AI-Based Extraction Difficulty Model Aids Surgical Planning for third molar extraction: Study

A new study published in the Annals of Biomedical Engineering developed a scoring model that helps dentists assess the difficulty and risk of third molar (3M) extraction, which supports in better surgical planning. Further research with more diverse data and enhanced artificial intelligence (AI) performance is needed for effective clinical implementation.

Wisdom tooth removal is one of the most common oral surgeries, yet the difficulty can vary significantly depending on factors like tooth position, depth of impaction, and proximity to critical anatomical structures like nerves. Traditionally, dentists rely on experience and visual interpretation of X-rays to estimate the difficulty of surgery. However, this process can be subjective and may not fully capture subtle risk factors.

Thus, this research have now developed a novel scoring system that leverages AI-based technologies to analyze panoramic dental X-rays. The system focuses on 3 key anatomical components, namely the mandibular canal (which houses the inferior alveolar nerve), the inner area of alveolar bone, and the third molar itself. Using a deep learning model, this study successfully detected and segmented these structures across patients from 16 to 86 years old.

The system assigns scores based on 3 measurable criteria such as the angle or inclination of the wisdom tooth, the depth at which the tooth is impacted in the jawbone, and how close the tooth is to the mandibular canal. Each factor is scored individually, and the combined total places the extraction into one of the categories as very easy, easy, slightly difficult, or very difficult.

The detection accuracy for upper and lower third molars reached impressive precision levels of 0.93 and 0.97, respectively. Meanwhile, the algorithms used to assess inclination, impaction depth, and nerve proximity achieved accuracies of approximately 85%, 95%, and 90%. These figures suggest that AI can reliably interpret dental imaging and assist in clinical decision-making.

Overall, this approach could significantly improve patient outcomes by offering a standardized and data-driven difficulty score, where dentists can better anticipate surgical challenges, minimize complications, and reduce the risk of nerve damage. This system may also enhance communication between dentists and patients. The patients can be better informed about the procedure, recovery expectations, and potential complications with clear risk assessments.

Source:

Ko, J., Sooksatra, S., Kim, S., Tang, S., Ha, J.-E., Han, D.-H., Lee, K. H., Jung, Y.-J., & Kim, M.-J. (2026). Quantitative assessment of third molar extraction difficulty and nerve injury risk using artificial intelligence and image processing. Annals of Biomedical Engineering. https://doi.org/10.1007/s10439-026-04114-9