ReMix, a new technology in the works at MIT's Computer Science and Artificial Intelligence Laboratory, seeks to make progress toward a new paradigm of treatment – one that's less reliant on endoscopy and surgery, but instead relies on sensors and what scientists call an "in-body GPS" to help make diagnoses and guide drug administration.
The goal, researchers said, is a new approach to patient care that's less expensive, invasive and time-consuming than has been previously been possible.
MIT has always been a pioneer in envelope-pushing medical advances, of course. Recently much of its focus has been on AI and machine learning. CSAIL researchers, for example, have recently been exploring ways AI can be deployed to improve electronic health records, streamlining ICU data and helping with predictive models.
ReMix, which is being developed with help from with clinicians at Massachusetts General Hospital, depends on complex algorithms to work, but also leverages other technologies – spotting the location of ingestible implants by way of low-power wireless signals.
MIT professor Dina Katabi has pioneered the use of such signals to detect subtle movements and even vital signs from a distance. CSAIL researchers have tested the ReMix on animals and were able to show that the technology can track implants accurately, within a centimeter or so. The aim is for similar techniques to someday be used in human patients, helping target drugs to specific parts of the body.
In their initial testing of ReMix, the researchers implanted a small sensor in the animal and used a wireless device that reflects radio signals to track it. They deployed a specific algorithm to pinpoint the location of the marker – which doesn't have to transmit a signal of its own, only to reflect the one aimed at it by the wireless device.
The challenge, MIT scientists said, is that wireless signals can bounce off many parts of a human body – and are significantly more powerful reflecting off the skin than on any implantable markers. So they developed a small semiconductor that helps differentiate different signals by combining them and allowing the filtering of irrelevant frequencies.
That said, there's plenty of room for ReMix to be honed and fine-tuned before its ready to use in clinical settings – its location accuracy would need to be down to millimeters, rather than centimeters – but CSAIL researchers say they're interested in how this "GPS" could help locate tumors, for instance, or one day guide the precision application of proton therapy for a wider array of cancers than is currently possible.
The plan next is to combine wireless signals with other data such as MRI scans, helping hone the system's accuracy, and to do further research to assess the how the algorithm can account for the diverse complexities among different patients' physiologies.
"We want a model that’s technically feasible, while still complex enough to accurately represent the human body," Deepak Vasisht, a PhD student a CSAIL and lead author of a new paper on MIT's ReMix research. "If we want to use this technology on actual cancer patients one day, it will have to come from better modeling a person’s physical structure."
Email the writer: [email protected]
Source: Read Full Article