Korean Journal of Radiology

  • 제24권11호
  • /
  • Pages.1151-1163
  • /
  • 2023
  • /
  • 1229-6929(pISSN)
  • /
  • 2005-8330(eISSN)
대한영상의학회 (The Korean Society of Radiology)
권호 표지
DOI QR코드

DOI QR Code

Bone Age Assessment Using Artificial Intelligence in Korean Pediatric Population: A Comparison of Deep-Learning Models Trained With Healthy Chronological and Greulich-Pyle Ages as Labels

  • Pyeong Hwa Kim (Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Hee Mang Yoon (Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Jeong Rye Kim (Department of Radiology, Dankook University Hospital, Dankook University College of Medicine) ;
  • Jae-Yeon Hwang (Department of Radiology, Research Institute for Convergence of Biomedical Science and Technology, Pusan National University Yangsan Hospital, Pusan National University School of Medicine) ;
  • Jin-Ho Choi (Department of Pediatrics, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Jisun Hwang (Department of Radiology, Ajou University Hospital, Ajou University School of Medicine) ;
  • Jaewon Lee (VUNO Inc.) ;
  • Jinkyeong Sung (VUNO Inc.) ;
  • Kyu-Hwan Jung (VUNO Inc.) ;
  • Byeonguk Bae (VUNO Inc.) ;
  • Ah Young Jung (Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Young Ah Cho (Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Woo Hyun Shim (Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Boram Bak (University of Ulsan Foundation for Industry Cooperation, Asan Medical Center, University of Ulsan College of Medicine) ;
  • Jin Seong Lee (Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine)
  • 투고 : 2022.07.08
  • 심사 : 2023.08.06
  • 발행 : 2023.11.01
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

Objective: To develop a deep-learning-based bone age prediction model optimized for Korean children and adolescents and evaluate its feasibility by comparing it with a Greulich-Pyle-based deep-learning model. Materials and Methods: A convolutional neural network was trained to predict age according to the bone development shown on a hand radiograph (bone age) using 21036 hand radiographs of Korean children and adolescents without known bone development-affecting diseases/conditions obtained between 1998 and 2019 (median age [interquartile range {IQR}], 9 [7-12] years; male:female, 11794:9242) and their chronological ages as labels (Korean model). We constructed 2 separate external datasets consisting of Korean children and adolescents with healthy bone development (Institution 1: n = 343; median age [IQR], 10 [4-15] years; male: female, 183:160; Institution 2: n = 321; median age [IQR], 9 [5-14] years; male: female, 164:157) to test the model performance. The mean absolute error (MAE), root mean square error (RMSE), and proportions of bone age predictions within 6, 12, 18, and 24 months of the reference age (chronological age) were compared between the Korean model and a commercial model (VUNO Med-BoneAge version 1.1; VUNO) trained with Greulich-Pyle-based age as the label (GP-based model). Results: Compared with the GP-based model, the Korean model showed a lower RMSE (11.2 vs. 13.8 months; P = 0.004) and MAE (8.2 vs. 10.5 months; P = 0.002), a higher proportion of bone age predictions within 18 months of chronological age (88.3% vs. 82.2%; P = 0.031) for Institution 1, and a lower MAE (9.5 vs. 11.0 months; P = 0.022) and higher proportion of bone age predictions within 6 months (44.5% vs. 36.4%; P = 0.044) for Institution 2. Conclusion: The Korean model trained using the chronological ages of Korean children and adolescents without known bone development-affecting diseases/conditions as labels performed better in bone age assessment than the GP-based model in the Korean pediatric population. Further validation is required to confirm its accuracy.

키워드

과제정보

This study was supported by a grant (2020IE0016-1) from Asan Medical Center Children's Hospital (Heart Institute), Seoul, Korea. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (2020R1G1A1004591). This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI22C1723). This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HR20C0026).

참고문헌

  1. Oh MS, Kim S, Lee J, Lee MS, Kim YJ, Kang KS. Factors associated with advanced bone age in overweight and obese children. Pediatr Gastroenterol Hepatol Nutr 2020;23:89-97
  2. Kim D, Cho SY, Maeng SH, Yi ES, Jung YJ, Park SW, et al. Diagnosis and constitutional and laboratory features of Korean girls referred for precocious puberty. Korean J Pediatr 2012;55:481-486
  3. Kelly PM, Dimeglio A. Lower-limb growth: how predictable are predictions? J Child Orthop 2008;2:407-415
  4. Creo AL, Schwenk WF 2nd. Bone age: a handy tool for pediatric providers. Pediatrics 2017;140:e20171486
  5. Greulich WW, Pyle SI. Radiographic atlas of skeletal development of the hand and wrist. 2nd ed. Redwood City: Stanford University Press, 1999
  6. Lee BD, Lee MS. Automated bone age assessment using artificial intelligence: the future of bone age assessment. Korean J Radiol 2021;22:792-800
  7. Roche AF, Rohmann CG, French NY, Davila GH. Effect of training on replicability of assessments of skeletal maturity (Greulich-Pyle). Am J Roentgenol Radium Ther Nucl Med 1970;108:511-515
  8. Bull RK, Edwards PD, Kemp PM, Fry S, Hughes IA. Bone age assessment: a large scale comparison of the Greulich and Pyle, and Tanner and Whitehouse (TW2) methods. Arch Dis Child 1999;81:172-173
  9. Kim JR, Shim WH, Yoon HM, Hong SH, Lee JS, Cho YA, et al. Computerized bone age estimation using deep learning based program: evaluation of the accuracy and efficiency. AJR Am J Roentgenol 2017;209:1374-1380
  10. Eng DK, Khandwala NB, Long J, Fefferman NR, Lala SV, Strubel NA, et al. Artificial intelligence algorithm improves radiologist performance in skeletal age assessment: a prospective multicenter randomized controlled trial. Radiology 2021;301:692-699
  11. Booz C, Yel I, Wichmann JL, Boettger S, Al Kamali A, Albrecht MH, et al. Artificial intelligence in bone age assessment: accuracy and efficiency of a novel fully automated algorithm compared to the Greulich-Pyle method. Eur Radiol Exp 2020;4:6
  12. Lee KC, Lee KH, Kang CH, Ahn KS, Chung LY, Lee JJ, et al. Clinical validation of a deep learning-based hybrid (Greulich-Pyle and modified Tanner-Whitehouse) method for bone age assessment. Korean J Radiol 2021;22:2017-2025
  13. Alshamrani K, Messina F, Offiah AC. Is the Greulich and Pyle atlas applicable to all ethnicities? A systematic review and meta-analysis. Eur Radiol 2019;29:2910-2923
  14. Zhang A, Sayre JW, Vachon L, Liu BJ, Huang HK. Racial differences in growth patterns of children assessed on the basis of bone age. Radiology 2009;250:228-235
  15. Ontell FK, Ivanovic M, Ablin DS, Barlow TW. Bone age in children of diverse ethnicity. AJR Am J Roentgenol 1996;167:1395-1398
  16. Hwang J, Yoon HM, Hwang JY, Kim PH, Bak B, Bae BU, et al. Re-assessment of applicability of Greulich and Pyle-based bone age to Korean children using manual and deep learning-based automated method. Yonsei Med J 2022;63:683-691
  17. Yeon KM. Standard bone-age of infants and children in Korea. J Korean Med Sci 1997;12:9-16
  18. King DG, Steventon DM, O'Sullivan MP, Cook AM, Hornsby VP, Jefferson IG, et al. Reproducibility of bone ages when performed by radiology registrars: an audit of Tanner and Whitehouse II versus Greulich and Pyle methods. Br J Radiol 1994;67:848-851
  19. Mongan J, Moy L, Kahn CE Jr. Checklist for artificial intelligence in medical imaging (CLAIM): a guide for authors and reviewers. Radiol Artif Intell 2020;2:e200029
  20. Nejedly N. Normal and abnormal growth in the pediatric patient. Curr Probl Pediatr Adolesc Health Care 2020;50:100771
  21. Kim JH, Yun S, Hwang SS, Shim JO, Chae HW, Lee YJ, et al. The 2017 Korean National Growth Charts for children and adolescents: development, improvement, and prospects. Korean J Pediatr 2018;61:135-149
  22. Iglovikov VI, Rakhlin A, Kalinin AA, Shvets AA. Paediatric bone age assessment using deep convolutional neural networks. In: Stoyanov D, Taylor Z, Carneiro G, Syeda-Mahmood T, Martel A, Maier-Hein L, et al., eds. Deep learning in medical image analysis and multimodal learning for clinical decision support. DLMIA ML-CDS 2018. Lecture notes in computer science, vol 11045. Cham: Springer, 2018:300-308
  23. Halabi SS, Prevedello LM, Kalpathy-Cramer J, Mamonov AB, Bilbily A, Cicero M, et al. The RSNA pediatric bone age machine learning challenge. Radiology 2019;290:498-503
  24. Wang J, Sun K, Cheng T, Jiang B, Deng C, Zhao Y, et al. Deep high-resolution representation learning for visual recognition. IEEE Trans Pattern Anal Mach Intell 2020;43:3349-3364
  25. He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR); 2016 Jun 27-30; Las Vegas, USA: IEEE; 2016:770-778
  26. Deng J, Dong W, Socher R, Li LJ, Li K, Fei-Fei L. ImageNet: a large-scale hierarchical image database. Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition; 2009 Jun 20-25; Miami, USA: IEEE; 2009:248-255
  27. Gao BB, Zhou HY, Wu J, Geng X. Age estimation using expectation of label distribution learning. Proceedings of the 27th International Joint Conference on Artificial Intelligence (IJCAI-18); 2018 Jul 13-19; Stockholm, Sweden: IJCAI; 2018:712-718
  28. Larson DB, Chen MC, Lungren MP, Halabi SS, Stence NV, Langlotz CP. Performance of a deep-learning neural network model in assessing skeletal maturity on pediatric hand radiographs. Radiology 2018;287:313-322
  29. Lea WW, Hong SJ, Nam HK, Kang WY, Yang ZP, Noh EJ. External validation of deep learning-based bone-age software: a preliminary study with real world data. Sci Rep 2022;12:1232
  30. Martin DD, Wit JM, Hochberg Z, Savendahl L, van Rijn RR, Fricke O, et al. The use of bone age in clinical practice - part 1. Horm Res Paediatr 2011;76:1-9
  31. Richmond EJ, Rogol AD. Diagnostic approach to children and adolescents with short stature [accessed on April 12, 2022]. Available at: https://www.uptodate.com/contents/diagnostic-approach-to-children-and-adolescents-with-short-stature
  32. Gilsanz V, Ratib O. Hand bone age: a digital atlas of skeletal maturity. 1st ed. Berlin, Heidelberg: Springer-Verlag, 2005