Exploring the Potential of the PerioAI System to Support Periodontal Clinical Decision Making: A Proof-of-Principle...

Area of Science:

• Periodontology
• Artificial Intelligence in Dentistry
• Radiographic Imaging Analysis

Background:

• Periodontal disease diagnosis often relies on radiographic imaging.
• Accurate measurement of periodontal parameters is crucial for effective treatment planning.
• Integrating AI with dental imaging offers potential for improved diagnostic capabilities.

Purpose of the Study:

• To evaluate PerioAI's ability to measure gingival margin-to-bone distance (GBD) and derive AI-probing depth (AI-PD).
• To assess if AI-PD provides additional value for periodontal clinical decision-making using radiographic data.
• To determine the impact of AI-PD on prognosis and treatment planning accuracy.

Main Methods:

• A cross-sectional study involving 53 periodontitis patients (1298 teeth).
• PerioAI integrated intraoral scans and cone-beam CT to calculate GBD, converted to AI-PD.
• Clinical decisions (prognosis, treatment planning) were compared across three conditions: OPG+chart, OPG-only, and OPG+AI-PD.

Main Results:

• The OPG+AI-PD condition showed significantly higher agreement with reference clinical decisions compared to OPG-only.
• Patient-level agreement rates for prognosis and treatment planning increased from ~78% to ~84% (p<0.05).
• AI-PD integration reduced overtreatment risk by 42.3% and tooth extraction risk by 98.5%.

Conclusions:

• PerioAI, when combined with radiographic information, offers valuable incremental data for periodontal clinical decision-making.
• AI-derived probing depth shows promise in improving diagnostic accuracy and treatment planning.
• Further validation in larger, diverse populations is recommended to integrate additional periodontal parameters.