Researchers develop near real-time brain tumor diagnosis

Sunday, January 12, 2020 - 3:07pm


design by Sherry Chen

In a study led by Michigan Medicine, researchers combined laser imaging with artificial intelligence to help predict brain tumor diagnosis within minutes.

A new publication in Nature Medicine highlights how the NIO Imaging System from Invenio Imaging allows surgeons to get a real-time image of a piece of tissue from surgery and use artificial intelligence to get a rough diagnosis of the tissue’s composition. 

Todd Hollon, chief neurosurgery resident at the University of Michigan, was first author of the publication. He conducted his research through Adjunct Assistant Professor Daniel Orringer’s lab, which focuses on using medical imaging to improve outcomes. According to data, this method is much faster than the current standard of care, which involves pathologists interpreting frozen sections of tissue.

“I think there was a real response from surgeons in particular,” Hollon said. “Everyone realized this is a big shift in the workflow.”

Chris Freudiger, co-founder of California tech startup Invenio Imaging, said he met Orringer at a conference in Sweden in 2009 where they were both presenting their work. From there, the two collaborated to utilize the technology Freudiger was presenting.

While the research’s clinical trial occurred at three universities, Hollon worked with the patients and surgeons at Michigan Medicine to assist with this leg of the trials.

“I think the patients certainly appreciated that at the University of Michigan we’re using new technologies to assist brain tumor diagnosis,” Hollon said. 

Balaji Pandian, a medical student at the University of Michigan and author on the publication, wrote some of the original code that started the artificial intelligence process and helped engineer the image processing software. Pandian explained he joined this lab because of his desire to get involved with translational research, which uses biological principles and scientific techniques to better health outcomes.

As this technology continues to develop, researchers hope to expand its use beyond the neurosurgical field. Arjun Adapa, a third-year medical student at the University of Michigan, was also an author on the publication and sees potential for the work to become more widespread.

“The beauty is that this kind of technology would be able to be applied across disciplines, across medical specialties,” Adapa said. “It would be able to improve intraoperative diagnosis for other types of tumors, such as lung tumors or breast cancer tumors.”

Adapa noted that researchers also want to expand this technology to understaffed areas that don’t have many trained neuropathologists.

“This study does have the potential to change medicine in general and surgery in ways that even now, even with the publishing of this study, we might not be able to fully grasp yet,” Adapa said. “The potential for this technology and our study to be a more ubiquitous thing that is being used across the country, I think that’s something that can happen in the next five to fifteen years, and I think that’s what makes it so exciting.”