Abnormal maxillary sinus diagnosing on CBCT images via object detection and 'straight-forward' classification deep learning strategy.
cone-beam computed tomography
deep learning
dental implant
mass screening
paranasal sinus disease
Journal
Journal of oral rehabilitation
ISSN: 1365-2842
Titre abrégé: J Oral Rehabil
Pays: England
ID NLM: 0433604
Informations de publication
Date de publication:
Dec 2023
Dec 2023
Historique:
revised:
06
04
2023
received:
23
02
2023
accepted:
18
08
2023
medline:
3
11
2023
pubmed:
4
9
2023
entrez:
4
9
2023
Statut:
ppublish
Résumé
Pathological maxillary sinus would affect implant treatment and even result in failure of maxillary sinus lift and implant surgery. However, the maxillary sinus abnormalities are challenging to be diagnosed through CBCT images, especially for young dentists or dentists in grassroots medical institutions without systematical education of general medicine. To develop a deep-learning-based screening model incorporating object detection and 'straight-forward' classification strategy to screen out maxillary sinus abnormalities on CBCT images. The large area of background noise outside maxillary sinus would affect the generalisation and prediction accuracy of the model, and the diversity and imbalanced distribution of imaging manifestations may bring challenges to intellectualization. Thus we adopted an object detection to limit model's observation zone and 'straight-forward' classification strategy with various tuning methods to adapt to dental clinical need and extract typical features of diverse manifestations so that turn the task into a 'normal-or-not' classification. We successfully constructed a deep-learning model consist of well-trained detector and diagnostor module. This model achieved ideal AUROC and AUPRC of 0.953 and 0.887, reaching more than 90% accuracy at optimal cut-off. McNemar and Kappa test verified no statistical difference and high consistency between the prediction and ground truth. Dentist-model comparison test showed the model's statistically higher diagnostic performance than dental students. Visualisation method confirmed the model's effectiveness in region recognition and feature extraction. The deep-learning model incorporating object detection and straightforward classification strategy could achieve satisfying predictive performance for screening maxillary sinus abnormalities on CBCT images.
Sections du résumé
BACKGROUND
BACKGROUND
Pathological maxillary sinus would affect implant treatment and even result in failure of maxillary sinus lift and implant surgery. However, the maxillary sinus abnormalities are challenging to be diagnosed through CBCT images, especially for young dentists or dentists in grassroots medical institutions without systematical education of general medicine.
OBJECTIVES
OBJECTIVE
To develop a deep-learning-based screening model incorporating object detection and 'straight-forward' classification strategy to screen out maxillary sinus abnormalities on CBCT images.
METHODS
METHODS
The large area of background noise outside maxillary sinus would affect the generalisation and prediction accuracy of the model, and the diversity and imbalanced distribution of imaging manifestations may bring challenges to intellectualization. Thus we adopted an object detection to limit model's observation zone and 'straight-forward' classification strategy with various tuning methods to adapt to dental clinical need and extract typical features of diverse manifestations so that turn the task into a 'normal-or-not' classification.
RESULTS
RESULTS
We successfully constructed a deep-learning model consist of well-trained detector and diagnostor module. This model achieved ideal AUROC and AUPRC of 0.953 and 0.887, reaching more than 90% accuracy at optimal cut-off. McNemar and Kappa test verified no statistical difference and high consistency between the prediction and ground truth. Dentist-model comparison test showed the model's statistically higher diagnostic performance than dental students. Visualisation method confirmed the model's effectiveness in region recognition and feature extraction.
CONCLUSION
CONCLUSIONS
The deep-learning model incorporating object detection and straightforward classification strategy could achieve satisfying predictive performance for screening maxillary sinus abnormalities on CBCT images.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
1465-1480Subventions
Organisme : Guangdong Financial Fund for High-Caliber Hospital Construction
Organisme : National Undergraduate Training Program for Innovation and Entrepreneurship (202210780)
Organisme : Science and Technology Program of Guangzhou, China (2023B03J1232)
Organisme : Special Funds for the Cultivation of Guangdong College Students' Scientific and Technological Innovation ("Climbing Program" Special Funds, pdjh2023b0013)
Informations de copyright
© 2023 John Wiley & Sons Ltd.
Références
Anand VK. Epidemiology and economic impact of rhinosinusitis. Ann Otol Rhinol Laryngol Suppl. 2004;193:3-5. doi:10.1177/00034894041130s502
Hastan D, Fokkens WJ, Bachert C, et al. Chronic rhinosinusitis in Europe-an underestimated disease. A GA2LEN study. Allergy. 2011;66(9):1216-1223. doi:10.1111/j.1398-9995.2011.02646.x
Shi JB, Fu QL, Zhang H, et al. Epidemiology of chronic rhinosinusitis: results from a cross-sectional survey in seven Chinese cities. Allergy. 2015;70(5):533-539. doi:10.1111/all.12577
Psillas G, Papaioannou D, Petsali S, Dimas GG, Constantinidis J. Odontogenic maxillary sinusitis: a comprehensive review. J Dent Sci. 2021;16(1):474-481. doi:10.1016/j.jds.2020.08.001
Niu L, Wang J, Yu H, Qiu L. New classification of maxillary sinus contours and its relation to sinus floor elevation surgery. Clin Implant Dent Relat Res. 2018;20(4):493-500. doi:10.1111/cid.12606
De Fauw J, Ledsam JR, Romera-Paredes B, et al. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nat Med. 2018;24(9):1342-1350. doi:10.1038/s41591-018-0107-6
Shanbhag S, Karnik P, Shirke P, Shanbhag V. Cone-beam computed tomographic analysis of sinus membrane thickness, ostium patency, and residual ridge heights in the posterior maxilla: implications for sinus floor elevation. Clin Oral Implants Res. 2014;25(6):755-760. doi:10.1111/clr.12168
Haixiang G, Yijing L, Shang J, Mingyun G, Yuanyue H, Bing G. Learning from class-imbalanced data: review of methods and applications. Expert Syst AppL. 2017;73:220-239. doi:10.1016/j.eswa.2016.12.035
Miracle AC, Mukherji SK. Cone beam CT of the head and neck, part 1: physical principles. AJNR Am J Neuroradiol. 2009;30(6):1088-1095. doi:10.3174/ajnr.A1653
Fokkens WJ, Lund VJ, Hopkins C, et al. European position paper on Rhinosinusitis and nasal polyps 2020. Rhinology. 2020;58(Suppl S29):1-464. doi:10.4193/Rhin20.600
Redmon J, Divvala S, Girshick R, Farhadi A. You only look once: unified. Real-Time Object Detection IEEE. 2016;27-30(2016):779-788. doi:10.1109/CVPR.2016.91
Ruder S. An overview of gradient descent optimization algorithms. Working Paper arXiv. 2016. doi:10.48550/arXiv.1609.04747
Lin Y, Shi M, Xiang D, et al. Construction of an end-to-end regression neural network for the determination of a quantitative index sagittal root inclination. J Periodontol. 2022;93(12):1951-1960. doi:10.1002/jper.21-0492
Kingma D, Ba J. Adam: a method for stochastic optimization. arXiv. 2014. doi:10.48550/arXiv.1412.6980
Sabanayagam C, Xu D, Ting DSW, et al. A deep learning algorithm to detect chronic kidney disease from retinal photographs in community-based populations. Lancet Digit Health. 2020;2(6):e295-e302. doi:10.1016/s2589-7500(20)30063-7
Ozenne B, Subtil F, Maucort-Boulch D. The precision-recall curve overcame the optimism of the receiver operating characteristic curve in rare diseases. Article. J Clin Epidemiol. 2015;68(8):855-859. doi:10.1016/j.jclinepi.2015.02.010
Jeon Y, Lee K, Sunwoo L, et al. Deep learning for diagnosis of paranasal sinusitis using multi-view radiographs. Diagnostics (Basel). 2021;11(2):250. doi:10.3390/diagnostics11020250
Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-CAM: visual explanations from deep networks via gradient-based localization. INT J COMPUT Vision. 2020;128(2):336-359. doi:10.1109/ICCV.2017.74
Zou Z, Shi Z, Guo Y, Ye J. Object detection in 20 years: a survey. Working Paper arXiv. 2019. doi:10.48550/arXiv.1905.05055
Wu X, Sahoo D, Hoi S. Recent advances in deep learning for object detection. Neurocomputing. 2020;396:39-64. doi:10.48550/arXiv.1908.03673
Everingham M, Gool LV, Williams CKI, Winn J, Zisserman A. The Pascal visual object classes (VOC) challenge. Int J Comput Vision. 2010;88(2):303-338. doi:10.1007/s11263-009-0275-4
Lee KS, Kwak HJ, Oh JM, et al. Automated detection of TMJ osteoarthritis based on artificial intelligence. J Dent Res. 2020;99(12):1363-1367. doi:10.1177/0022034520936950
Xie S, Girshick R, Dollár P, Tu Z, He K. Aggregated residual transformations for deep neural networks. Working Paper. arXiv. 2016;5987-5995. doi:10.1109/CVPR.2017.634
Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the inception architecture for computer Vision. IEEE. 2016;2818-2826. doi:10.1109/CVPR.2016.308
Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions. Working Paper IEEE. 2014:1-9. doi:10.1109/CVPR.2015.7298594
Chan HP, Samala RK, Hadjiiski LM, Zhou C. Deep learning in medical image analysis. Adv Exp Med Biol. 2020;1213:3-21. doi:10.1007/978-3-030-33128-3_1
Emilio S, José D. Handbook of Research on Machine Learning Applications and Trends. IGI Publishing; 2009;11(Transfer Learning) (11):242-264. doi:10.4018/978-1-60566-766-9
Nakkiran P, Kaplun G, Bansal Y, Yang T, Barak B, Sutskever I. Deep Double Descent: Where Bigger Models and More Data Hurt. arXiv. 2019. doi:10.48550/arXiv.1912.02292
Hastie T, Montanari A, Rosset S, Tibshirani RJ. Surprises in high-dimensional Ridgeless least squares interpolation. Ann Stat. 2019;50(2):949-986. doi:10.48550/arXiv.1903.08560
He K, Girshick R, Dollár P. Rethinking ImageNet Pre-Training IEEE. 2018:4917-4926. doi:10.1109/ICCV.2019.00502
Vranckx M, Van Gerven A, Willems H, et al. Artificial intelligence (AI)-driven molar angulation measurements to predict third molar eruption on panoramic radiographs. Int J Environ Res Public Health. 2020;17(10):3716. doi:10.3390/ijerph17103716
Yu HJ, Cho SR, Kim MJ, Kim WH, Kim JW, Choi J. Automated skeletal classification with lateral Cephalometry based on artificial intelligence. J Dent Res. 2020;99(3):249-256. doi:10.1177/0022034520901715
Connor S, Taghi MK. A survey on image data augmentation for deep learning. Article. J Big Data. 2019;6(1):1-48. doi:10.1186/s40537-019-0197-0
Xiao W, Huang X, Wang JH, et al. Screening and identifying hepatobiliary diseases through deep learning using ocular images: a prospective, multicentre study. Lancet Digit Health. 2021;3(2):e88-e97. doi:10.1016/s2589-7500(20)30288-0
Zhang K, Liu X, Xu J, et al. Deep-learning models for the detection and incidence prediction of chronic kidney disease and type 2 diabetes from retinal fundus images. Nat Biomed Eng. 2021;5(6):533-545. doi:10.1038/s41551-021-00745-6
Balasubramanian VN. Toward explainable deep learning. Commun ACM. 2022;65(11):68-69. doi:10.1145/3550491
Lipton ZC. The mythos of model interpretability. Commun ACM. 2018;61(10):36-43. doi:10.1145/3233231
Ribeiro MT, Singh S, Guestrin C. "why should I trust you?": explaining the predictions of any classifier. ACM. 2016:1135-1144. doi:10.48550/arXiv.1602.04938
Shan T, Tay FR, Gu L. Application of artificial intelligence in dentistry. J Dent Res. 2021;100(3):232-244. doi:10.1177/0022034520969115
Lei S, Zhang H, Wang K, Su Z. How Training Data Affect the Accuracy and Robustness of Neural Networks for Image Classification. openreview; 2018.