Related Experiment Video
Updated: Jan 13, 2026
07:32
Author Spotlight: 3D Movement Assessment of Maxillary Posterior Teeth in Clear Aligner Treatment
Published on: February 23, 2024
1.8K
Accuracy of dentalmonitoring's artificial intelligence in detecting aligner tracking issues: a retrospective
Julie Fahl McCray1, William Dabney2, Dylan Handlin1
1Department of orthodontics, Center of Advanced Dental Education, Saint-Louis University, 3320 Rutger St., St. Louis, MO, 63104, USA.
BMC Oral Health
|January 8, 2026
Summary
DentalMonitoring's (DM) artificial intelligence (AI) accurately detects aligner tracking issues, showing high reliability in identifying unseat events. The AI system demonstrated strong performance, especially in ruling out significant aligner misfits.
Area of Science:
- Orthodontics
- Artificial Intelligence in Healthcare
- Dental Technology
Background:
- Aligner tracking issues are common in orthodontic treatment.
- Accurate detection of aligner tracking problems is crucial for treatment success.
- DentalMonitoring (DM) has developed an AI system to address these issues.
Purpose of the Study:
- To evaluate the performance of DentalMonitoring's AI in detecting aligner tracking issues.
- To compare AI-driven assessments with expert orthodontic evaluations.
- To quantify the accuracy of the AI in identifying different severities of aligner unseating.
Main Methods:
- A multicenter retrospective study analyzed 3,323 patient assessments.
- DM's AI was evaluated using binary (seated/unseated) and three-level (seated, slight unseat, noticeable unseat) models.
- AI outputs were compared against a reference standard from orthodontic residents; sensitivity, specificity, PPV, and NPV were calculated.
Main Results:
- For binary classification, AI achieved 93.2% sensitivity and 86.2% specificity.
- In the three-level model, noticeable unseats had 91.1% sensitivity and 90.5% specificity.
- High negative predictive values (NPV) indicated reliability in ruling out significant unseat events.
Conclusions:
- DM's AI system shows high sensitivity and NPV for detecting aligner unseat events.
- The AI reliably differentiates between slight and noticeable unseats.
- The AI performed reliably, minimizing false negatives for clinically significant misfits, warranting further validation.
