Dr. Iltifat Husain’s physician take on how Watson could be used in Emergency Settings is at the end of this article
In its ever-growing quest to expand IBM Watson’s “physical senses,” the company has moved beyond the supercomputer’s ability to understand voice and text and now plans to teach Watson to “see,” enabling it to analyze video and still images in healthcare.
“Medical professionals will be able to tap Watson for analyzing images like X-rays, MRIs, angiograms, electrocardiograms, for example” says John Kelly, Senior Vice President, Solutions Portfolio and Research at IBM. In August, the company announced that it will start using Watson’s advanced image analytics and cognitive capabilities by joining forces with Merge Healthcare, which runs a large medical imaging management platform that’s currently used by more than 7,500 U.S. healthcare sites, as well as several clinical research institutes and drug companies.
In a recent press release about the acquisition, IBM pointed out that medical images are the largest and fastest-growing source of medical data, accounting for about 90% of all medical data currently available. This massive payload is virtually impossible for clinicians and researchers to gain real insights from without help. According to the company: “IBM plans to leverage the IBM Watson Health Cloud to analyze and cross-reference medical images against a deep trove of lab results, electronic health records, genomic tests, clinical studies and other health-related data sources, already representing 315 billion data points and 90 million unique records.” If they are successful, this will allow the healthcare organizations that work with Merge to compare new medical images with older images stored in a patient’s records, and even tap large data sources that include populations of patients with similar conditions. That in turn may enable clinicians to detect changes and anomalies that would otherwise be overlooked.
The Watson team also hopes to teach the computer to filter clinical and diagnostic imaging information to help clinicians develop recommendations and identify abnormalities, perhaps serving as a diagnostic aid for radiologists, and others, responsible for interpreting medical images.
In addition to helping clinicians make diagnostic and treatment decisions, the IBM/Merge partnership also hopes to: “Support researchers and healthcare professionals as they advance the emerging discipline of population health…”
Dr. Iltifat Husain’s take on this:
Imaging modalities are a huge part of medical decision making, and often a rate limiting step to care and intervention. Key ways Watson could be used would be to compare a patient’s CT images from prior, and point out to a radiologist if there are any obvious significant changes. The most ideal scenario would be Watson being used as a screening mechanism for life threatening pathology as soon as the image is completed. For example, picking up hemoperitoneum on a CT of the abdomen or pneumothorax on chest x-ray. Obviously – these types of patients often have physical exam manifestations that enable the diagnosis even without imaging, but there are times when presentations can be more subtle. Speeding up the ability to get a preliminary diagnosis would be instrumental in Emergency workflow and patient care — if Watson could be used to do that — it would change everything (ER length of stays, waiting room times, etc).