, analysis, action and learning.
“Digital Twins”
AI represents for researchers a state-of-the-art assistant capable of analyzing a multitude of data in record time better than any human brain. While it still takes twelve years of R&D and 2.5 billion euros on average to develop a new drug, AI could make it possible to go fifteen times faster, and reduce development costs by 30%. and bringing a drug to market while reducing the risk of failure – which is currently around 93%. In 2020, it took less than a year to develop several drugs for obsessive compulsive disorder (OCD) or certain cancers and to start testing them in humans. Indeed, instead of following a complex process that requires testing a panoply of molecules for each therapeutic target, often randomly, in animals and then in humans, AI considerably increases the productivity of experiments.
” READ ALSO – What ethics for the use of artificial intelligence in health?
After integrating and analyzing hundreds of billions of parameters, AI uses its artificial neural network to design and optimize drug candidates that interact with these therapeutic targets. It also makes it possible to reposition old drugs in new indications, as was the case during the Covid pandemic with anti-inflammatories in particular. Baricitinib, initially used against rheumatoid arthritis, has thus made it possible to reduce the mortality of hospitalized patients by 38% by acting on the pulmonary inflammation induced by Sars-CoV-2. Another example of the power of AI: the supercomputer at Oak Ridge National Laboratory was able to model the molecular structure of 2 billion molecules in one day, instead of three months previously for a single molecule… It took only forty-six days for the start-up Insilico Medicine thanks to GENTLR, its artificial intelligence, to develop a drug against renal fibrosis, at the price of only 150,000 dollars.
Leaving the bench for the computer, researchers are also beginning to test the efficacy and safety of drug candidates on virtual patient models, called “digital twins”. We thus create virtual human organs (heart, brain, and soon lung and liver), just as we can model the effect of a placebo in clinical studies. A digital twin of the human heart, jointly developed by the American drug agency (FDA) and Dassault Systèmes, has made it possible to predict the effectiveness of pacemakers and stents. AI is also needed to design and evaluate in silico (by computer) totally innovative drugs, with the ultimate prospect of avoiding animal testing and reducing clinical studies in humans. By integrating high-resolution biological, clinical and medical imaging data, AI reports on the heterogeneity of diseases, to identify in detail the particularities of patients and provide them with personalized treatment. Known as “computational”, the medicine of the future will even make it possible to produce a digital duplicate of each patient that the doctor will learn to know better and treat better, even from a distance.
” READ ALSO – How to regulate artificial intelligence in medicine?
AI also optimizes the entire drug chain, from the factory to the patient. Sensors monitor the quality of drug manufacturing, packaging and distribution, and visualization tools help with supply chain management. At the time of dispensing, AI helps automatically identify patients at risk, and detect drug incompatibilities in prescriptions. In the United States, smart tablets, such as the Helios pill from Proteus, already incorporate a chip providing information on their nature and their dosage to facilitate compliance by the elderly.
Thanks to digital transformation, hospitals can better exploit health data and automate tasks, which improves the patient care journey, favoring personalized, non-invasive and ambulatory treatments. Finally, once the drug is on the market, AI makes it possible to optimize pharmacovigilance, i.e. to improve the monitoring of drugs in terms of efficacy and safety at the population level. .
AI-augmented medicine is therefore no longer science fiction, and truly personalized precision medicine is on the way to becoming reality.
Thematic session “Smart medicine” at the Academy of Pharmacy on February 1, 2023. www.acadpharm.org