Segmentation, accuracy validation of mandibular molar, pulp cavity on cone-beam CT images by U-net neural network

doi:10.19439/j.sjos.2022.05.002

Abstract

Abstract: PURPOSE: To realize the automatic segmentation of mandibular molar and pulp cavity on cone-beam CT (CBCT) images by U-net convolutional neural network, and to use the 3D models reconstructed by Micro-CT data as the ground truth to validate its accuracy. METHODS: Twenty groups of small field of view(FOV) CBCT data containing complete mandibular molars were collected from the Department of Radiology, Affiliated Stomatology Hospital of Tongji University. After preprocessing, an endodontic specialist labeled teeth and pulp cavities by MITK Workbench software. These data were used as the training set for training U-net neural network. In addition, five mandibular molars and corresponding small FOV CBCT data were collected. These five CBCT were processed in the same way and used as the testing set. Then, teeth and pulp cavities on CBCT images of the testing set were segmented and reconstructed by U-net neural network and the same specialist. The isolated teeth were scanned by a Micro-CT machine after preprocessing and the results were reconstructed to 3D models, which were used as the ground truth. Then the 3D models reconstructed by the specialist’s labeling, U-net network segmentation results, and the ground truth in the testing set were compared. Dice similarity coefficient(DSC), average symmetric surface distance (ASSD), Hausdorff distance (HD), and morphological analysis were used to evaluate the results. SPSS 20.0 software package was used for statistical analysis. RESULTS: Compared with the ground truth, the segmentation accuracy of the U-net neural network measured by DSC, ASSD, and AHD was (95.30±1.01)%, (0.11±0.02) mm, and (1.05±0.31) mm in teeth and (81.21±2.27)%, (0.15±0.05) mm, and (3.29±1.85) mm in the pulp cavity, respectively. Morphological analysis results showed that the U-net network segmentation results were similar to the ground truth in tooth and pulp chamber. As for the segmentation results of root canals, only thick root canals could be segmented rather than the thin root canals, such as the canals in the apical third and lateral root canals. CONCLUSIONS: Under the experimental conditions, the U-net neural network trained by the specialist’s labeling realized the automatic and accurate segmentation of mandibular molar and their pulp chamber on CBCT images. For the segmentation of root canals, the results need to be further improved.

Key words: U-net convolutional neural network, Cone-beam CT, Micro-CT, Image segmentation, Mandibular molars

CLC Number:

R781.3

LIN Xiang, FU Yu-jie, REN Gen-qiang, WEN Jia-huan, CHEN Yu-fei, ZHANG Qi. Segmentation, accuracy validation of mandibular molar, pulp cavity on cone-beam CT images by U-net neural network[J]. Shanghai Journal of Stomatology, 2022, 31(5): 454-459.

References

[1] Pauwels R, Araki K, Siewerdsen J, et al.Technical aspects of dental CBCT: state of the art[J]. Dentomaxillofac Radiol, 2015, 44(1): 20140224.
[2] No authors listed. AAE and AAOMR joint position statement: use of cone-beam computed tomography in endodontics 2015 update[J]. J Endod, 2015, 41(9): 1393-1396.
[3] Kruse C, Spin-Neto R, Wenzel A, et al.Cone-beam computed tomography and periapical lesions: a systematic review analysing studies on diagnostic efficacy by a hierarchical model[J]. Int Endod J, 2015, 48(9): 815-828.
[4] Kfir A, Telishevsky-Strauss Y, Leitner A, et al.The diagnosis and conservative treatment of a complex type 3 dens invaginatus using cone beam computed tomography (CBCT) and 3D plastic models[J]. Int Endod J, 2013, 46(3): 275-288.
[5] Byun C, Kim C, Cho S, et al.Endodontic treatment of an anomalous anterior tooth with the aid of a 3-dimensional printed physical tooth model[J]. J Endod, 2015, 41(6): 961-965.
[6] Torres A, Lerut K, Lambrechts P, et al.Guided endodontics: use of a sleeveless guide system on an upper premolar with pulp canal obliteration and apical periodontitis[J]. J Endod, 2021, 47(1): 133-139.
[7] Harris M, Qi A, Jeagal L, et al.A systematic review of the diagnostic accuracy of artificial intelligence-based computer programs to analyze chest X-rays for pulmonary tuberculosis[J]. PLoS One, 2019, 14(9): e0221339.
[8] Litjens G, Kooi T, Bejnordi BE, et al.A survey on deep learning in medical image analysis[J]. Med Image Anal, 2017, 42: 60-88.
[9] Cui Z, Li C, Wang W.ToothNet: automatic tooth instance segmentation and identification from cone-beam CT images[C]. Proc IEEE/CVF Conf Comput Vis Pattern Recognit (CVPR), 2019: 6361-6370.
[10] Li Q, Chen K, Han L, et al.Automatic tooth roots segmentation of cone beam computed tomography image sequences using U-net and RNN[J]. J Xray Sci Technol, 2020, 28(5): 905-922.
[11] Zheng Q, Ge Z, Du H, et al.Age estimation based on 3D pulp chamber segmentation of first molars from cone-beam-computed tomography by integrated deep learning and level set[J]. Int J Legal Med, 2021, 135(1): 365-373.
[12] Duan W, Chen Y, Zhang Q, et al.Refined tooth and pulp segmentation using U-Net in CBCT image[J]. Dentomaxillofac Radiol, 2021: 20200251.
[13] Aksoy U, Küçük M, Versiani MA, et al. Publication trends in Micro-CT endodontic research: a bibliometric analysis over a 25-year period[J]. Int Endod J, 2021, 54(3): 343-353.
[14] Wolf TG, Stiebritz M, Boemke N, et al.3-dimensional analysis and literature review of the root canal morphology and physiological foramen geometry of 125 mandibular incisors by means of Micro-computed tomography in a German population[J]. J Endod, 2020, 46(2): 184-191.
[15] Shaheen E, Khalil W, Ezeldeen M, et al.Accuracy of segmentation of tooth structures using 3 different CBCT machines[J]. Oral Surg Oral Med Oral Pathol Oral Radiol, 2017, 123(1): 123-128.
[16] Galibourg A, Dumoncel J, Telmon N, et al.Assessment of automatic segmentation of teeth using a watershed-based method[J]. Dentomaxillofac Radiol, 2018, 47(1): 20170220.
[17] Michetti J, Basarab A, Diemer F, et al.Comparison of an adaptive local thresholding method on CBCT and μCT endodontic images[J]. Phys Med Biol, 2017, 63(1): 015020.
[18] Lahoud P, Ezeldeen M, Beznik T, et al.Artificial intelligence for fast and accurate 3D tooth segmentation on CBCT[J]. J Endod, 2021, 47(5): 827-835.
[19] Qu L, Zhang Y, Wang S, et al. Synthesized 7T MRI from 3T MRI via deep learning in spatial and wavelet domains[J]. Med Image Anal, 2020, 62: 101663.e1-101663.e28.
[20] Chun IY, Zheng X, Long Y, et al.BCD-Net for low-dose CT reconstruction: acceleration, convergence, and generalization[C]//Shen D, Liu T, Peters TM, et al. Medical Image Computing and Computer Assisted Intervention-MICCAI 2019. Shenzhen: Proceedings of 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, 2019: 31-40.
[21] Yin X, Zhao Q, Liu J, et al.Domain progressive 3D residual convolution network to improve low-dose CT imaging[J]. IEEE Trans Med Imaging, 2019, 38(12): 2903-2913.
[22] Fu Y, Deng Q, Xie Z, et al.Coronal root canal morphology of permanent two-rooted mandibular first molars with novel 3D measurements[J]. Int Endod J, 2020, 53(2): 167-175.
[23] Zehnder MS, Connert T, Weiger R, et al.Guided endodontics: accuracy of a novel method for guided access cavity preparation and root canal location[J]. Int Endod J, 2016, 49(10): 966-972.
[24] Liang X, Liao W, Cai H, et al.3D-printed artificial teeth: accuracy and application in root canal therapy[J]. J Biomed Nanotechnol, 2018, 14(8): 1477-1485.