Surgery

Sept. 22, 2026

While artificial intelligence (AI) and machine learning (ML) are advancing rapidly at Mayo Clinic in Rochester, Minnesota, their adoption in oral and maxillofacial surgery (OMS) remains in its early stages. Oral and maxillofacial surgeons John M. Nathan, M.D., D.D.S., and Kyle S. Ettinger, M.D., D.D.S., use AI-ML to streamline image analysis, surgical planning and clinical documentation. Rather than replacing physician expertise, the technology improves efficiency while supporting clinical decision-making.

Dr. Nathan, Dr. Ettinger and other Mayo Clinic researchers are helping advance new AI-ML tools. But widespread adoption will require collaboration across multiple institutions to ensure they are safe, clinically useful and reliable across real-world surgical settings.

Why surgeons must help shape AI-ML

Before AI-ML tools can be adopted, they must be carefully developed, tested and validated. "It's a methodical process, similar to how we scrutinize and adopt any innovation," Dr. Nathan says. "The difference is that AI technology is not widely understood among clinicians."

In a recent Journal of Oral and Maxillofacial Surgery article, Dr. Nathan and Dr. Ettinger review emerging AI-ML applications in OMS and emphasize the role surgeons play in their development.

"Oral and maxillofacial surgeons need an understanding of AI to help validate the models and shape how they're implemented responsibly and optimally," Dr. Ettinger says. That understanding helps surgeons recognize when the outputs are useful, when they need correction and when they should not influence care.

Current applications of AI-ML in OMS

AI-ML is already helping oral and maxillofacial surgeons work more efficiently in defined parts of care delivery, including:

Automated image segmentation

Image segmentation involves identifying and outlining anatomical structures, such as the mandible, on CT images by assigning each pixel to a specific structure. It is time-consuming but essential for diagnosis and virtual surgical planning. New AI-ML-powered segmentation tools are reducing processing time, enabling faster diagnosis and surgical planning for patients.

Specialists at Mayo Clinic use these tools to generate the segmentation. Then they review and refine the result before it is used. In one study reviewed by the authors, an AI-ML model achieved 97.4% agreement with expert mandibular segmentations while completing the task in just seven seconds.

"This efficiency is essential, especially for cases involving cancer or trauma, where care needs to be performed expeditiously," Dr. Ettinger says.

Landmark identification

Oral and maxillofacial surgeons are using commercially available AI-ML software in jaw surgery to automatically identify anatomical landmarks on X-ray and CT images and calculate cephalometric measurements that inform diagnosis, occlusal analysis and operative planning. "Historically, we placed each point and performed the analyses manually," Dr. Ettinger says. "By automating these tasks, we can spend more time developing the most appropriate treatment plan for each patient."

Research reviewed by the authors has demonstrated clinically meaningful accuracy for automated landmark detection. Even so, Dr. Nathan says surgeons must continue to verify the landmarks and measurements before incorporating them into the surgical plan.

Ambient listening

Ambient listening is one of the most widely used generative AI applications in healthcare. The technology, which is based on large language models, helps physicians focus more on patients during visits and significantly reduces documentation time. However, physicians must still review every note for accuracy, completeness and clinical nuance before it becomes part of the patient's record.

Large language models may eventually play a role in patient education and postoperative communication. "These tools still require physician oversight because they can generate inaccurate or misleading information," Dr. Nathan says.

Emerging AI-ML applications in OMS

AI-ML development in OMS is focused on automating individual steps within clinical workflows while surgeons remain responsible for output. Instead of creating a fully autonomous surgical planning system, researchers are creating tools that can simplify specific tasks.

Promising applications include:

• Virtual surgical planning. In addition to identifying landmarks and segmenting images, AI-ML may eventually automate other planning steps, such as identifying pathology, determining resection margins and generating preliminary surgical plans. It also could help engineers design patient-specific implants more efficiently by automating portions of the design process.
• Predictive models. Researchers are developing AI-ML tools that may one day help estimate cancer recurrence, treatment response and surgical risk. These applications remain investigational because they require large, diverse datasets and extensive validation.
• Augmented reality in the operating room. Researchers are exploring systems that would align CT images with the patient's anatomy in real time, providing enhanced guidance during procedures. This technology is experimental because maintaining alignment in the dynamic surgical field during OMS is technically challenging.

Responsible adoption of AI-ML requires rigorous validation

Although studies of AI-ML in oral and maxillofacial surgery have produced encouraging results, many are proof-of-concept studies that require additional clinical validation. Like any medical innovation, AI-ML must perform reliably across patient populations, imaging protocols and practice settings and improve care delivery beyond a research setting.

"Software engineers can develop a great model, but without clinician input, it may not be implemented in the most meaningful way to improve patient care," Dr. Ettinger says.

At Mayo Clinic, oral and maxillofacial surgeons collaborate with engineers and industry partners to bring new AI-ML applications into clinical use. Dr. Nathan is also uniquely trained with a master's degree in artificial intelligence and healthcare. That combination of technical knowledge and clinical experience helps the team identify applications with real clinical potential.

"A human using AI will outperform a human alone," Dr. Nathan says. The future of AI-ML in OMS depends on thoughtful collaboration among clinicians, engineers and researchers to adopt tools that strengthen clinical judgment, improve efficiency and preserve surgeon accountability.

For more information

Nathan JM, et al. Artificial intelligence for the oral and maxillofacial surgeon. Journal of Oral and Maxillofacial Surgery. 2026;84:608.

Refer a patient to Mayo Clinic.