Mandible Segmentation of Dental CBCT Scans Affected by Metal Artifacts Using Coarse-to-Fine Learning Model.

Bingjiang Qiu,Max Johannes Hendrikus Witjes,Jiapan Guo,Peter M A Van Ooijen,Ronald J H Borra,Hylke Van Der Wel,Joep Kraeima,Haye Hendrik Glas

doi:10.3390/jpm11060560

Bingjiang Qiu, Max Johannes Hendrikus Witjes + Show 6 more

Open Access

https://doi.org/10.3390/jpm11060560

Copy DOI

Abstract

Accurate segmentation of the mandible from cone-beam computed tomography (CBCT) scans is an important step for building a personalized 3D digital mandible model for maxillofacial surgery and orthodontic treatment planning because of the low radiation dose and short scanning duration. CBCT images, however, exhibit lower contrast and higher levels of noise and artifacts due to extremely low radiation in comparison with the conventional computed tomography (CT), which makes automatic mandible segmentation from CBCT data challenging. In this work, we propose a novel coarse-to-fine segmentation framework based on 3D convolutional neural network and recurrent SegUnet for mandible segmentation in CBCT scans. Specifically, the mandible segmentation is decomposed into two stages: localization of the mandible-like region by rough segmentation and further accurate segmentation of the mandible details. The method was evaluated using a dental CBCT dataset. In addition, we evaluated the proposed method and compared it with state-of-the-art methods in two CT datasets. The experiments indicate that the proposed algorithm can provide more accurate and robust segmentation results for different imaging techniques in comparison with the state-of-the-art models with respect to these three datasets.

Highlights

Three-dimensional (3D) virtual surgical planning (VSP) technique is commonly used for orthodontic diagnosis, orthognathic diagnosis and surgery planning because it allows for pre- or post-operative simulation of surgical options [1]
We aim to develop an accurate mandible segmentation algorithm to overcome the inaccurate prediction for 3D virtual surgical planning (3D VSP) in cone-beam computed tomography (CBCT)
We propose a hybrid method which consists of a coarse stage and fine stage, in which the coarse stage makes use of 3D convolutional neural networks (CNN) for predicting the mandible-like organ and the fine stage utilizes the recurrent segmentation CNN for fine mandible segmentation in CBCT images which are mostly affected by metal artifacts

Summary

Introduction

Three-dimensional (3D) virtual surgical planning (VSP) technique is commonly used for orthodontic diagnosis, orthognathic diagnosis and surgery planning because it allows for pre- or post-operative simulation of surgical options [1]. 3D mandible surface models in 3D VSP are created and superimposed to demonstrate the orthodontic changes both visually and quantitatively (including pre- and post- operation). Teeth, tooth fillings, and dental braces in orthodontic treatment and metal implants in orthognathic treatment are high attenuation materials which cause high noise and low contrast in visual impressions of CBCT images. It is challenging for orthodontic or orthognathic VSP to accurately perform mandibular segmentation in CBCT. The patient-specific orthodontic or orthognathic treatment planning is restricted and delayed by this time-consuming procedure

Objectives

Methods

Results

Discussion

Conclusion