Department of Dentistry, Graduate School, Kyung Hee University, Seoul, South Korea.
Department of Convergence Medicine, Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea.
Am J Orthod Dentofacial Orthop. 2022 Aug;162(2):e53-e62. doi: 10.1016/j.ajodo.2022.01.011. Epub 2022 May 31.
This study aimed to evaluate a 3-dimensional (3D) U-Net-based convolutional neural networks model for the fully automatic segmentation of regional pharyngeal volume of interests (VOIs) in cone-beam computed tomography scans to compare the accuracy of the model performance across different skeletal patterns presenting with various pharyngeal dimensions.
Two-hundred sixteen cone-beam computed tomography scans of adult patients were randomly divided into training (n = 100), validation (n = 16), and test (n = 100) datasets. We trained the 3D U-Net model for fully automatic segmentation of pharyngeal VOIs and their measurements: nasopharyngeal, velopharyngeal, glossopharyngeal, and hypopharyngeal sections as well as total pharyngeal airway space (PAS). The test datasets were subdivided according to the sagittal and vertical skeletal patterns. The segmentation performance was assessed by dice similarity coefficient, volumetric similarity, precision, and recall values, compared with the ground truth created by 1 expert's manual processing using semiautomatic software.
The proposed model achieved highly accurate performance, showing a mean dice similarity coefficient of 0.928 ± 0.023, the volumetric similarity of 0.928 ± 0.023, precision of 0.925 ± 0.030, and recall of 0.921 ± 0.029 for total PAS segmentation. The performance showed region-specific differences, revealing lower accuracy in the glossopharyngeal and hypopharyngeal sections than in the upper sections (P <0.001). However, the accuracy of model performance at each pharyngeal VOI showed no significant difference according to sagittal or vertical skeletal patterns.
The 3D-convolutional neural network performance for region-specific PAS analysis is promising to substitute for laborious and time-consuming manual analysis in every skeletal and pharyngeal pattern.
本研究旨在评估一种基于三维（3D）U-Net 的卷积神经网络模型，用于在锥形束计算机断层扫描中对区域性咽兴趣区（VOI）进行全自动分割，以比较模型在呈现不同咽尺寸的不同骨骼模式下的性能准确性。
将 216 例成人锥形束计算机断层扫描随机分为训练（n=100）、验证（n=16）和测试（n=100）数据集。我们使用 3D U-Net 模型对咽 VOI 及其测量值（鼻咽、腭咽、舌咽和下咽部分以及总咽气道空间（PAS））进行全自动分割。根据矢状和垂直骨骼模式对测试数据集进行细分。使用半自动软件由 1 名专家手动处理创建的真实值来评估分割性能，通过 Dice 相似系数、体积相似性、精度和召回值进行评估。
所提出的模型表现出高度准确的性能，总 PAS 分割的平均 Dice 相似系数为 0.928±0.023，体积相似性为 0.928±0.023，精度为 0.925±0.030，召回率为 0.921±0.029。性能表现出区域特异性差异，在上部区域比在舌咽和下咽区域表现出较低的准确性（P<0.001）。然而，根据矢状或垂直骨骼模式，每个咽 VOI 的模型性能准确性没有显著差异。
3D 卷积神经网络在特定区域的 PAS 分析中的性能有望替代费力且耗时的手动分析，适用于各种骨骼和咽模式。
