All Mercy services are open. See safe options for care and the latest COVID-19 vaccine information.
Mercy Research is one of the nation’s largest fully integrated, community health system-based research organization. We conduct research in all our available service lines, and we manage more than 600 clinical trials with data from locations throughout our system. We offer:
“A physician is not only in a position to improve health of a singular person but also in a position to advance medical therapies through research and education to improve the health of a population. Mercy Research has seamlessly allowed us to start physician-initiated trials in Missouri, and, importantly, participate in large international trials to advance the care of patients not only in the United States but the world.” – Christopher Stout, MD, Mercy Vascular and Endovascular Surgery Clinic
Dr. Johnson Thomas was named Mercy Researcher of the Year for 2019. He is honored for his dedication to research and for exemplifying:
He is section chair of the Department of Endocrinology at Mercy Hospital in Springfield, MO. Dr. Thomas completed his endocrinology fellowship and internal medicine residency at Nassau University Medical Center in New York, where he received the Kenneth Simcic Award for Academic Excellence in Endocrinology. He received his medical degree from TD Medical College in Kerala, India.
Thyroid nodules are common. About half of all people develop them by age 60, and more than 90% of nodules are benign. An artificial intelligence (AI) tool researched and developed by Dr. Johnson Thomas, section chair of the Department of Endocrinology at Mercy Hospital Springfield and our 2019 Researcher of the Year, can potentially reduce the number of unnecessary biopsies.
Dr. Thomas presented his study, “AiBx, a Deep Learning Model to Classify Thyroid Ultrasound Images,” at the American Thyroid Association’s 89th Annual Thyroid Conference in fall 2019—and his research is being published in Thyroid Journal.
AiBx is an AI model that allows physicians to compare ultrasound images of undiagnosed and diagnosed thyroid nodules that share similar characteristics. This adds objectivity to the diagnostic process, but doctors still determine whether a biopsy is needed. “From a physician standpoint, I can use these images to see why the AI is telling me that a nodule is cancer or not cancer, and then make my decision,” Dr. Thomas says. “Many medical AI algorithms are black boxes, but our tool highlights the reasons behind the predictions, which increases physicians’ trust.”
AiBx was built using Epic ultrasound images of thyroid nodules from 482 Mercy patients who underwent either biopsy or thyroid surgery from 2012 to 2017. Later, 103 thyroid nodules were tested against the AiBx model to see how accurately it predicts thyroid cancer. The model performed well at predicting both negative (benign) and positive (cancerous) nodules.
American Thyroid Association (ATA) and American College of Radiology (ACR), the AiBx model has similar negative predictive value but greater sensitivity, specificity and positive predictive value. “The ATA and ACR systems both have low positive predictive values. That means when they say a nodule is cancer, many times it’s actually not cancer. With these guidelines, we’re still doing a lot of unnecessary biopsies. AI models like ours with high positive and negative predictive values can help avoid that,” Dr. Thomas says.
In addition to the AiBx study, Dr. Thomas was selected as Researcher of the Year for his significant contributions to other research on treatment-resistant thyroid cancer and indeterminate thyroid nodules.
Get information on becoming a Research Investigator.