“When my father was misdiagnosed and administered the wrong medication placing him in a coma nearly 20 years ago, I saw firsthand the need for technology to help physicians make accurate decisions,” said Tanveer Syeda-Mahmood, IBM Fellow and Chief Scientist of the Medical Sieve Radiology Grand Challenge Project at IBM Research – Almaden in San Jose, Calif.
Dr. Tanveer Syeda-Mahmood, IBM Fellow & Chief Scientist, Medical Sieve Radiology
This week in Chicago, Dr. Syeda-Mahmood’s mission meets the real world as IBM Research debuts a new Watson-powered demo that shows the future of Artificial Intelligence (AI) in radiology. The demo is the result of a shared vision by Dr. Syeda-Mahmood and Dr. Eugene Walach from IBM Research – Haifa to help radiologists make accurate patient diagnoses quickly and easily.
In any given day, radiologists can review up to thousands of medical images to make health diagnoses. To date, accuracy has relied mainly on medical professionals piecing together multiple sources of clinical information visually and manually to make critical decisions, including electronic health records, research publications and other data. A recent explosion in the amount and variety of medical data available digitally has made it virtually impossible for physicians to keep pace.
This week in Chicago, radiologists can experience how IBM researchers have harnessed the cognitive computing power of IBM Watson to analyze large amounts of imaging and text in electronic health records. As part of the educational sessions at the Radiological Society of North America (RSNA) annual meeting, a radiologist can select a sample patient case from breast and cardiac radiology specialties and see how a Watson-powered prototype, called Medical Sieve, surfaces insights from the case as it understands, reasons and learns from text and imaging data in real time.
The educational demo was developed at the request of and in collaboration with the Radiological Society. Based on more than a decade of work by IBM researchers and data scientists, the demo uses sophisticated medical imaging, deep learning, and clinical inference technologies to create a model of a single-view, compact summary report for clinicians. Ultimately, the work will inform a commercial solution that aims to increase efficiency in provider workflows and support doctors in making accurate, speedy diagnoses.
“We designed this Watson-based demonstration to show physicians that soon they can navigate an abundance of digital data – structured and unstructured, text and images – and make informed decisions based on relevant and current information,” said Dr. Syeda-Mahmood. “More importantly, we can analyze a broad array of medical data and derive Watson-powered insights that are meaningful to doctors.”
Added Flora Gilboa-Solomon, research manager for Medical Sieve at IBM Research – Haifa, “The Watson technology we are demonstrating this week to the radiology community is unique in its ability to learn and improve over time, similar to a human assistant who gains broader experience by analyzing new data.”
Medical Sieve is just one example of the IBM Research efforts exploring the potential for cognitive computing to help doctors analyze medical imaging. As a result of many years of developing deep learning algorithms globally, IBM Research is developing new ways for machine learning to help physicians visually detect disease from medical images. IBM Research teams in Switzerland and Brazil are participating in hackathons to use machine learning to help physicians diagnose cancer from biopsy tissue samples as tiny as a pinhead. In Yorktown Heights, N.Y. and Melbourne, Australia, IBM Research is using machine vision to help physicians identify skin cancer. IBM Research – Australia also applies cognitive visual recognition to assist ophthalmologists with the early detection of eye diseases such as diabetic retinopathy and glaucoma.
The work of IBM Researchers is informing new offerings in development by Watson Health Imaging and by Merge Healthcare. At RSNA 2016, seven products will be previewed at the medical meeting.
In our recent paper, "Optimal periodic closure for minimizing risk in emerging disease outbreaks," published in PLoS One, we developed a technique to calculate the optimal duration of a periodic lockdown during an outbreak of an infectious disease where there is no cure or vaccine. Our findings are different from the lockdown duration being widely applied, today.
In a newly published paper “Quantitative language features identify placebo responders in chronic back pain” in the peer-reviewed journal PAIN, we report the first proof-of-concept that uses AI to analyze patients’ clinical trial experiences. The AI quantifies a placebo response in patients with chronic pain and distinguishes those who respond to placebo from those who do not.