A study to compare the accuracy of volume measurement between a commercially available 3D facial imaging system (di4d snap) and an “app” based system (bellus3d)

Desai, Rajeev Ramarao (2022). A study to compare the accuracy of volume measurement between a commercially available 3D facial imaging system (di4d snap) and an “app” based system (bellus3d). University of Birmingham. M.Sc.

Preview

Desai2023MScbyRes_Redacted.pdf
Text - Redacted Version
Available under License All rights reserved.
Download (6MB) | Preview

Abstract

AIM: The aim of the study is to compare the accuracy of volume measurement between a commercially available 3D facial imaging system (Di4D SNAP) and an “App” based system (Bellus3D). STUDY DESIGN: Single centre, prospective trial.

MATERIALS & METHODS: Thirty-three volunteers comprising of staff and students at Birmingham Dental Hospital and School were recruited. Volunteers had to meet the inclusion and exclusion criteria, no recent dental or medical interventions and no known craniofacial conditions. Baseline 3D facial images at rest were taken using Di4D SNAP and Bellus3D. 2cm3 polyvinylsiloxane putty simulated volume explants were added to the chin, upper lip and right and left paranasal regions and the volunteers re-imaged using Di4D SNAP and Bellus3D. Following facial image capture, the simulated volume images were superimposed on the baseline images and the volume difference in the chin, lip, right and left paranasal regions were measured, using the Tetrahedral Construction Method (TCM), and compared to the gold standard known volume of the putty (2cm3). In addition, the accuracy of using TCM and a Coons patch, to measure volume, was assessed using only the Di4D SNAP facial images.

RESULTS: Overall, the mean percentage error for Bellus3D was an over estimation of 10.4% (95% CI 2.5 to 18.2%). Overall, the mean percentage error for Di4D SNAP was an underestimation of 1.6% (95% CI -4.4 to 1.1%). The median volume difference between Bellus3D and Di4D SNAP was statistically significantly different (p<0.05) in all areas except the left paranasal region. When comparing the Coons patch and TCM for volume measurement, overall as a percentage, the Coons patch overestimated the mean volume by 9.4% (95% CI 5.0 to 13.6%). Whilst the TCM algorithm underestimated the mean volume by 1.6% (95% CI -4.4 to 1.1%).

CONCLUSIONS: Across the 4 regions investigated there was close to a 10% error in mean volume measurement using Bellus3D and a 2% error using Di4D SNAP. This was a statistically significant difference (p<0.05) based on the actual volume differences. The results of the present study would suggest that Bellus3D captures simulated volume changes of the left side of the face with greater accuracy than the right side of the face. Based on the Di4D SNAP images, across the 4 regions investigated there was close to a 10% error in mean volume measurement using Coons patch and a 2% error using TCM.

Type of Work:

Thesis (Masters by Research > M.Sc.)

Award Type:

Masters by Research > M.Sc.

Supervisor(s):