Age-group determination of living individuals using first molar images based on artificial intelligence

Seunghyeon Kim, Yeon-Hee Lee, Yung-Kyun Noh, Frank C Park, Q-Schick Auh, Seunghyeon Kim, Yeon-Hee Lee, Yung-Kyun Noh, Frank C Park, Q-Schick Auh

Abstract

Dental age estimation of living individuals is difficult and challenging, and there is no consensus method in adults with permanent dentition. Thus, we aimed to provide an accurate and robust artificial intelligence (AI)-based diagnostic system for age-group estimation by incorporating a convolutional neural network (CNN) using dental X-ray image patches of the first molars extracted via panoramic radiography. The data set consisted of four first molar images from the right and left sides of the maxilla and mandible of each of 1586 individuals across all age groups, which were extracted from their panoramic radiographs. The accuracy of the tooth-wise estimation was 89.05 to 90.27%. Performance accuracy was evaluated mainly using a majority voting system and area under curve (AUC) scores. The AUC scores ranged from 0.94 to 0.98 for all age groups, which indicates outstanding capacity. The learned features of CNNs were visualized as a heatmap, and revealed that CNNs focus on differentiated anatomical parameters, including tooth pulp, alveolar bone level, or interdental space, depending on the age and location of the tooth. With this, we provided a deeper understanding of the most informative regions distinguished by age groups. The prediction accuracy and heat map analyses support that this AI-based age-group determination model is plausible and useful.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Overall workflow.
Figure 2
Figure 2
Classification of a panoramic dental X-ray using learned networks to predict age-group of the patient.
Figure 3
Figure 3
Conversion of the age-group prediction result of the five age-groups to the three age-groups.
Figure 4
Figure 4
Example of correctly classified teeth patches and their Grad-CAM results. The figures show the original molar image (left), Grad-CAM for three age-group prediction (middle) and Grad-CAM for five age-group prediction (right). Note that the softmax score for every first molar image in the figure is above 0.99.
Figure 5
Figure 5
Confusion matrix of the age-group estimation results of the predictions for the three age-groups and five age-groups. In (a), the network trained for predicting three age-groups. In (b), the network trained for predicting five age-groups is used for three age-group prediction. Each element of the confusion matrix represents the average number of patients over five-fold cross validation.
Figure 6
Figure 6
A comparison of average ROC curves over five-fold cross validation between the networks to obtain predictions for three age-groups and five age-groups. Each curve is obtained by binary classification to discriminate one age-group from others. The network that learned to predict five age-groups is used to predict the three age-groups. The AUC of each age-group shows (a) ages 0–19: Three-age group = 0.98 ± 0.01 and five-age group = 0.98 ± 0.01 (p value = 0.97), (b) ages 20–49: three-age group = 0.95 ± 0.01 and five-age group = 0.94 ± 0.01 (p value = 0.12), (c) ages over 50: three-age group = 0.97 ± 0.01 and five-age group = 0.97 ± 0.01 (p value = 0.98).

References

    1. Sajid M, Taj IA, Bajwa UI, Ratyal NI. Facial asymmetry-based age group estimation: role in recognizing age-separated face images. J. Forensic Sci. 2018;63:1727–1749. doi: 10.1111/1556-4029.13798.
    1. Sykes L, Bhayat A, Bernitz H. The effects of the refugee crisis on age estimation analysis over the past 10 years: a 16-country survey. Int. J. Environ. Res. Public Health. 2017;14:630. doi: 10.3390/ijerph14060630.
    1. Ren X, et al. Regression convolutional neural network for automated pediatric bone age assessment from hand radiograph. IEEE J. Biomed. Health Inform. 2019;23:2030–2038. doi: 10.1109/jbhi.2018.2876916.
    1. Mihail RP, Liang G, Jacobs N. Automatic hand skeletal shape estimation from radiographs. IEEE Trans. Nanobiosci. 2019;18:296–305. doi: 10.1109/tnb.2019.2911026.
    1. Manzoor Mughal A, Hassan N, Ahmed A. Bone age assessment methods: a critical review. Pak. J. Med. Sci. 2014;30:211–215. doi: 10.12669/pjms.301.4295.
    1. Rozylo-Kalinowska I, Kolasa-Raczka A, Kalinowski P. Relationship between dental age according to Demirjian and cervical vertebrae maturity in Polish children. Eur. J. Orthod. 2011;33:75–83. doi: 10.1093/ejo/cjq031.
    1. Moorrees CF, Fanning EA, Hunt EE., Jr Age variation of formation stages for ten permanent teeth. J. Dent. Res. 1963;42:1490–1502. doi: 10.1177/00220345630420062701.
    1. Cameriere R, et al. Age estimation in children by measurement of open apices in teeth with Bayesian calibration approach. Forensic Sci. Int. 2016;258:50–54. doi: 10.1016/j.forsciint.2015.11.005.
    1. Gotmare SS, et al. The coronal pulp cavity index: a forensic tool for age determination in adults. Dent. Res. J. (Isfahan) 2019;16:160–165. doi: 10.4103/1735-3327.255743.
    1. Sarajlić N, Topić B, Brkić H, Alajbeg IZ. Aging quantification on alveolar bone loss. Coll. Antropol. 2009;33:1165–1170.
    1. Ruquet M, Saliba-Serre B, Tardivo D, Foti B. Estimation of age using alveolar bone loss: forensic and anthropological applications. J. Forensic Sci. 2015;60:1305–1309. doi: 10.1111/1556-4029.12827.
    1. Koh KK, et al. Age estimation from structural changes of teeth and buccal alveolar bone level. J. Forensic Leg. Med. 2017;48:15–21. doi: 10.1016/j.jflm.2017.03.004.
    1. Cameriere R, Ferrante L, Cingolani M. Precision and reliability of pulp/tooth area ratio (RA) of second molar as indicator of adult age. J. Forensic Sci. 2004;49:1319–1323.
    1. Juneja M, Devi YB, Rakesh N, Juneja S. Age estimation using pulp/tooth area ratio in maxillary canines-A digital image analysis. J. Forensic Dent. Sci. 2014;6:160–165. doi: 10.4103/0975-1475.137047.
    1. Shah PH, Venkatesh R. Pulp/tooth ratio of mandibular first and second molars on panoramic radiographs: an aid for forensic age estimation. J. Forensic Dent. Sci. 2016;8:112. doi: 10.4103/0975-1475.186374.
    1. Huttner EA, Machado DC, de Oliveira RB, Antunes AG, Hebling E. Effects of human aging on periodontal tissues. Special Care Dent. 2009;29:149–155. doi: 10.1111/j.1754-4505.2009.00082.x.
    1. Chen H, et al. A deep learning approach to automatic teeth detection and numbering based on object detection in dental periapical films. Sci. Rep. 2019;9:3840. doi: 10.1038/s41598-019-40414-y.
    1. Vinayahalingam S, Xi T, Bergé S, Maal T, de Jong G. Automated detection of third molars and mandibular nerve by deep learning. Sci. Rep. 2019;9:9007. doi: 10.1038/s41598-019-45487-3.
    1. Krois J, et al. Deep learning for the radiographic detection of periodontal bone loss. Sci. Rep. 2019;9:8495–8495. doi: 10.1038/s41598-019-44839-3.
    1. Demirjian A, Goldstein H, Tanner JM. A new system of dental age assessment. Hum. Biol. 1973;45:211–227.
    1. Merdietio-Boedi R, et al. Effect of lower third molar segmentations on automated tooth development staging using a convolutional neural network. J. Forensic Sci. 2020;65:481–486. doi: 10.1111/1556-4029.14182.
    1. Mathew DG, et al. Adult forensic age estimation using mandibular first molar radiographs: a novel technique. J. Forensic Dent. Sci. 2013;5:56–59. doi: 10.4103/0975-1475.114552.
    1. He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, 2016, 770–778, 10.1109/CVPR.2016.90.
    1. Russakovsky O, Deng J, Su H, et al. ImageNet large scale visual recognition challenge. Int. J. Comput. Vis. 2015;115:211–252. doi: 10.1007/s11263-015-0816-y.
    1. Tajbakhsh N, et al. Convolutional neural networks for medical image analysis: full training or fine tuning? IEEE Trans. Med. Imaging. 2016;35(5):1299–1312. doi: 10.1109/TMI.2016.2535302.
    1. Shin H-C, et al. Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning. IEEE Trans. Med. Imaging. 2016;35:1285–1298. doi: 10.1109/TMI.2016.2528162.
    1. Chang, J., Yu, J., Han, T., Chang, H-J., & Park, E. (2017). A method for classifying medical images using transfer learning: a pilot study on histopathology of breast cancer. In 2017 IEEE 19th International Conference on e-Health Networking, Applications and Services, Healthcom 2017 (pp. 1–4). (2017 IEEE 19th International Conference on e-Health Networking, Applications and Services, Healthcom 2017; Vol. 2017-December). Institute of Electrical and Electronics Engineers Inc. 10.1109/HealthCom.2017.8210843
    1. Dubey SR, et al. diffGrad: an optimization method for convolutional neural networks. IEEE Trans. Neural Netw. Learn. Syst. 2019 doi: 10.1109/tnnls.2019.2955777.
    1. Iizuka T, Fukasawa M, Kameyama M. Deep-learning-based imaging-classification identified cingulate island sign in dementia with Lewy bodies. Sci. Rep. 2019;9:8944. doi: 10.1038/s41598-019-45415-5.
    1. Ohannessian R, et al. Heat map for data visualization in infection control epidemiology: an application describing the relationship between hospital-acquired infections, Simplified Acute Physiological Score II, and length of stay in adult intensive care units. Am. J. Infect. Control. 2017;45:746–749. doi: 10.1016/j.ajic.2017.02.012.
    1. Jia L, Zhang W, Chen X. Common methods of biological age estimation. Clin. Interv. Aging. 2017;12:759–772. doi: 10.2147/CIA.S134921.
    1. Mohammed RB, et al. Digital radiographic evaluation of mandibular third molar for age estimation in young adults and adolescents of South Indian population using modified Demirjian's method. J. Forensic Dent. Sci. 2014;6:191–196. doi: 10.4103/0975-1475.137068.
    1. Ilayaraja V, et al. Digitized morphometric analysis using maxillary canine and mandibular first molar for age estimation in South Indian population. Open Dent. J. 2018;12:762–769. doi: 10.2174/1745017901814010762.
    1. Garamendi PM, Landa MI, Botella MC, Alemán I. Forensic age estimation on digital X-ray images: medial epiphyses of the clavicle and first rib ossification in relation to chronological age. J. Forensic Sci. 2011;56(Suppl 1):S3–12. doi: 10.1111/j.1556-4029.2010.01626.x.
    1. Limdiwala PG, Shah JS. Age estimation by using dental radiographs. J. Forensic Dent. Sci. 2013;5:118–122. doi: 10.4103/0975-1475.119778.
    1. Rawat W, Wang Z. Deep convolutional neural networks for image classification: a comprehensive review. Neural Comput. 2017;29:2352–2449. doi: 10.1162/NECO_a_00990.
    1. Tang Z, et al. Interpretable classification of Alzheimer's disease pathologies with a convolutional neural network pipeline. Nat. Commun. 2019;10:2173–2173. doi: 10.1038/s41467-019-10212-1.
    1. Krishan K, Kanchan T, Garg AK. Dental evidence in forensic identification—an overview, methodology and present status. Open Dent. J. 2015;9:250–256. doi: 10.2174/1874210601509010250.
    1. Liao H, Yan Y, Dai W, Fan P. Age estimation of face images based on CNN and divide-and-rule strategy. Math. Probl. Eng. 2018;2018:1712686. doi: 10.1155/2018/1712686.
    1. Baltruschat IM, Nickisch H, Grass M, Knopp T, Saalbach A. Comparison of deep learning approaches for multi-label chest x-ray classification. Sci. Rep. 2019;9:6381. doi: 10.1038/s41598-019-42294-8.
    1. Sakata A, Takemura N, Yagi Y. Gait-based age estimation using multi-stage convolutional neural network. IPSJ Trans. Comput. Vis. Appl. 2019;11:4. doi: 10.1186/s41074-019-0054-2.
    1. Karaarslan B, Karaarslan ES, Ozsevik AS, Ertas E. Age estimation for dental patients using orthopantomographs. Eur. J. Dent. 2010;4:389–394. doi: 10.1055/s-0039-1697858.
    1. Jain S, et al. Tooth coronal index and pulp/tooth ratio in dental age estimation on digital panoramic radiographs-A comparative study. Forensic Sci. Int. 2017;277:115–121. doi: 10.1016/j.forsciint.2017.05.006.
    1. Hansen LK, Salamon P. Neural network ensembles. IEEE Trans. Pattern Anal. Mach. Intell. 1990;12:993–1001. doi: 10.1109/34.58871.
    1. Irvine N, Nugent C, Zhang S, Wang H, Ng WWY. Neural network ensembles for sensor-based human activity recognition within smart environments. Sensors (Basel) 2019;20:216. doi: 10.3390/s20010216.
    1. Abuassba AOM, Zhang D, Luo X, Shaheryar A, Ali H. Improving classification performance through an advanced ensemble based heterogeneous extreme learning machines. Comput. Intell. Neurosci. 2017;2017:3405463. doi: 10.1155/2017/3405463.
    1. Schmidt J, Marques MRG, Botti S, Marques MAL. Recent advances and applications of machine learning in solid-state materials science. Comput. Mater. 2019;5:83. doi: 10.1038/s41524-019-0221-0.

Source: PubMed

Sponsorzy i współpracownicy

Warunki medyczne

Interwencje lekowe

3
Subskrybuj