Researchers from the Indian Institute of Science (IISc), in collaboration with the Aster-CMI hospital, following a study on carpal tunnel syndrome (CTS), have developed a artificial intelligence can identify the median nerve in ultrasound videos and detect CTS.
The study was published in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control.
Carpal tunnel syndrome: here's what new research says
Carpal tunnel syndrome occurs when the median nerve, which runs from the forearm to the hand, becomes compressed in the carpal tunnel part of the wrist, causing numbness, tingling, or pain.
It is one of the most common nerve-related disorders, particularly affecting people who perform repetitive hand movements, such as office staff who work with keyboards, assembly line workers and athletes.
Doctors currently use ultrasound to visualize the median nerve and evaluate its size, shape, and any abnormalities. “But unlike X-rays and MRI scans, it is difficult to detect what is happening in ultrasound images and videos,” explains Karan R Gujarati, first author and former MTech student in the Department of Computational and Data Sciences (CDS), IISc.
“At the wrist, the nerve is quite visible, its boundaries are clear, but if you go down to the elbow region, there are many other structures, and the boundaries of the nerve are not clear.” Monitoring the median nerve is also important for treatments that require doctors to administer local anesthesia to the forearm or block the median nerve to provide relief from pain caused by carpal tunnel syndrome.
To develop their tool, the team turned to a machine learning model based on the transformer architecture, similar to the one that powers ChatGPT. The model was originally developed to detect dozens of objects simultaneously in YouTube videos.
The team eliminated computationally expensive elements of the model to speed it up, and reduced the number of objects it could track to just one: the median nerve, in this case.
They collaborated with Lokesh Bathala, consultant neurologist at Aster-CMI Hospital, to collect and annotate ultrasound videos of both healthy participants and people with carpal tunnel syndrome, to train the model. Once trained, the model was able to segment the median nerve into individual frames of the ultrasound video.
“Imagine a video of an autonomous car. If the car moves on the road, you want to follow it,” explains corresponding author Phaneendra K Yalavarthy, professor at CDS. “Similarly, we are able to track nerves throughout the video.”
The model was also able to automatically measure the cross-sectional area of the nerve, used to diagnose carpal tunnel syndrome. This measurement is performed manually by a sonographer.
“The tool automates this process. It measures the cross-sectional area in real time,” explains Bathala. It was able to report the cross-sectional area of the median nerve with greater than 95% accuracy in the wrist region, the researchers say.
While many machine learning models have been developed for screening CT and MRI scans, very few have been developed for video ultrasound, especially nerve ultrasound, explains Yalavarthy.
“We initially trained the model on a nerve. Now we will extend it to all the nerves of the upper and lower limbs,” says Bathala. He adds that he has already been used as a pilot test in hospitals for the detection of carpal tunnel syndrome.
“We have an ultrasound machine connected to an additional monitor running the model. I can observe the nerve, and at the same time, the software tool also outlines the nerve. We can see its performance in real time.”
Bathala says the next step for a therapy for carpal tunnel syndrome would be to look for ultrasound machine manufacturers who can integrate it into their systems. “This type of tool can help any doctor. It can reduce inference time,” she says. “But obviously the final diagnosis will have to be made by the doctor.”
#Carpal #tunnel #syndrome #arrives