October 2, 2017 — Thirty radiologists from seven of Strategic Radiology’s 26 member practices will contribute to work undertaken by the Watson Health Medical Imaging Collaborative to advance natural language processing capabilities used in future cognitive medical imaging solutions. The radiologists will provide expert annotation of de-identified medical images that will be used by IBM to train Watson to recognize various medical imaging terms that would be found in a patient’s electronic medical record (EMR), as well as providing feedback on solution design.
“Machine learning is here, and we want to be central to understanding how this important new tool can be leveraged to improve quality and value in radiology,” said A. Van Moore Jr, M.D., FACR, chair and CEO of Strategic Radiology. “We look forward to working with IBM to develop a role for artificial intelligence in medical imaging.”
Just as search engines categorize information in order to index and rank content, Watson is being trained by radiologists to review patient records that include a radiology report and categorize the relevant text from those records into groupings. For example, if Watson can be taught to recognize relevant clinical information in free text notes, a time-consuming task could be performed by Watson and provided in advance to the interpreting radiologist.
Strategic Radiology is a member of Watson Health’s medical imaging collaborative, a global initiative comprised of leading health systems, academic medical centers, private radiology practices, ambulatory radiology providers and imaging technology companies that are finding ways to use medical imaging to identify and predict the risk of cancer, diabetes, and diseases of the eye, brain and heart and related conditions.
Watson is a cloud-based system that analyzes high volumes of data, understands complex questions in basic language and proposes evidence-based answers. Watson can continuously learn, gaining knowledge over time from previous interactions.
For more information: www.ibm.com/watson
May 01, 2024 
