Medicine has seen some major advancements in 2022 thanks to AI-powered tools and research, from a new way to detect Parkinson’s, type 2 diabetes, birth defects, and more. It’s clear that researchers are going all in with AI to empower medicine so that patients can receive the best care possible, and medical professionals are armed with the best technology. Here is the top AI healthcare/medical news of 2022!
New AI Model Predicts the Onset of Type 2 Diabetes Within 12 Hours
It was reported at the beginning of December that a new AI model was able to predict the onset of type 2 diabetes within 12 hours. What makes this a remarkable advancement is the traditional method would take three months!
The way it would work is that a blood test would be used to measure a person’s average blood sugar level over that period. But in regions where hospital input is incomplete or scarce, such as less densely populated regions and developing nations, this could become a game changer in preventative medicine, affecting the lives of millions worldwide if proven to be able to be scaled.
Scientists at NTU are Using AI to Identify Mental Illnesses to Help with Proper Diagnosing and Treatment
Like with many physical diseases, mental and neurological disorders often are best treated if caught early and a plan of action is put into place. But unlike with physical diseases, it’s extremely rare to see physical manifestations of these illnesses until it’s far too late. But thanks to the world of researchers in Singapore, it could be a thing of the past.
A group of researchers from Singapore’s Nanyang Technological University (NTU) is working with emerging AI tools to fight mental illnesses when symptoms are too mild to detect, but when treatment could be most effective. So far, they’re attempting to process vast datasets in a bit to identify the biomarkers that point out mental illness manifesting in people. Though the research is still early, it looks quite promising.
New Neural Network Can Detect Parkinson’s Disease by Breathing Patterns
A new neural network from a team at MIT can now detect Parkinson’s Disease early before symptoms become apparent, by simply watching your breathing patterns. This milestone came thanks to Dina Katabi, PhD, a professor of electrical engineering and computer science and principal investigator at the MIT Jameel Clinic.
Her team at MIT developed an artificial intelligence model through a series of connected algorithms that mimic the way the human brain operates. They were able to demonstrate that the AI assessment of Parkinson’s could be done nightly, in a patient’s home while they slept using a device that is about the size of a Wi-Fi router.
This is a milestone in preventative medicine that has the potential to expand the treatment options of those affected by Parkinson’s due to early detection.
DeepMind’s AlphaFold Discovers Nearly Every Known Protein
During the latter part of the summer, DeepMind discovered nearly every single protein currently known to science, making major AI healthcare news. Using AlphaFold, which was developed back in 2018, the open-source program uses machine learning algorithms to predict a protein’s three-dimensional structure.
Cardiologist Eric Topol from the Scripps Research Translational Institute explains in a statement about the news, “Determining the 3D structure of a protein used to take many months or years, it now takes seconds…With this new addition of structures illuminating nearly the entire protein universe, we can expect more biological mysteries to be solved each day.”
This is a breakthrough for the microbiology community as well as the medical community and only time will tell the changes and advancements that will be ushered in by this discovery.
Researchers Use AI and Deep Learning to Identify Potential Birth Defects
Back in July, researchers at the Univesity of Ottawa used an artificial intelligence-based deep learning model to assist doctors in rapidly reading ultrasound images. This rapid read assisted doctors in locking possible birth defects in the earliest periods of pregnancy. The study’s goal was to demonstrate the potential of using a new deep learning tool to bot locking and properly identify cystic hygroma in first-trimester ultrasound scans.
This was a breakthrough due to the difficulty of identifying these issues this early in pregnancy due to the lack of detail in images, the size of the baby, and normal fetal movements. But with this tool, the Canadian team’s model was able to achieve an accuracy of over 93% overall, which is an amazing first step with this new tool.
Deep Learning Could Aid Medical Professionals in Improved Medical Ventilator Control, Google AI Proposes
Back in March, we reported a breakthrough that could fill in gaps where current ventilator control methods have difficulty. Currently, ventilators use the PID (Proportional, Integral, Differential) controls. They adjust the device based on discrepancies between the measured and target air pressures. But though they have reliability, they are prone to over and undershooting targets due to differences between patients. Because of this, the current method requires continuous monitoring to maintain effectiveness.
With Google’s ventilator control solution, they use machine learning instead of manual controls. This design requires much less manual intervention by predicting air pressure changes based on real-time data. Armed with this, it can adjust to prevent drastic pressure changes instead of responding to them.
This new design can not only potentially save lives, but free up much-needed man-hours and reduce human errors as the data sets grow.
Well, these are our top AI Healthcare/Medical news of 2022. The year has been an exciting year with many advances within the field of medicine, laying the foundations for future advancement in medical care empowered by artificial intelligence. If you’re looking to make the news in the future (hopefully in a non-controversial way!) then learn more about what’s big in data science at ODSC East 2023 this May 9th-11th, currently 70% off!