New Assessment Method of Alveolar Bone Grafting Using Automatic Registration and AI-based Segmentation

Area of Science:

• Craniofacial surgery
• Medical imaging analysis
• Artificial intelligence in medicine

Background:

• Secondary alveolar bone grafting is crucial for cleft lip and palate patients.
• Accurate assessment of graft volume and density is essential for treatment success.
• Existing methods for evaluating bone grafts can be time-consuming and subjective.

Purpose of the Study:

• To introduce a new assessment method for alveolar bone grafts.
• To utilize automatic registration and AI-based segmentation for graft evaluation.
• To determine the reliability of this novel assessment technique.

Main Methods:

• Computed tomography (CT) scans of 7 patients were analyzed before and after secondary alveolar bone grafting.
• Automatic rigid image registration was used to superimpose pre- and post-operative CT images.
• AI-based segmentation identified and quantified the volume and Hounsfield units (HUs) of the grafted bone.

Main Results:

• The new method demonstrated high inter-rater reliability for volume (ICC=0.95) and Hounsfield units (ICC=0.99) immediately after surgery.
• Reliability for volume (ICC=0.81) and HUs (ICC=0.57) was moderate at 6 months post-surgery.
• The technique allows for simple and rapid assessment of residual grafted bone.

Conclusions:

• The developed AI-based method provides a straightforward and efficient way to evaluate alveolar bone grafts.
• The technique shows promising results for assessing graft outcomes after secondary alveolar bone grafting.
• This method has the potential to improve post-operative monitoring in cleft care.