AI as an Ally

Dr. Ahmed Sultan on Advancing Dental Diagnostics 

What if your dentist had a second set of eyes — powered by AI — to help spot cancer earlier and improve patient outcomes? 

That’s the driving question behind the work of Ahmed Sultan, BDS, PhD, a pioneer in the use of artificial intelligence (AI) in oral health care. Sultan directs the Division of Artificial Intelligence Research at the University of Maryland School of Dentistry, the first such division at a dental school in the United States. His work is advancing a new era of diagnostics, where AI tools assist clinicians in detecting disease earlier, improving outcomes, and closing care gaps. 

“We’re seeing more administrative burden and clinician fatigue, which increases the risk of missing things,” Sultan explained. “AI can triage cases and flag abnormalities early, helping dentists and pathologists prioritize serious conditions before they’re overlooked.” 

At the heart of Sultan’s research is a curated data set that’s one of the largest of its kind in the nation, featuring cone beam CT scans and biopsy images of the head and neck cancer. Combined with partnerships across computer science, his team is building AI models that can support dentists in identifying early warning signs of oral cancer and other serious conditions — especially in locations with limited access to specialists. 

In a new video Q&A, Sultan discusses how AI is being developed responsibly to support—not replace—clinicians, and how his team has created cutting-edge tools, including an AI chatbot designed to analyze diagnostic images and guide users toward safe next steps. 

Questions

What makes the University of Maryland School of Dentistry uniquely suited to lead this kind of AI research?

Noting that the Division of AI Research at the University of Maryland School of Dentistry is the first of its kind at a dental school in the United States, Sultan also emphasized the advantage of being based at the world’s oldest dental school. “We have the one of the largest curated data sets of head and neck, cone, beam, CT radiographs and whole slide oral biopsy histopathology images,” he explained. “So the data set – plus our collaborations with technical experts in computer science and computer vision, and our multi center consortium of images from different sites – allows us to have different image formats and patient images that are diverse, so that our models can have real world generalizability in terms of their applicability. There's a huge potential to combine these with dental records to come up with personalized dentistry through AI to help patient outcomes, identify diseases early and help flag things where dentists can take a second look at to improve patient outcomes overall.”

What kind of diseases or conditions are you focusing on with AI assisted diagnostics?  

Noting that AI is good at flagging subtle abnormalities, Sultan said his focus is on conditions where dentists and oral pathologists play a key role in diagnosing the patient. “These are the early detection of precursor lesions of the head and neck in the oral maxillofacial field,” he explained. “If you detect a precursor lesion, like a white patch or a leukoplakia, there is a chance that then it could be removed or treated preventing head and neck cancer.” 

What should clinicians keep in mind as they begin to use these tools?

“The old adage that those who do not use AI will be replaced by those who do is standing the test of time. So, what you would want to do is read about the different technologies available, their limitations, read about the ADA standards for augmented intelligence, and truly understand where AI is effective in very narrow use cases and where there's limitations for it,” he said.