The distinguished professor of Mathematics at the engineering school of the University of Paris-Saclay and founder, president and CEO of TheraPanacea, Nikos Paragios, describes in “K” the possibilities of artificial intelligence in Medicine
For several years now, but even more intensely in the last decade, researchers around the world have been looking at how artificial intelligence can improve Public Health.
Hundreds of funded programs and many interesting studies showed the “power” of the algorithm over medical methods.
But today anyone still looking for ways to improve public health through artificial intelligence is left far behind, as there are already researchers around the world who clearly know the dynamics of algorithms in medicine and are already applying them.
For example, Hungary has a robust breast cancer screening program and one of the largest testing sites for the technology in real patients. More than 35,000 X-rays are performed annually in five hospitals and clinics in the country, as artificial intelligence systems were introduced from 2021 and now help screen for signs of cancer that may have been “missed” by the radiologist.
On the other hand, the New York Department of Health approved a new digital pathological diagnostic test for breast cancer based on artificial intelligence. The algorithms will be able to accurately ‘classify’ invasive breast cancer into early, low and high-risk stages, as well as calculate the likelihood of recurrence within six years of primary diagnosis.
It is therefore clear that artificial intelligence can not only “infiltrate” doctors’ offices and hospitals worldwide, as it already does, but it can completely change the way diseases are diagnosed. In the future, an algorithm will be able to show which patient will be able to undergo a specific treatment and what success rate it will have. The image of hospitals is expected to change drastically in five years, but with it the responsibilities of doctors.
The distinguished professor of Mathematics at the engineering school of the University of Paris-Saclay and founder, president and CEO of TheraPanacea (a European research center), Nikos Paragios, describes in “K” the possibilities of artificial intelligence in Medicine and gives us a promising and optimistic picture from the future especially regarding the diagnosis and treatment of cancer.
“What we can say with certainty now is that artificial intelligence is useful in three different areas in Medicine. The first area is the diagnosis, the second is the treatment and the third is the prediction of the response that a patient will have to a treatment”, points out Mr. Paragios.
Artificial intelligence useful tool in diagnosis and treatment
“In the field of breast cancer diagnosis, AI can process millions of data from past mammograms over the past ten years and create algorithms and logic that replicate best practices in the field. The goal is therefore through artificial intelligence to have a much safer and more accurate diagnosis of the patient, which could also be used in hospitals where the necessary expertise is not available”, says the professor of the University of Paris-Saclay.
The second area in which artificial intelligence can help is that of therapy. Let’s take cancer as an example, for the treatment of which there are currently four treatment options. The first is to have surgery, the second to undergo radiation therapy, the third to chemotherapy and the fourth to immunotherapy.
“With the help of artificial intelligence, we can better target cancer cells and avoid healthy ones, so that we have the least destruction of the healthy cells that are around the cancer. To date, radiation therapy to patients is done manually by the doctor through CT scan processing. But now the algorithms can tell us more precisely where the robot should throw the beam exactly”, says Mr. Paragios and adds that “one of the applications we work on at TheraPanacea does exactly that and today it is installed in over 100 hospitals in the world (including “Alexandra” hospital in Greece).
The third area is a bit more distant but at the same time the most interesting, as it relates to the prognosis on immunotherapy – or any other treatment.
An algorithm will tell us the success rate for cancer treatments
“Today the problem with immunotherapy is that on the one hand it is very expensive (€100,000 per patient) and on the other hand the success rate reaches about 20%. Imagine in the near future an algorithm being able to tell us which patient will be able to undergo the treatment and if it will be successful. In this way, the cost of the treatment would be reduced and we would know which patients would have a response, while at the same time we would direct the rest to alternative treatments, saving significant time – and in the fight against cancer, the time factor is very important”, says the scientist.
From the above, it is understood that the innovations that artificial intelligence will bring to health systems will affect both the diagnosis and treatment of cancer, as well as the role that the doctor will have in the future.
The role of the doctor will be boiled down to the very basics
“Each AI computing solution can process a lot of data from different clinics. A doctor will see more limited medical phenomena in his life. So we are led to the conclusion that a computational solution using artificial intelligence will have much more data than a doctor and even more interestingly it could have data and learn from the best practices and the best university centers. That is why I believe that the role of the doctor will be channeled in the coming years to the very difficult or rare cases in terms of diagnosis. Essentially, artificial intelligence in five years will reach the level of an average doctor”, says Mr. Paragios and points out that currently in the US 500 applications have been certified in artificial intelligence.
As for the treatment part, in the future we will have many micro-robots that with the help of doctors and data will give more precise targeting and delivery of treatment.
However, as Mr. Paragios points out, although in America there is a significant amount of investment in this sector, in Europe the procedures are more complex.
“In America things are much easier. If we have an application there that does its job, it is very easy to “come out” from the laboratory to the market. In Europe, each country has its peculiarities. So there are obstacles to the rapid spread of the models in the EU. Despite the “general umbrella”, this heterogeneity in Europe puts a brake on the circulation of such applications”, he points out.
Artificial intelligence in Greek hospitals in the horizon of a decade
As for Greece, the use of these applications would be particularly important. Due to the idiosyncratic distribution of the population, as well as the geography and the large number of regional hospitals, AI applications can offer specialized diagnoses across the territory.
“University hospitals could host such applications in the future, so that they can then spread to the rest.
In fact, consider that in the future, through artificial intelligence, a patient could have a guaranteed and safe diagnosis, whether he was in the hospital of rural Karpathos, where I come from, or in Evangelismos,” said Mr. Paragios.
However, as he points out, the implementation of such solutions in our country is extremely difficult, as it requires a financial model that is currently difficult to make available, when there are much more important problems regarding the financing of the health system.
“I believe that in a decade’s time, Greece will be ready to welcome such applications and the doors of hospitals will be opened for artificial intelligence applications. However, I believe that this will be done much earlier in the hospitals of France, Germany, Switzerland, the northern European and Baltic countries”, he concludes.
Source kathimerini.gr