Smile News: New AI Tool Can Spot Early Warning Signs in MRIs
In a world where Artificial Intelligence (AI) is becoming more and more normalized, there is a lot of fear and controversy surrounding it. Many condemn AI, seeing it as something that will take away jobs, harm the arts, and destroy the planet. All of these concerns are valid and real, but recent developments with AI in the medical world has also shown that it can bring a certain level of good to the world when used correctly.
A new Artificial Intelligence tool has been developed, aiming to catch early warning signs for brain diseases in brain MRIs. Researchers at Mass General Brigham have been tirelessly working on this tool, which they hope will be able to read brain MRI scans much more accurately than the human eye. The tool looks at scans in a much broader way, searching for patterns that point to current or future diseases before symptoms even arise.
This new tool is named BrainIAC, which stands for Brain Imaging Adaptive Core. Researchers have trained it using almost 49,000 diverse brain MRI scans, teaching the tool to spot important neurological health signs and red flags.
Mass General Brigham’s Dr. Benjamin Kann, leader of this development, expressed his enthusiasm for what this tool can do, stating, “Anything that you could see from an MRI that you want to know more about it, this model has the potential to help you get there.”
The current model of this tool has the ability to detect brain age, dementia risk, brain tumor mutations, and the chance of survival from brain cancer. While hopeful, the team also remains cautious and aware that the tool has not been proven to be perfect.
“I think with these AI tools, we do have to be careful,” Kann added. “The results can look really promising and look really accurate when you study them, and then, when you actually roll them out into patient care, we have to make sure that the accuracy is still the same.”
The research team will continue working on the tool until they have enough data to ensure everything is running as expected. Only then will they roll it out to use with patients.
