IBM Watson Health wowed clinicians and physicians at RSNA with a demo of Watson cognitive assistant applications currently in development.
One platform — planned for release in 2017 — discovers discrepancies in cases of aortic stenosis (AS). The cognitive peer review tool sorts through EHRs and flags AS cases in which the diagnosis code and the EHR problem list, and the clinical evidence in the physician’s case notes, do not match. Clinicians then have the final say in accepting or rejecting Watson’s recommendation.
"We're working across six business pillars and imaging is one of them," Kurt Hammond, VP, sales and marketing told HCB News. "We currently have 20 clinical and vendor partners within the Medical Imaging Collaborative, who we think will each bring unique value in teaching Watson, and will assist in developing solutions."
AS is a common and serious heart valve disease, affecting 1.5 million Americans each year. According to IBM, a recent study found that nearly half of the diagnoses and findings from imaging studies and associated reports never make it into the problem list of billable diagnoses in a patient's EHR.
"Only half of AS patients are alive two years after diagnosis. It doesn't have to be this way ... if caught earlier," Dr. Ricardo Cury, FSCCT, director of cardiac imaging at Baptist Hospital of Miami, said in a statement. "Out of the gate the Watson cognitive tool could potentially provide big benefits."
Merge intrigued audiences with another common body part imaging procedure to benefit from big data and artificial intelligence: mammography. The Marktation work in progress allows the physician to label findings on an image using text or speech recognition. The text label is simultaneously stored on the image and pushed into the clinical report, helping to improve reading speed and accuracy.
According to IBM researchers, radiology images account for 90 percent of medical data sources and continue to increase every year. Most information still has to be extracted manually. Artificial intelligence can help reduce errors and allow radiologists to spend more time consulting with referring physicians and their patients.
IBM is also exhibiting caution to ensure that Watson platforms will play nice across existing systems.
"We know this has to be done without a forklift upgrade," Michael Klozotsky, Marketing Leader, Watson Health Imaging told HCB News. "No hospital administrator wants to have to replace their PACS system to accommodate a new quality platform. So all of our Watson partnerships must offer a vendor-neutral platform."
Merge and Watson presented several other developments at RSNA, including:
- A cognitive data summarization tool that aims to see expansively all available patient data sources, filter and present contextually relevant information in one view, customizable for radiologists and referring physicians.
- A cognitive physician support tool that seeks to recommend probability-driven differential diagnosis options upon analysis of vast amounts of patient, population and medical research data, to help inform physicians' decisions for the patient.
- The MedyMatch "Brain Bleed" App, a cognitive image review tool intended to help ER physicians diagnose a stroke or brain bleed in a trauma patient by identifying relevant evidence in a patient record.
- Watson Clinical Integration Module, a cloud application for radiologists that aims to help increase reader efficiency and counteract common causes of errors in medical imaging. This includes such factors as base rate neglect, anchoring, bias, framing bias and premature closure.
- A Merge lesion segmentation and tracking module, designed to help radiologists increase the speed with which they interpret and report comparison exams in cancer patient conditions that require longitudinal tracking.
IBM Watson is the first commercially-available cognitive computing capability representing a new era in computing. The system, delivered through the cloud, analyzes high volumes of data, understands complex questions posed in natural language, and proposes evidence-based answers. Watson continuously learns from previous interactions, gaining in value and knowledge over time.