RSNA Session offers a Radiology Tutorial of NLP & Machine Learning Technologies

Natural language processing and machine learning technologies are making their way into the healthcare industry, and radiologists need to become educated on what these technologies do and how they can help improve the care delivered.


Natural language processing and machine learning technologies are making their way into the healthcare industry, and radiologists need to become educated on what these technologies do and how they can help improve the care delivered.

Session SSA11 at the RSNA Conference will feature nine presentations in a 90-minute deep dive on Sunday, November 29 at 10:30 a.m. George Shih, M.D., associate professor of radiology at Weill Cornell Medical College and vice chair of informatics for the radiology department is one of three moderators and will present a brief keynote address to start the presentations.

Dating sites and self-driving cars are among the services now using NLP and machine learning technology; the job of Shih and other presenters is to explain how those technologies fit in with radiology. For radiologists skilled in technology, the session will offer technical abstracts. But the keynote talk will help educate the non-techies on how these technologies work and give examples of them being used in radiology, Shih says.

For instance, computer-aided diagnosis has been used in mammography for several years and more recently is being used to find nodules in lungs. Now, when a radiologist orders a CT scan of the head—of which there may be 10 to 15 different types of protocols—machine learning technology can potentially select the most appropriate type of scan.

Natural language processing and machine learning technologies also can “read” radiology reports and classify them by disease to quantify and track the diseases a resident should be seeing when on their rotations so that each resident is seeing enough of the diseases, Shih explains.

“My main goal is to introduce these technologies to non-technical radiologists and inspire them to become more familiar with the technology,” Shih says. “They will be affected by this technology and should better understand it so we can influence its development.”

More for you

Loading data for hdm_tax_topic #better-outcomes...