In medical research they are actively introduced digital doubles created using artificial intelligence. The technology allows you to test medical devices, such as stents and valves prostheses, on virtual models before conducting tests in public.
ADSILICO has developed a system of digital doubles of the heart that can simulate various biological characteristics, including age, gender, pressure, as well as health and ethnicity. This helps to conduct tests on a more diverse virtual patients than is possible in traditional clinical studies. The use of AI allows you to take into account the anatomical features and physiological reactions, which makes testing more accurate and safe.
The effectiveness of such approaches becomes obvious against the background of data on the safety of medical devices: from 2008 to 2018, more than 83 thousand deaths and 1.7 million injuries associated with their use were recorded. Careful testing with digital doubles can significantly reduce these indicators, as well as reduce the number of trials.
The testing process in Adsilico includes the creation of a digital copy of the device, which is then tested in the virtual model of the heart. This allows you to carry out thousands of simulations, modeling different conditions – from high pressure to various stages of diseases. This approach is cheaper than traditional clinical trials, and opens more opportunities for analysis.
Similar technologies are used in the development of drugs. For example, the pharmaceutical company Sanofi uses digital doubles to reduce the test time by 20%, which can lead to significant savings. AI systems create virtual patients who participate in control and placebo groups, and also model as the drug is absorbed by the body. This helps predict patient reactions and increase research accuracy.
Nevertheless, the development of digital doubles is faced with problems related to the quality of the source data. The insufficient representation of marginalized groups in medical data can lead to the preservation of some prejudices in the simulations. Companies are trying to solve this problem using data from external sources, such as biobanks and electronic medical records.
The technology of digital doubles has the potential to replace testing on animals, which is still used in the medical industry. The virtual model of the heart is already closer to the real human organ than the heart of the animal, which makes it more promising for research.