Comparison of automatic and visual methods used for image segmentation in Endodontics: a microCT study

Abstract

To calculate root canal volume and surface area in microCT images, an image segmentation by selecting threshold values is required, which can be determined by visual or automatic methods. Visual determination is influenced by the operator's visual acuity, while the automatic method is done entirely by computer algorithms.Objective:To compare between visual and automatic segmentation, and to determine the influence of the operator's visual acuity on the reproducibility of root canal volume and area measurements.Material and methods:Images from 31 extracted human anterior teeth were scanned with a μCT scanner. Three experienced examiners performed visual image segmentation, and threshold values were recorded. Automatic segmentation was done using the “Automatic Threshold Tool” available in the dedicated software provided by the scanner's manufacturer. Volume and area measurements were performed using the threshold values determined both visually and automatically.Results:The paired Student's t-test showed no significant difference between visual and automatic segmentation methods regarding root canal volume measurements (p=0.93) and root canal surface (p=0.79).Conclusion:Although visual and automatic segmentation methods can be used to determine the threshold and calculate root canal volume and surface, the automatic method may be the most suitable for ensuring the reproducibility of threshold determination.

Untitled

“…In a study, 31 extracted human anterior teeth were scanned by a μCT scanner. Results showed no significant difference between visual and automatic segmentation methods regarding root canal volume measurements and root canal surface (Queiroz, Rovaris, Santaella, Haiter-Neto and Freitas, 2017).…”

“…Since the knowledge of the root canal anatomy is essential in endodontics, many studies focus on a morphometric analysis of teeth for evaluating aspects such as the shaping ability of endodontic instruments, simulated root canal abnormalities, decalcification and sectioning techniques (Queiroz, Rovaris, Santaella, Haiter-Neto and Freitas, 2017). Moreover, several methods have been proposed to identify the canal anatomy, such as radiographies, diaphonization, computed tomography (CT), and more recently, micro-CT (Santa-Rosa, de Sousa-Neto, Versiani, Nevares, Xavier, Romeiro, Cassimiro, Leoni, de Menezes and Albuquerque, 2016).…”

“…The micro-CT has been considered as the gold standard for laboratory studies in endodontics (Parekh, Shah and Joshi., 2011;Queiroz, Rovaris, Santaella, Haiter-Neto and Freitas, 2017). However, studies with upper molars showed no difference between the images obtained with micro-CT and CBCT in terms of canal detection (Domark, Hatton, Benison and Hildebolt, 2013).…”

“…To characterize the efficiency of various instrumentation techniques, micro-CT based imaging studies has been used for the evaluations the following parameters: surface area change (ΔSA), volume change (ΔV), ratio of surface area change to volume change (ΔSA/ΔV), structure model index change (ΔSMI), centre of mass change (CM shift) (Juhasz., Hegedus., Marton, Benyo, Orhan., and Dobó-Nagy, 2019;Bergmans, Van Cleynenbreugel, Wevers, and Lambrechts, 2001;Hartmann, Barletta, Camargo Fontanella and Vanni, 2007;Peters, Peters, Schonenberger and Barbakow, 2003;Queiroz, Rovaris, Santaella, Haiter-Neto and Freitas, 2017).…”

Root canal segmentation in cone-beam computed tomography

“…Such findings may be attributed to the high resolution and low noise produced by micro-CT, what makes identification of dentin borders accurate. In fact, it has been shown that micro-CT has a unique potential of showing detailed root canal morphological features in an accurate manner, without destruction of the tooth, while offering reproducible data in three-dimension 10,12,18 . Regarding the segmentation performed on CBCT images, the results showed that, when compared to the gold standard, both the "automatic CBCT" and "visual CBCT" were statistically different compared to the gold standard, revealing a limitation of the accurate determination of root canal volume using this image modality (with acquisition parameters used in this study).…”