When Hend-Arkaderi was researching how saliva can predict the risk of diabetes and the severity of coronavirus infections, he measured hundreds of bacterial samples for each patient and collected large numbers of saliva samples, thousands of samples.
She and her colleagues couldn’t analyze all the bacteria in each saliva sample, so they had to choose which ones to use. It was a difficult decision to make because we knew it might not end up being a representative sample.
A few years ago, Arcadelli learned that some of his colleagues were using machine learning, a type of artificial intelligence, to analyze large amounts of data. “I was interested,” said Arcadelli, an assistant professor in the Department of Public Health and Community Services at Tufts University School of Dentistry.
She attended a conference and learned that machine learning can analyze thousands of samples and predict diseases in minutes. Arcaderi was hooked. She took two courses on AI at MIT three years ago and continues to analyze large datasets. Mr. Arcadelli currently serves as director of the Dental AI Lab, a joint appointment with the Tufts Institute for Artificial Intelligence (TIAI). In that role, she teaches dental students how to use artificial intelligence through the Dental AI Lab and the new Artificial Intelligence in Dentistry class.
Although Arkadelli first became interested in AI through research, she also appreciates the technology’s ability to enhance patient care in dental practices. AI helps with scheduling, billing, insurance, patient records, and documentation, all of which are time-consuming tasks that support dental care.
Dentists “spend a lot of time documenting every procedure and treatment plan,” Arcadelli says. “If we can use AI to free up a little bit of dentists’ time, they can focus more on what matters most: treating patients and communicating with them.”
Predictive power of AI
AI can also be a powerful tool in preventive oral health care.
Although we know some risk factors, such as socioeconomic status, diet, and smoking, we do not yet have the ability to build software that can predict diseases before they occur. ”
Hend Alkaderi, Assistant Professor, Department of Public Health and Community Services, Tufts University School of Dentistry
That’s where AI comes in, analyzing electronic dental records and creating algorithms that predict which patients are most likely to develop cavities, oral cancer, and other diseases.
Arcadelli likens the system to recommendations made for users by Netflix, Google, or Amazon. Algorithms derived from what you watch, search, and buy drive recommendations of things you might like.
The vision of a streamlined dental office powered by AI to provide pre-diagnostic care is planting the seeds for an ongoing project at the Tufts Dental AI Lab. The lab, which soft-launched in July with a research and education mission, is part of TIAI and the Department of Public Health and Community Services. It is funded by Beyond Limits, a California-based technology company.
For now, the young lab is focused on research and publishing academic papers. “In the future, we want to translate our research into practice. We want to see all these projects come to fruition,” Arcadelli says. When lab members build AI models to predict disease, they ultimately want to build software that determines which patients will develop certain symptoms and integrates into the office’s electronic health system.
One of our current projects is developing an AI model that can read X-rays and detect anything suspicious. “The dentist can cross-check or verify after the AI, which saves a lot of time,” says Arcadelli.
AI in curriculum
Each project involves students and faculty from the School of Dentistry, as well as data and computer scientists from the Tufts Institute for Artificial Intelligence. “We are experts in our field and they are data scientists,” Arcadelli says. Additionally, while TUSDM personnel do not build AI models, they do need to understand the basics of AI analysis in order to collaborate and collaborate with more technical personnel.
To this end, Mr. Arcadelli led Tufts’ first dental artificial intelligence course during the fall semester. There, 240 third-year dental students attended 10 lectures covering key concepts in AI, ethics, dental treatment planning, and how AI can increase access to dental care for people living in remote areas.
Students evaluate existing dental AI applications according to criteria such as ethics and data quality. They are also collaborating with the same scientists in the Dental AI Lab to work as a team on a capstone project.
AI is being introduced into dentistry for patient care, research and academia, Arcadelli said. “We want our students to be prepared to deal with AI in clinical settings, consider ethical considerations, and critique and use AI.”
AI organization for dental students
The role of AI in dentistry is widely misunderstood compared to other fields such as engineering, systemic medicine, and business, says Yash Brahmbhatt, D27, founder of the Association for Artificial Intelligence in Dental Research and Education.
Brahmbhatt, the first Tufts dental student intern at the Tufts Institute for Artificial Intelligence, started the club to encourage dental students and faculty to learn more about AI and explore more avenues for AI-based dental research. “When starting the conference, I was inspired by the lack of an organization focused on combining dentistry, research, and AI,” he says.
“We hope more people understand that AI will not replace dentists, but rather enhance our skills to improve diagnosis, treatment planning, and patient outcomes,” Brahmbhatt says.
“I strongly believe in the power of AI to help patients, and AI is already shaping healthcare. I also believe that properly trained AI models have the power to greatly aid patient outcomes, often catching diagnoses that humans may miss due to external factors such as fatigue, stress, or inadvertent oversights,” he added.
