Home / Health / The AI ​​could be the key to catching type 1 diabetes much sooner

The AI ​​could be the key to catching type 1 diabetes much sooner



For the millions of people with type 1 diabetes in the world, everyday reality involves significant self-monitoring. Without this monitoring, the pancreas does not produce enough insulin, which is used to transfer blood sugar that provides energy to the body's cells. Whether it is insulin injections daily or nutrition and exercise programs, it is essential that patients remain very vigilant about their health.

In the United States alone, about 1.25 million people have type 1 diabetes, with about 40,000 new diagnoses each year, according to the American Diabetes Association. In these circumstances, you may be surprised to find that there is no standardized screening process for the disease to detect it quickly. Doctors usually test based on family history and other known risk factors, which means that type 1 diabetes may appear under the radar. This can lead to sudden emergency visits and unexpected diagnoses, making the development of better screening tests a major priority for physicians.

Doctor or nutritionist explaining the glycemic index of a diabetic patient

Now, here is where AI comes in. At the 79th Scientific Sessions of the American Diabetes Association in early June, IBM and JDRF (formerly known as the Juvenile Diabetes Research Foundation), a non-profit organization that leads type 1 diabetes research, have unveiled an AI predictive tool mapped the presence of type 1 diabetes antibodies in the blood to determine exactly when and how the course of the disease could occur. Jianying Hu, fellow scientist and world-leading AI scientist for health care at IBM Research, told Engadget that AI had powered data from more than 22,000 people from the United States, from Sweden and Finland.

The program highlighted similarities between people with specific antibodies for the disease and the chronology of progression of their type 1 diabetes.

"One of the biggest potentials of this type of work in building type 1 diabetes learning models is to be able to better identify who to monitor and how often to monitor them," Hu said. team worked on this project with JDRF for over a year. "At the present time, even the little we know, these antibodies are ubiquitous in the progression of type 1 diabetes, but no one knows who is most likely to develop them and when."

She said that these models of AI could give doctors "a more personalized time" for how to monitor people and how often they should be tested.

In the past, type 1 diabetes was called juvenile diabetes because it is usually diagnosed in children, adolescents, and young adults, according to Centers for Disease Control and Prevention (CDC). However, it can affect people of all ages, said Utpal Pajvani, an endocrinologist and assistant professor at the Columbia University Medical Center.

Pajvani, who is not affiliated with this project, explained that the general practice is to filter only those with "high risk". This means that they have a first-degree family member who has been diagnosed. Since it is so rare, he said that it was not something that would justify a screening of the general population.

He cautioned that many screening methods like this could lead to many false positives.

"If you're looking for a relatively uncommon disease, you're going to end up with a lot of false positives, if you test everyone, [including] those who are [at a] relatively low risk of developing it if there is no family history or other autoimmune disease, you will have a much higher rate of identification of people likely to undergo a test positive antibody screening but [at] low risk of having the disease, "Pajvani told Engadget.

Blood sample for the diabetes test "data-caption =" Blood sample for the diabetes test in a blood tube on the glucose control card. "Data-credit =" Gam1983 via Getty Images "data-credit-link-back =" "data-dam-provider =" Creative Getty "data-local-id =" local-5-5238662-1560782155355 "data-media- id = "cc5b7ed9-0be6-365e-ad14-dab9c8a4f757" data-original-url = "https: / /s.simy.com/os/creatr-images/2019-06/fe588540-910c-11e9-b77f-8bb561e2a1f9" data-title = "Blood Sample for Diabetes Screening" src = "https: //o.aolcdn. com / images / dims? crop = 4928% 2C3169% 2C0% 2C54 & quality = 85 & format = jpg & resize = 1600% 2C1029 & image_uri = https% 3A% 2F% 2Fs.yimg.com% 2Fos% 2FFreatr-images% 2F2019-06% 2Ffe588540- 910- 11e9-b77f-8bb561e2a1f9 & customer = a1acac3e1b3290917d92 & signature = a8fb55df99ff92be4ad8e834f8a6e400232634d2 "/></p><p>The risk of applying a general screening test for a rare condition like this means that you can also needlessly worry people about an illness that they are unlikely to contract. Essentially, no perfect screening test exists without false positives, he added.</p><p>Despite these criticisms, Pajvani sees a future for this type of technology. People who are at high risk of diabetes and who have the presence of these antibodies still do not know what will be the chronology of the course of the disease. This type of AI tool could give doctors a road map necessary to monitor the evolution of the disease, he explained.</p><p>Hu Jintao said his team would soon add more data from Germany, which would be sorted by the AI. She added that another important element of the project was to work closely with the doctors to see how they could apply this solution and how the information collected through AI could be used in clinical studies.</p><script async src=

"I like a lot that people are thinking about this, and I like the idea of ​​advancing the important clinical question of who will develop an illness, especially one as serious as the type 1 diabetes, "added Pajvani. "J & # 39; hope [I’d] love to see us in a place where technology can better understand how we help people. "

Pajvani said that as a clinician, he had not yet seen IA come out of the theoretical and practical treatment of patients. For her part, Ms. Hu said that she thought that the presence of machine learning in medicine would continue to "accelerate" and that it was a problem. an "extremely important" job that could be an indispensable tool for doctors.

This AI does not provide a definitive screening method today, but it offers a track on how machine learning tools could be used for faster and more likely type 1 diabetes diagnoses. Lives.

Images: Jovanmandic via Getty Images (doctor and patient); Gam1983 via Getty Images


Source link