UTHealth Houston schools team up to create AI system to prevent wrong-tooth extractions

Simon Young, DDS, MD, PhD, of UTHealth Houston School of Dentistry (second from right) is creating an AI system to prevent wrong-tooth extractions. (Photo by UTHealth Houston)

Simon Young, DDS, MD, PhD, professor and acting chair of the Bernard and Gloria Pepper Katz Department of Oral and Maxillofacial Surgery at UTHealth Houston School of Dentistry, has received a $75,000 research support grant from the Oral and Maxillofacial Surgery Foundation to develop an artificial intelligence system to prevent wrong-tooth extractions.

The one-year grant will support a multidisciplinary project designed to identify high-risk situations before a procedure begins, providing clinicians with an additional safety layer.

Young is leading the School of Dentistry team as co-principal investigator in partnership with faculty from the McWilliams School of Biomedical Informatics at UTHealth Houston, where Sayali Tungare, BDS, MPH, PhD, serves as principal investigator and leads the project’s artificial intelligence development. Co-principal investigator Ming Huang, PhD, also contributes AI expertise, while Muhammad Walji, PhD, Chair in Dental Informatics who also holds the D. Bradley McWilliams Professorship, provides leadership in dental informatics. The effort represents a broader collaboration between the School of Dentistry and the McWilliams School of Biomedical Informatics.

Addressing a persistent clinical challenge

Wrong-site surgery, including removal of the incorrect tooth, is considered a “never event” — an error that should not occur but still happens occasionally despite established safeguards.

Young said the idea for the project emerged during a school retreat focused on emerging technologies and AI.

Across dentistry, certain clinical scenarios are known to increase the risk of extracting the wrong tooth, particularly when multiple teeth are impacted or orthodontic plans are complex.

“There are well-known situations where mistakes can happen,” said Young, who holds the Young OMS Research Professorship. “The goal is to identify those higher-risk cases before treatment begins.”

The proposed system would analyze patient records and imaging data to flag potential concerns, similar to how electronic health records alert providers to allergies or medication interactions.

“Preventing wrong-tooth extraction requires more than checklists alone,” Walji said. “By combining clinical data, imaging, and workflow information, we can build intelligent systems that identify risk before an irreversible error occurs. It would function as a warning system, prompting clinicians to double-check when a situation carries elevated risk.”

Harnessing data and collaboration

Because wrong-tooth extractions are relatively rare, developing an effective predictive model requires sophisticated methods to train the AI system.

The research team plans to use existing clinical data to identify patterns associated with past errors, then generate simulated cases to expand the dataset for training and validation.

“You have to train AI on enough information,” Young said. “Even if real cases are limited, you can use synthetic scenarios to help the system learn what high-risk situations look like.”

Young emphasized that the project depends on collaboration across disciplines, bringing together clinicians, data scientists, and engineers.

“You need people with different expertise — clinicians who understand the problem and experts who can build the technology,” he said.

UTHealth Houston’s integrated environment, including access to large dental datasets and artificial intelligence specialists, made the project feasible.

“Artificial intelligence is particularly well suited to rare but high-impact events,” Tungare said. “By learning patterns across imaging, documentation, and clinical context, the multimodal approach can flag situations that may otherwise be difficult to recognize during routine care. If we can identify high-risk cases early, we can prevent harm rather than respond to it.”

Improving patient safety nationwide

If successful, the system could eventually be implemented in electronic dental records and clinical workflows to help prevent errors before they occur.

Young noted that even rare mistakes can have serious consequences, making prevention efforts critical.

“It’s another layer of protection,” he said. “Something that helps ensure the right procedure is performed on the right patient every time.”

The OMS Foundation Research Support Grant is intended to generate preliminary data to support larger external funding applications and future implementation studies. Additional funding has expanded the project’s scope, bringing total support to $175,000, including $100,000 from the UT REAL Health AI Pilot Program.

The UT REAL Health AI Pilot Program supports collaborative AI initiatives across institutions in The University of Texas System. Led by principal investigator Tungare, the team includes Young, co-principal investigator; Huang and Walji, co-investigators; Yojana Vadnerkar, graduate research assistant; and Joshua Stone, DDS, MD, a collaborator at UT Health San Antonio.

The 21-month study began Jan. 1, 2026, and will continue through Sept. 30, 2027.

Young said the project illustrates how innovative ideas can grow from institutional collaboration and targeted seed funding.

“This started as a simple idea during a discussion about AI,” he said. “Now, we have the opportunity to develop a tool that could improve surgical safety on a much broader scale.”