The AI ​​healthcare revolution begins

This article is part of Future Explored, a weekly guide to world-changing technology. Receive articles like this every Thursday morning straight to your inbox.
subscribe here.

Artificial intelligence could be the next world-changing medical advancement.

From autonomous surgical robots to software that finds signs of disease doctors might miss, advanced AI is proving to have a powerful impact on medicine, but patients embrace Dr. AI. Is not it?


A doctor needs to know the cause of a patient’s illness in order to treat him, and diagnosis is one of them, if not more. of — The most important part of medicine. It may also be where AI can have the biggest impact.

By training AI with a wealth of existing data, such as patient records and medical scans, researchers have developed a system that can analyze new cases and predict what might go wrong. Often faster and with equal or better accuracy than human experts.

One of the latest examples is AI trained to identify small fractures. This indicates that the patient has osteoporosis, a disease in which bones become weak and brittle. If detected early, patients can take steps to strengthen their bones, but often the disease persists until it undergoes a painful and sometimes disabling break. Not diagnosed.

“Our AI automatically examines the spine.”

Ellen Yilmaz

Developed by German researchers, the new AI looks for small fractures indicative of osteoporosis in computed tomography (CT) scans ordered for other reasons, such as examining a patient’s lungs, and ranks their severity. To do.

“Our program can run in the background during such inspections,” says lead author Eren Yilmaz. “It automatically examines the spine and shows vertebral fractures that otherwise could not be detected.”

The team tested AI on 159 CT scans of the spine featuring a total of 170 fractures and correctly classified 90% of scans with fractures and 87% of scans without fractures. The AI ​​is still in development, but Yilmaz said that if it were introduced into routine radiology, it could be an “early warning system to prevent the serious consequences of osteoporosis.”

AI correctly classified both fractures on this CT scan.Credit: Ellen Ilmaz

drug discovery

Drug discovery usually begins with identifying targets, usually proteins, that play a role in whatever you want to treat. If the target is ‘druggable’, i.e. has a structure that can be bound by another molecule, the next step is to find a drug that favors the target (and No major problems anywhere else).

For example, protein cyclooxygenase produces chemicals that cause pain, making it a good target for analgesics. Aspirin binds with cyclooxygenase to prevent the production of those chemicals, thereby reducing pain.

With an estimated 3,000 druggable proteins in our bodies and an almost infinite number of drugs that can affect them, it’s important to find combinations that do what we want. is. safely This is a slow, expensive process, and one that AI is currently disrupting.

“We are proud to be one of the few AI-driven biotech companies to reach the clinical stage.”

Alice Chan

In October 2022, biotech company Verge Genomics began one of the first clinical trials of an AI-discovered drug candidate aimed at treating ALS, also known as Lou Gehrig’s disease.

After feeding AI with more than 11 million data points from a genetics database and tissue samples from ALS patients, the Verge identified the drug as a molecule called VRG50635. The system then identified a protein called PIKfyve that may be involved in this disease, suggesting that VRG50635 should be able to inhibit it.

“We are not only one of the few AI-driven biotech companies to reach the clinical stage, but also one of the first to propose novel clinical compounds against novel targets that are fully discovered and developed. Verge CEO Alice Zhang said:

“This is an important milestone for both longevity research and the application of AI to drug discovery.”

Felix Wong

More recently, the biotech company Integrated Biosciences sought to find drugs that could wipe out senescent cells (“zombie” cells that have stopped growing but are not dead) associated with many age-related diseases. announced that he had trained

Some scientists suspect that drugs that can get rid of them, called “senolytics,” could help us live longer and healthier. However, the field is still relatively new, with only a few candidates reaching clinical trials.

Integrated’s AI has been able to identify three new senolytic drugs with desirable properties, such as indications of high oral bioavailability, meaning that they can be taken as tablets, from a collection of over 800,000 molecules. I made it. One drug was shown to eliminate senescent cells as expected in experiments with aged mice.

“This study result is an important milestone for both longevity research and the application of artificial intelligence to drug discovery,” said Felix Wong, co-founder of Integrated Biosciences.

“These data suggest that we have computationally explored the chemical space and identified multiple anti-aging compound candidates with greater potential for clinical success compared to the most promising examples of this kind being studied today. It shows what can be done,” he continued.


Today, rather than just bending over and swinging a scalpel, many surgeons stand behind a computer monitor and use a joystick to precisely guide an on-screen robotic arm into the patient’s body.

These robots make it easier for doctors to perform complex and delicate surgeries while minimizing risk and pain, but for now they still rely on the know-how and skill of the doctor operating the joystick.

AI robots did their jobs more accurately and faster than human surgeons.

In January 2022, Johns Hopkins University (JHU) reported that AI-powered smart tissue-autonomous robots (STARs) performed delicate surgery on live pigs by suturing both ends of their intestines.

Not only is this the first example of a robot performing laparoscopic surgery autonomously, the robot performed the job more accurately and quickly than a human surgeon.

“By incorporating novel suturing tools, imaging systems, machine learning algorithms, and robotic control, the STAR system has the capability to overcome the challenges of autonomous laparoscopic surgery in soft tissue,” said senior author Axel Krieger. said.

STAR robot is active.Credit: Jiawei Ge / IMERSE Institute / Johns Hopkins University

big picture

The JHU team says STAR could make surgery more predictable and affordable. Patient outcomes are no longer dependent on the skills of local hospital surgeons. But the company’s robots face the same hurdles as any other AI in medicine, he notes. Citizens are wary of AI’s contribution to their healthcare.

In February 2023, the Pew Research Center reported findings measuring American acceptance of AI in healthcare, with 60% of respondents saying they would be uncomfortable if their physicians relied on AI to diagnose illness or make treatment decisions. I answered that I feel Only 38% believed it would improve their health.

“In many ways, our research highlights potential blind spots for AI researchers.”

Sanjay Aneha

Meanwhile, a similar survey conducted by Yale Cancer Center in 2022 found that more than 70% of respondents were at least somewhat uncomfortable receiving a diagnosis from an AI, even if it could not explain how it reached certain conclusions. replied. It was 90% accurate.

A Yale University study also found that patient comfort varied depending on the AI’s behavior. His 55% of respondents said he was at least somewhat comfortable having an AI read his chest x-ray, but if the AI ​​was diagnosing cancer, that figure was 31.2% for him. decreased to .

To maximize the potential benefits of AI in medicine, developers should not only focus on AI training, but also on what patients want to get from AI and how best to alleviate their concerns. You also need to think about how to

“In many ways, our study highlights a potential blind spot for AI researchers,” said Sanjay Aneja, senior author of the Yale study. We need to address this as we move forward.”

We look forward to hearing from you! If you have any comments about this article, or tips for future Freethink stories, please email us at:

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *