PerioAI: A digital system for periodontal disease diagnosis from an intra-oral scan and cone-beam CT image

Abstract

Periodontal disease diagnosis and treatment planning are critical for preventing bone and tooth loss. Clinically, dentists manually measure periodontal pocket depth with probes while integrating bone structure from imaging to assess periodontal status, a process that is subjective, invasive, and cognitively burdensome. Here, we propose PerioAI, an accurate, automatic, and non-invasive system that directly measures the gingiva-bone distance (GBD) and provides soft and hard tissue information digitally. PerioAI is a full-stack process comprising four key components: intra-oral scan (IOS) segmentation, cone-beam computed tomography (CBCT) image segmentation, multimodal data fusion, and digital probing measurement. We evaluated PerioAI on multicenter cohorts comprising 2,507 patients. Outstanding IOS and CBCT segmentation performances ensure accuracy throughout the full-stack process. Moreover, digital probing achieves remarkable precision with only 0.040mm error. This approach has the potential to substantially improve clinical workflows in periodontal disease management, offering a more precise, patient-friendly method for diagnosis and treatment decision-making.