AI in Orthodontics, Where Are We And Where Are We Going 10 MINUTE SUMMARY

AI in Orthodontics, Where Are We And Where Are We Going 10 MINUTE SUMMARY

Click here for youtube video

Join me for a podcast summary looking at Ai in orthodontics and its clinical application. A growing topic in orthodontics, and one of the most featured topics at this years AAO. This summary is based on 3 lectures from this year’s summer meeting by Juan Francisco Gonzalez & Jean Marc Retrouvey, Tarek ElShebiny , Jonas Bianchi and Lucia Cevidanes. We will look what Ai is, the way it works and its clinical application, as well as a critical view on this young field.

What is Ai:

1. Technology that enables computers and machines to simulate human intelligence, perform 1 task very well, e.g. voice command, Youtube recommendations
2. Predictive modelling, makes calculations, convert information into numbers or categories and recognise patterns

Levels of Ai: Machine learning, Neural Networks and Deep Learning

1. Machine learning
   1. The ability for a machine to learn from data and past experience to identify patterns and make predictions

2. Neural Networks
   1. Specific model which relies on interconnected nodes, which perform a mathematical calculation of associations , patterns, and probabilities

3. Deep learning
   1. Is a complex version of neural networks

Virtual patient

• CBCT segment + STL file – segmentation of the teeth and roots, with labelling of different stuctures
  • Can print model, visualise ideal vector and calculate ideal vector
  • However clinician still required to establish biomechanics

• CBCT integration for aligner cases, Unpublished thesis Khalid Alotaibi:
  • Treatment planning confidence increased 50%, least change was treatment planning modification

 Diagnostic data:

• Ai cephalometric tracing
  • 46% of 24 landmarks 2.0mm within
  • 4 different programmes Iortho, Webceph, Orthodc, cephx
  • All landmarks had good overall agreement but variation in identification

• Facial Analysis
• Automated 3D facial asymmetry analysis using machine learning Adel 2025
  • Study – 7 landmarks
  • Identified manually and with deep learning
  • 5 accurate, 2 significant difference but not clinically relevant

Diagnostic accuracy of photos

• Clinical photos assessment by Ai, and compared to clinical examination
• Sensitivity 72%, specificity 54% Vaughan & Ahmed 2025

Growth prediction

• Poor agreement age 9

Comparison between direct, virtual and AI bonding

• DIBs – uses Ai for bonding
• Compare Ai Vs user modified indirect bonding Vs direct bonding (gold standard), 0.5mm significant
• Incisors accurate
• Premolars and lower laterals inaccurate

Monitoring

Previous podcast exploring the accuracy of remote monitoring

• with Ferlito 2022 80% repeatability from 2 scans 7% repeatability and reproducibility 

Bracket removal from scan and retainer fit

Tarek Assessment of virtual bracket removal by artificial intelligence and thermoplastic retainer fit AJODO 2024

• Retainers for both – clinically acceptable

FDA approval of Ai in dentistry

• FDA – Software of Medical Diagnosis
  • 4 dental:
    • Dental Monitoring
    • Ray Co
    • X-Nav technologies
    • Densply Sirona

What’s next

• More data learning to train AI model
• Robotics customising appliances per patient