IBM Debuts Watson Health Cognitive Imaging Solution

Essentially, it uses cognitive text analytics to read both structured and unstructured information in cardiology medical reports, combines it with a variety of data from other sources, and extracts the relevant information needed to verify that key data, including the diagnosis, is accurately reflected throughout the electronic medical record (EMR).

It subsequently compares the clinical data with the EMR problem list and medical coding to empower users with accurate information. The first application introduced is designed to identify cardiovascular disease (CVD) patients suffering from aortic stenosis, a debilitating condition that causes shortness of breath, tiredness, and chest pain. IBM plans to supplement the release of the initial offering with nine other cardiovascular conditions, such as myocardial infarction (MI), valve disorders, cardiomyopathy, and deep vein thrombosis (DVT).

“Apart from identifying patients who might need follow-up care, Watson will look across patient populations to figure out similar patients who could benefit from follow-up visits, even if their ultrasounds or other images weren't analyzed by Watson,” said Anne Le Grand, vice president of imaging for Watson Health. “Then, Watson will move into predictive care, helping to recommend treatment for patients who could be at risk based on the computer's AI analysis. The response is twofold: improving the quality of diagnosis and the consistency of diagnosis.”

“If you have an ultrasound image of a heart, the quality [of care] could be affected by how good the technology is and also by variations in how different physicians may interpret those images,” said cardiologist Jaime Murillo, MD, of Sentara Healthcare (Norfolk, VA, USA). “With Watson, we're looking at standardizing and improving accuracy of diagnostic interpretations that result in better patient care and more accuracy.”

Watson--named after Thomas Watson, the first president of IBM--incorporates a technology called Deep Question and Answer (DeepQA) that allows it to analyze human language and quickly process huge amounts of information. The same technology can be applied to help Watson sift through the equivalent of about a million books or roughly 200 million pages of data, analyze the information, and provide an answer in less than three seconds.