We are in an era of unprecedented medical data collection. While many companies have heeded the call of physicians looking to derive insights from data collected from EHRs or connected apps and devices, there is an important silo of unstructured data that is often overlooked: medical images.
Today Watson Health announced a new medical imaging collaborative—comprised of 16 health systems—that will use the cognitive computing platform to derive insights from image data in combination with data from other sources including clinical trials, lab results, and de-identified, HIPPA compliant patient information.
For over a decade, IBM has been training Watson health’s cognitive computing platform using ground truth from practicing physicians. Watson will now augment its knowledge by ingesting resources—including images and ground truth—made available by the collaborative. Initially it will be trained to assist with the diagnosis and treatment of a limited number of conditions, such as those pertaining to cardiovascular health and eye health.
“The aim is to bring cognitive imaging to daily practice,” says Anne Le Grand, vice president of imaging for Watson Health. The hope is that Watson Health will effectively provide physicians with what LeGrand calls “augmented intelligence” that helps them land on the right diagnosis or care pathway.
“As Watson did not go to med school and is not a practicing physician, Watson will never deliver a differential diagnosis,” explains Le Grand. Rather, Watson’s imaging algorithm will supplement physician knowledge by to interpreting image data in the context of other medical knowledge that it has ingested. “It’s like having second reads from world leading opinions that will help doctors derive their own differential diagnosis,” Le Grand says.
Send this to a friend