A Symposium on Artificial Intelligence organized by Fleni delved into the main findings and their application in the healthcare field. How it strengthens the doctor-patient relationship and the experts' conclusions.
To reflect on the scope of this new tool, the Artificial Intelligence/Human Intelligence Symposium was held on April 25th, where renowned experts from various universities and institutions in Argentina, Chile, and the United States discussed the most important findings regarding the application of AI in the healthcare field before an audience of healthcare professionals.
Fleni, an institution dedicated to the care, prevention, diagnosis, treatment, and rehabilitation of neurological diseases, together with Entelai, the leading healthcare AI development company in Latin America, held the symposium where current AI projects and applications in healthcare were discussed.
In an exclusive interview, Dr. Gustavo Sevlever, Director of Research and Teaching at Fleni, and Dr. LucÃa Crivelli, head of the Neuropsychology Department at the same institution, spoke, about the implications of this technological advancement in the field of healthcare and the doctor-patient relationship.
How AI Has Changed
Medicine
In the context of an increasingly aging population and an increase in chronic diseases, AI is presented as a solution to many of today's problems, providing tools that reduce data and image processing times, facilitate diagnoses, and facilitate communication between different healthcare centers, among many other functions.
In Argentina, artificial intelligence (AI) has been used in daily medical practice for several years, especially in neurological diseases, as a diagnostic tool. Furthermore, the two most widely used applications are those related to medical imaging assistance and patient care through virtual assistants.
At the opening of the symposium, Dr. Gustavo Sevlever described that so-called artificial intelligence consists of hardware and software-based systems that help manage large amounts of data. In the healthcare field, it contributes to understanding images and medical histories, among other aspects. He also emphasized the role of the physician in this context and stressed that the doctor-patient relationship is irreplaceable and fundamental to both diagnosis and treatment.
For her part, Dr. LucÃa Crivelli, head of Fleni's Neuropsychology Service, proposed a reflection on what truly makes us human in a world increasingly influenced by artificial intelligence, highlighting creativity, fragility, and consciousness as inherent human dimensions.
Another relevant figure participating in the meeting was Dr. Guido Falcone, associate professor of Neurology at Yale University School of Medicine, who researches and applies AI to the care of neurological patients.
The expert explained that artificial intelligence is improving healthcare in many ways. “One very concrete way is that it is increasing the speed and accuracy with which we interpret neuroimages. This has to do with neurology in particular and truly adds a layer, like a co-pilot to the general practitioner or neurologist in this case.”
Both Dr. Crivelli and Dr. Diego Fernández Slezak, creator and director of the Applied Artificial Intelligence Laboratory at the University of Buenos Aires (UBA) and co-founder of Entelai, who also spoke at the symposium, reflected on human and artificial intelligence, drawing parallels between the two and how they can learn from each other.
What is artificial
intelligence
Dr. Sevlever asked himself what artificial intelligence is: “Is it a resource, a colleague, a companion, a therapy, or a diagnosis? It's all of them together. A bit of each. And it has an impact on society, because since the advent of the Internet, patients have had—what we joke about—“Doctor Google.” So today we have a different way of doing medicine than 30 or 40 years ago, and this reinforces that trend a bit,” he emphasized.
Dr. Crivelli also pointed out that artificial intelligence serves as “an assistant that makes tasks easier for us; that provides us with relevant information at the time of the consultation; that can assist with the diagnosis because there are very useful diagnostic applications, but at some point, the intervention of the person, the medical judgment, is always necessary.”
He continued: “We were talking today during the Symposium about the nonverbal components of a medical consultation or human interaction. The gaze, what one can hear, the tone of voice. All of this is part of a diagnostic process or a choice of treatment or an interaction between two people that chat or artificial intelligence tools still can't resolve or provide the information for. So at that point, it's as if artificial intelligence could replace some tasks, but it can't do so with people or roles. It does complement and assist them.”
The expert emphasized that if we take advantage of its advantages, "we can facilitate our task and focus on what's truly important. Today, when we talk about human intelligence, the characteristic of human beings has to do with contact with others, with our gaze, with skilled and sensitive listening. All of these things, if we have a data processor as wonderful as this one that can accelerate us and provide responses based on information more quickly, we can do them with greater quality and in more time," she emphasized.
Dr. Svelever also noted: “Since the emergence of Big Data, we have the ability to store vast amounts of data on diseases, epidemiology, genetics, and images that we cannot process with our cognitive capacity. And that's where we need this auxiliary, the Spanish word for computers, the ordenador. It's the computer on a massive scale. We must learn to use it to take full advantage of its capabilities and to preserve human roles. All this artificial intelligence has a point that revalues, redefines, and enhances what is specifically human, in the relationship with the patient, with those who suffer, with those who are ill. This has added value,” the expert emphasized.
The advances that AI currently brings
Image analysis. Dr. Svelever explained that with AI, diagnosis and all data are enhanced because all the information is at your fingertips. “For example, a product called Entelai for image analysis quantifies lesions, shows the lesion load, provides a description, and assists the radiologist. It doesn't replace them, but it helps better define what they're seeing and its implications.”
Chatbot for patient care. “We also have a chatbot from Entelai that frees the system from some tedious and complicated functions. It assigns appointments, interacts with the patient, asks a brief history, questions, determines if they have a headache, hypertension, and performs what's called triage. It tries to identify if they should come to the emergency room. The chatbot does all of that and is under development, meaning it continues to learn,” said Svelever.
Stroke software. “We have a commercial package used in stroke. When a person suffers a stroke, they have a window of opportunity for the doctor to insert a catheter, break up the clot, and restore blood flow. That window of opportunity is statistical; you know it only takes so many hours. Now there's software that analyzes the CT scan and tells you, 'No, in this patient, you can extend that window: you have 4 hours, or 5, or 8.' This allows you to intervene and decide if things are going well, that they don't deteriorate. That's artificial intelligence software,” the expert explained.
Dr. Falcone considered that “AI in stroke is helping to interpret data and make decisions in a timely manner, in a much more efficient manner. It's not the same to wake up three doctors at 3:00 a.m. than to have a single doctor directly receive information that's already preprocessed by artificial intelligence.”
The Spoke tool is presented as a “copilot” for the medical professional. Dr. Sergio Baranzini, Distinguished Professor of Neurology at the University of California, San Francisco (UCSF), explained that Spoke is a system that integrates dozens of specialized databases on different topics, ranging from genetics, immunology, clinical medicine, and different areas of biomedicine, and connects them in a graph, a database with millions of concepts linked by semantic relationships that can be interpreted by computers and humans, assisting in the interpretation and retrieval of information.
He added: "In reality, it's more of an autopilot than a copilot. Today's airplane autopilot can take off, navigate, and land without pilot intervention if desired, but there are always two pilots in the cockpit at all times. I believe that these tools being designed now will help physicians make faster, more accurate, and informed decisions, but they won't replace them." My prediction isn't that artificial intelligence will replace humans, it will probably replace humans who don't use it," he remarked.
A project to detect cognitive decline early
Dr. Crivelli explained that she is working on a joint research project with Diego Fernández Slezak, in which "they evaluate patients and controls with mild cognitive impairment and controls without the disease, and we have them perform several tasks while we film and record them. So, with a natural language analysis algorithm called Natural Language Processing and with software called Facemesh, which interprets facial gestures, we try to detect cognitive impairment early in patients without having to perform a cognitive assessment," the doctor described.
This would allow not only to advance the time of diagnosis, which is very important now for the application of new treatments, the doctor said, but also "to democratize it, because the idea is to later bring this software that we use on computers to small screens and be able, through the use of a cell phone, to record the voice, film the patient, and be able to perform triage, early screening, and detect possible cases of cognitive impairment earlier than we are doing now."
Dr. Crivelli noted that what they're doing is a multimodal analysis of the data: combining voice with gestures. "When you memorize something, when you try to recall something, you actually move your eyes in a special way. When you go back in time. So the hypothesis is that patients with cognitive impairment who will develop cognitive impairment will make these eye movements and gestures differently than people who won't develop them. So, with language and gestures, we can identify high-risk cases early," the expert described.
What is the importance of a timely diagnosis? The doctor responded: “Knowing early that we have this risk is essential for two reasons: because I can work on the risk factors. We know that if we work on cognitive stimulation, cardiovascular risk factors, nutrition, physical exercise, and socialization, we can prevent a large proportion of dementia cases, make them appear later, or prevent the person from dying from something else without them appearing.”
She added: “In the famous Finnish Finger study, using these five factors, cases were reduced by up to 30%. People at risk of dementia did not develop it when they would have if they were not treated differently.”
And the doctor added: “The other reason why prevention is important is because of the new treatments available for Alzheimer's disease, so we can triage and see which people I might eventually need to evaluate and treat with these new treatments when they arrive in Argentina.”
Finally, Dr. Sevlever concluded: “There's a tendency toward magical thinking when using technology, but just as we learned to drive cars, ride bicycles, and use computers, we have to learn to use AI. It's a bit of a demystification. That said, I think it has a usefulness that we're just beginning to see in its development. And we, as people who work with the brain, with patients, with people who are suffering and ill, always have to value empathy. We say: “Cure when you can, relieve if you can't cure, and always accompany.” And that is a task, as I said, irreducibly human and non-delegable. And it's our responsibility.”
No comments:
Post a Comment