Sub-Second Optical Coherence Tomography Angiography Protocol for Intraoral Imaging Using an Efficient...

Area of Science:

• Biomedical Imaging
• Optical Coherence Tomography
• Artificial Intelligence in Medicine

Background:

• Intraoral imaging for disease diagnosis requires high-resolution visualization of microvasculature.
• Traditional optical coherence tomography angiography (OCTA) methods can be time-consuming, leading to motion artifacts.
• Deep learning approaches show promise for enhancing OCTA image quality and reducing acquisition times.

Purpose of the Study:

• To develop a fast OCTA protocol for intraoral imaging using a novel deep learning model.
• To reduce image acquisition time and minimize motion artifacts in intraoral OCTA scans.
• To evaluate the performance of the proposed model in reconstructing high-resolution intraoral OCTA images.

Main Methods:

• Implementation of a 200 kHz swept-source OCT system for intraoral imaging.
• Development of the Intraoral Micro-Angiography Super-Resolution Transformer (IMAST) model for image reconstruction.
• Acquisition of intraoral OCTA data with reduced spatial sampling resolution (~0.3s acquisition time).

Main Results:

• The IMAST model successfully reconstructed high-resolution intraoral OCTA images from reduced-resolution scans.
• Experimental results demonstrated robust performance and significant image quality enhancement compared to existing methods.
• IMAST exhibited lower model complexity, faster inference time, and reduced computational cost.

Conclusions:

• The developed fast OCTA protocol with the IMAST model is suitable for clinical intraoral imaging.
• This approach offers potential for noninvasive oral disease diagnosis and early detection of malignancies.
• The protocol can reduce patient discomfort and improve the efficiency of oral assessments.