AI in Medicine Gets Personal
A DMEXCO column by Thilo Kölzer, COO of DocCheck AG, on AI in Medicine, Artificial Patients, and Digital Twins.
AI in Medicine: Artificial Patients and Digital Twins
Medicine without big data in combination with artificial intelligence will no longer exist in the future. And that is a good thing!
The German Association of Research-based Pharmaceutical Companies (VFA) sees AI in medicine as an opportunity to better understand diseases and develop ever more effective therapies. Nowadays, it takes around 13 years from the idea to the approval of a drug—an unthinkable amount of time in the age of AI! This time span will be significantly reduced with the help of AI in medicine in the future. And from more precise diagnosis to personalized treatment plans, doctors and patients will also benefit from AI-based technologies in the subsequent stages of medical treatment.
AI in Medicine Speeds Up Progress
The production and marketing of medicinal products involves the management of information at every stage, as 8,000 to 10,000 new scientific publications are published every day. This large amount of data, for example in the form of application studies, must of course first be processed. Finding the relevant data snippets and interpreting them is the central challenge here. Finally, the relevant findings must be recognized and didactically reduced in such a way that third parties can understand, comprehend, and draw conclusions from them.
A lot of work for humans—but the perfect area of application for an AI in medicine. In short: artificial intelligence forms the basis for the healthcare solutions of the future.
Data visualization with AI Patients and Digital Twins
What is often overlooked in the field of science, millions of data points and extensive evaluations is that this data is generated by real people—namely patients!
So what could be more obvious than visualizing it through people? And this is also where AI comes into the “health game”.
The generated patient data is real and reflects medical reality, but the people concerned are not always available for visualization.
So how about using AI to create and visualize artificial humans—or rather “human models”—that are based on real data and thus reflect reality 1:1?
POV: A rare disease called XYZ affects only 5,000 people in Germany. Only 40 percent of this group have been identified so far. The symptoms of this rare disease are: a blank look, weight loss, certain changes in facial shape and proportions.
In order to make other people, for example members of the risk group, aware of this disease, it would make sense to make those already affected be seen and heard and make the changes in people clearly visible.
The catch is that due to the small number of people affected, it will be difficult or even impossible to find someone who has the symptoms AND who is willing to speak out and stand in front of a camera. In addition, with many diseases it is useful to look at them over time, to be able to recognize developments or to assess whether a certain treatment works in a certain period of time and alleviates the symptoms or the course of the disease.
All of this is almost impossible to achieve with humans—and the logistical effort involved in observing a person in everyday life over several years or even decades is beyond any budget.
How can AI in Medicine help here?
Tens of thousands of data points relating to disease XYZ are available from the real patients identified—in other words big data. Based on these real, authentic data points, AI can now be used to develop images and videos of people who look like those affected by disease XYZ, but who do not exist or are not available in reality.
AI Creates Visibility for Rare Diseases in Medicine
It is important for patients to see and recognize whether they belong to the group of people at risk or whether they could already have the disease. AI patient images thus help to create visibility and awareness for rare diseases and, as a result, improve the chances of early detection.
There are other areas of application in which generative AI is useful in medicine, for example in dermatology.
AI in Dermatology: Better Visualization of Disease Progression
Visible changes to the skin can be visualized very well using AI. Treatment success after two, four or six weeks can also be demonstrated. Rosacea and acne for example are two skin conditions that can last for many years, have different degrees of severity and can develop in completely different directions.
With the help of AI, both diseases can be simulated very well on virtual skin—changes in the skin and appearance, even if they are only nuances, can be represented realistically.
AI in Medicine for Training and Patient Education
Such representations are not only suitable for patient education, but also for training specialist staff, e.g. in pharmacies.
Image-generative AI thus enables available data and information to be efficiently converted into authentic disease representations in various medical fields.
AI-visualized patient and disease images will thus help to improve the treatment of those affected. The future of healthcare is being shaped by AI in medicine—from faster diagnoses and personalized treatments to innovative visualization tools that support both patients and medical professionals.
Recommended Reading: The Future of Medicine is Personalized
Personalization in medicine goes beyond treatments—it also applies to communication! In his DMEXCO column, Thilo Kölzer, COO of DocCheck AG, explains why tailored approaches—such as CAR-T cell therapy—require targeted messaging.
