“The DNA of the university is really to do user inspired work,” says Gene Tunik, the director of AI + Health Sciences at Northeastern University.
Before doctors can make a cancer diagnosis, they must perform a series of exams and lab tests to get a fuller picture. It may take weeks or months before there is a clear determination.
Saeed Amal, a Northeastern University bioengineering professor and a faculty member of the Roux Institute, has developed AI technologies to help medical professionals speed up that process. He has created online web tools to assist in diagnosing various forms of cancers, including breast cancer and prostate cancer.
But this technology can only get better if it is used by medical professionals working in clinics and other health care centers, explains Gene Tunik, the director of AI + Health Sciences at Northeastern University’s Institute for Experiential AI.
That’s why partnerships like the one recently formed by Northeastern University in collaboration with Prima Care and Santovia Path AI will be key in helping improve these technologies in the long term and help better support clinicians, Amal and Tunik explain.
“The DNA of the university is really to do use-inspired work,” says Tunik, who is also a professor in the Bouvé College of Health Sciences. “Beyond just generating knowledge and information, that type of use-inspired work really requires getting all of your stakeholders at the table early.”
Prima Care is a southern New England medical health care company that has 180 providers throughout Fall River, Somerset, Tiverton and Westport. Santovia Path AI will provide primary funding for the project.
Over the next two years, Northeastern researchers will work with the companies to develop AI modeling technology to better “identify potentially cancerous cells in digitized images of biopsy samples,” the researchers explain.
This work will build off the web tools Amal has already created in his previous research that allows pathologists to upload biopsy images in a few short clicks.
“Our goal is to evaluate the algorithms together with these tools and their usability in a real clinic,” Amal says. “Eventually, we want to evaluate it with multiple pathologists from multiple centers.”