Segmentation and validation of mandibular canal and its bifurcation on cone beam CT based on deep learning

doi:10.19439/j.sjos.2025.02.002

Abstract

Abstract: PURPOSE: To train the U-net of convolutional neural network to establish a method for detecting and segmenting the mandibular canal and its bifurcation, and validate its accuracy based on the ground truth labeled by experts. METHODS: A total of 290 CBCT scans were collected from Shanghai Stomatological Hospital from Jan. 2022 to Dec. 2022, which were divided into training set of 200 scans and test set of 90 scans. Model training included two steps. In the first step, bilateral mandibular canals and its bifurcation of 50 CBCT scans were labeled in 3D Slicer image computing platform by investigators. Three dimensional U-net segmentation model were trained initially with data enhancement. A morphological post-processing method was applied to the predicted results. In the second step, pseudo label method was employed to help annotating the mandibular canal and corresponding bifurcations on remaining 150 CBCTs, which would be included in training set after revision. Three dimensional U-net model was trained based on these 200 data. During test phase, totally 90 scans were labeled by two doctors and U-net model respectively. Consistency check was conducted to evaluate the labels between two doctors. Dice similarity coefficient and Hausdorff distance were calculated to evaluate the labels between doctors and the model. The detection rate of bifurcation was calculated. SPSS 20.0 software package was used for data analysis. RESULTS: In 90 CBCT test set, the Kappa value between two dentists' annotations was 0.667. The average Dice and Hausdorff distance between predictions and labels of doctors were (0.739±0.068) and (0.988±1.14) mm. In bifurcation detection, the detection rate was 91.30% on scans with clear bifurcations. CONCLSIONS: The dentification and segmentation U-net model of mandibular canal on dental CBCT can be reliable and practical for its high segmentation precision and predicting speed.

Key words: Deep learning, Mandibular canal, Bifid mandibular canal, CBCT

CLC Number:

R781.05

Ye Ye, Fang Shuobo, Lu Huitong, Liu Mingqian, Wu Xueying. Segmentation and validation of mandibular canal and its bifurcation on cone beam CT based on deep learning[J]. Shanghai Journal of Stomatology, 2025, 34(2): 119-125.

References

[1] Arx T, Lozanoff S.Clinical oral anatomy[M]. New York: Springer Cham, 2017.
[2] Denio D, Torabinejad M, Bakland LK.Anatomical relationship of the mandibular canal to its surrounding structures in mature mandibles[J]. J Endod, 1992, 18(4):161-165.
[3] Orhan K, Aksoy S, Bilecenoglu B, et al.Evaluation of bifid mandibular canals with cone-beam computed tomography in a Turkish adult population: a retrospective study[J]. Surg Radiol Anat, 2011, 33(6): 501-507.
[4] Pria CM, Masood F, Beckerley JM, et al.Study of the inferior alveolar canal and mental foramen on digital panoramic images[J]. J Contemp Dent Tract, 2011, 12(4): 265-271.
[5] Bogdán S, Pataky L, Barabás J, et al.Atypical courses of the mandibular canal: comparative examination of dry mandibles and X-rays[J]. J Craniofac Surg, 2006, 17(3): 487-491.
[6] Kuczynski A, Kucharski W, Franco A, et al.Prevalence of bifid mandibular canals in panoramic radiographs: a maxillofacial surgical scope[J]. Surg Radiol Anat, 2014, 36(9): 847-850.
[7] Naitoh M, Hiraiwa Y, Aimiya H, et al.Observation of bifid mandibular canal using cone-beam computerized tomography[J]. Int J Oral Maxillofac Implants, 2009, 24(1): 155-159.
[8] Angelopoulos C, Thomas SL, Hechler S, et al.Comparison between digital panoramic radiography and cone-beam computed tomography for the identification of the mandibular canal as part of presurgical dental implant assessment[J]. J Oral Maxillofac Surg, 2008, 66(10): 2130-2135.
[9] Cantekin K, Sekerci AE, Miloglu O, et al.Identification of the mandibular landmarks in a pediatric population[J]. Med Oral Patol Oral Cir Bucal, 2014, 19(2): e136-e141.
[10] Isensee F, Jaeger PF, Kohl SAA, et al.nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation[J]. Nat Methods, 2021, 18(2):203-211.
[11] Miki Y, Muramatsu C, Hayashi T, et al.Classification of teeth in cone-beam CT using deep convolutional neural network[J]. Comput Biol Med, 2017, 80(1): 24-29.
[12] Vranckx M, Van Gerven A, Willems H, et al.Artificial intelligence (AI)-driven molar angulation measurements to predict third molar eruption on panoramic radiographs[J]. Int J Environ Res Public Health, 2020, 17(10): 3716.
[13] Milletari F, Navab N, Ahmadi SA.V-Net: fully convolutional neural networks for volumetric medical image segmentation[C]//2016 Fourth International Conference on 3D Vision (3DV). CA: Stanford, 2016: 565-571.
[14] Jaskari J, Sahlsten J, Järnstedt J, et al.Deep learning method for mandibular canal segmentation in dental cone beam computed tomography volumes[J]. Sci Rep, 2020, 10(1): 5842.
[15] Abdolali F, Zoroofi RA, Abdolali M, et al.Automatic segmentation of mandibular canal in cone beam CT images using conditional statistical shape model and fast marching[J]. Int J Comput Assist Radiol Surg, 2017, 12(4): 581-593.
[16] 郭毅, 张巧红, 韩晓倩, 等. 下颌管分支的锥形束CT影像学分析[J]. 华西口腔医学杂志, 2015, 33(2): 158-160.
[17] 孙芹,杨永进,李广. 双下颌管分支的CBCT观察[J]. 牙体牙髓牙周病学杂志, 2017, 27(12): 713-716.
[18] Haghighat A, Jafari Z, Hasheminia D, et al.Comparison of success rate and onset time of two different anesthesia techniques[J]. Med Oral Patol Oral Cir Bucal, 2015, 20(4): e459-e463.
[19] Neves FS, Nascimento MC, Oliveira ML, et al.Comparative analysis of mandibular anatomical variations between panoramic radiography and cone beam computed tomography[J]. Oral Maxillofac Surg, 2014, 18(4): 419-424.
[20] Zhang YQ, Zhao YN, Liu DG, et al.Bifid variations of the mandibular canal: cone beam computed tomography evaluation of 1000 Northern Chinese patients[J]. Oral Surg Oral Med Oral Pathol Oral Radiol, 2018, 126(5): e271-e278.
[21] 余蕾, 朱钢. 基于CBCT 的西南地区汉族人双下颌管形态学研究[J]. 口腔医学, 2018, 38(8): 713-716.
[22] Yang X, Lyu C, Zou D.Bifid mandibular canals incidence and anatomical variations in the population of Shanghai area by cone beam computed tomography[J]. J Comput Assist Tomogr, 2017, 41(4): 535-540.