The use of artificial intelligence (AI) and machine learning (ML) in the field of otolaryngology is in its infancy; most current initiatives are in the research stage. But the benefits of employing these technologies show great promise.
Explore This IssueFebruary 2019
AI requires large datasets of patient characteristics, disorders, and outcomes. The technology can be used in almost any subspecialty in otolaryngology. “The most direct application is to develop clinical decision support systems from data,” said David R. Friedland, MD, PhD, professor and vice chair of otolaryngology and communication sciences at the Medical College of Wisconsin in Milwaukee. This could be as simple as recommending specific tests, given a specific complaint and symptomatology. AI would mine prior data and identify which test provides the most clinically and cost-effective benefit, and would use real patient data, in real time, to constantly upgrade its evidence-based recommendations.
Andrés Bur, MD, assistant professor of otolaryngology–head and neck surgery at the University of Kansas in Kansas City, believes that AI’s biggest impact may be to increase personalized care in otolaryngology. Companies such as Amazon and Google use ML technologies to understand their customers based on data. “We can use AI to increase personalization in otolaryngology and in health care in general by doing the same thing,” he said. ML can be used to predict clinically important outcomes and determine the best treatment for an individual patient based on data. These types of predictive algorithms could, in the future, be used to build clinical decision support to make personalized treatment recommendations supported by data.
“By analyzing how similar patients have responded to past treatments, ML can provide information based on many more patient experiences than any individual physician could incorporate into their medical decision making,” Dr. Bur said.
AI could also revolutionize how otolaryngologists interact with electronic health records (EHR). According to recent estimates, for every hour that physicians provide face-to-face clinical care to patients in the outpatient setting, they spend nearly two additional hours on EHR documentation and desk work (Ann Intern Med. 2016; 165:753-760). “By recording and automatically extracting content from clinical encounters using natural language processing, virtual scribes have the potential to reduce the burden of clinical documentation,” Dr. Bur said.
In a review article that has been accepted for publication in Otolaryngology – Head and Neck Surgery, Dr. Bur and his team of investigators identified 54 articles in the otolaryngology literature focused on AI. “Interestingly, more than half were published in the past two years, highlighting the recent explosion of interest in AI research,” he said. “There is great interest in AI and ML in our field; articles published in every specialty in our field center around using AI technologies.”
Aaron C. Moberly, MD, an assistant professor of otolaryngology-head and neck surgery at The Ohio State University Wexner Medical Center in Columbus, and colleagues have used AI to develop software called “Auto-Scope” to help clinicians perform more accurate diagnoses for the ear using digital otoscopy. “This initially involved classifying images as normal or abnormal, but will eventually expand to build enhanced composite images and provide more specific information on particular types of pathology,” he said.
Dr. Friedland and his team are studying ML and AI components in vestibular disorders. Specifically, they are looking at methods to aid otolaryngologists and primary care providers in formulating better differential diagnoses when a patient has dizziness. They use a direct-to-patient survey to learn about their symptoms. “By correlating large numbers of patient responses with an expert-provided diagnosis, we hope to identify patterns, which may improve diagnostic accuracy,” he said.
Another area Dr. Friedland has studied is using natural language processing, a form of AI that interprets natural speech or written documentation, to identify expert descriptors of specific vestibular conditions. “This may help non-experts develop language usage that better obtains a patient history and allows for more accurate formulation of a differential diagnosis,” he said.
Naweed Chowdhury, MD, assistant professor of otolaryngology at Vanderbilt University Medical Center in Nashville, and colleagues have investigated applications of AI and ML in rhinology and skull base surgery. They showed that a convolutional neural network called “Inception” could be retrained to classify the patency of the osteomeatal complex on a computerized tomography scan with about 85% accuracy (Int Forum Allergy Rhinol. 2019;9:46-52). More recently, they used a ML model known as a random forest to demonstrate that the mucus cytokines IL-5 and IL-13 were predictive of baseline olfactory function in chronic sinusitis patients.
In rural areas, digital slides read elsewhere could be used to assist in decision making and counseling patients.