# Accuracy assessment of three-dimensional abutment tooth data construction using swept-source optical coherence tomography

## TL;DR

This study evaluates the accuracy of using swept-source optical coherence tomography (SS-OCT) for creating 3D dental models, finding it less precise than current intraoral scanners.

## Contribution

The paper introduces a proof-of-principle application of SS-OCT in dental imaging and identifies specific limitations and improvements needed.

## Key findings

- TRIOS3 intraoral scanner showed superior accuracy compared to SS-OCT.
- SS-OCT accuracy decreases with depth and lacks automatic margin correction in CAD software.
- Improvements in CAD compatibility and specialized probes are needed for SS-OCT to be effective.

## Abstract

A limitation of widely used intraoral scanners (IOSs) is their inability to capture finish lines at the subgingival marginal area, as they only extract surface information. Swept-source optical coherence tomography (SS-OCT) captures high-speed, high-resolution cross-sectional images of soft and hard tissues. Integrating this technology can overcome clinical IOS limitations. Therefore, this study was conducted to fabricate crowns from three-dimensional images scanned with SS-OCT as a proof-of-principle for its application in IOSs and to evaluate fit accuracy. TRIOS3 was used for comparison, with both SS-OCT and TRIOS3 scanned three times, and crowns were fabricated using the same digital workflow. Internal gaps were measured using scanning electron microscopy, and marginal fit was evaluated via microscopy. Results showed that TRIOS3 had superior accuracy. SS-OCT can image solely in the occlusal direction, with accuracy decreasing at greater depths, which reduces precision around the margin. Additionally, SS-OCT lacks automatic correction of surface information in computer-aided design (CAD) software. To improve SS-OCT accuracy for abutment tooth measurements, automatic margin correction, improved CAD compatibility, and specialized probes for capturing tooth features are needed.