การศึกษาสภาพปัจจุบัน ปัญหา และความต้องการในการใช้เครื่องมือช่วยวินิจฉัย อายุกระดูกและติดตามการเจริญเติบโตในเด็กไทย

Ratikorn Chaisiwamongkol; Nipaporn Tewattanarat; Wilairat Thawande; Wichuda Chaisiwamongkol; Chanakarn Poonpol; Chatparin Pansukrada; Sarun Paisarnsrisomsuk; Panawit Hanpinitsak; Khwankaow Tangprasert; Pitchaya Wiratchotisatian

Authors

Ratikorn Chaisiwamongkol Department of Pediatrics, Faculty of Medicine, Khon Kaen University, Khon Kaen, 40002, Thailand
Nipaporn Tewattanarat Faculty of Medicine, Khon Kaen University
Wilairat Thawande Faculty of Medicine, Khon Kaen University
Wichuda Chaisiwamongkol Department of Statistics, Faculty of Science, Khon Kaen University
Chanakarn Poonpol Faculty of Science, Khon Kaen University
Chatparin Pansukrada Faculty of Science, Khon Kaen University
Sarun Paisarnsrisomsuk Faculty of Engineering, Khon Kaen University
Panawit Hanpinitsak Faculty of Engineering, Khon Kaen University
Khwankaow Tangprasert Faculty of Engineering, Khon Kaen University
Pitchaya Wiratchotisatian Department of Statistics, Faculty of Science, Khon Kaen University

Keywords:

Bone age assessment, Growth, Secondary sexual development, Diagnostic support tool

Abstract

Background: Bone age assessment (BAA) is essential for diagnosing and monitoring growth in children. While the Greulich-Pyle (GP) method is widely used, it is limited by concerns regarding accuracy and inter-rater variability. Although other countries utilize Artificial Intelligence (AI) to enhance assessment efficiency, Thailand currently lacks data on the current status, problems, and needs for implementing AI in this context.

Objective: The study aims to investigate the current status, problems, and needs for utilizing diagnostic support tools in bone age assessment and growth monitoring.

Methods: This was a qualitative study involving document review and in-depth interviews with two user groups: 6 physicians (3 pediatricians, 3 radiologists) and 6 involved parents. Data were collected at Srinagarind Hospital, Khon Kaen University, and subsequently analyzed using content analysis.

Results: Current Status: All interviewed physicians confirmed using the GP method due to its convenience and speed. Problems: The GP method's reliance on physician experience in certain age ranges causes inter-rater variability. Foreign AI systems like BoneXpert are costly and not yet widespread in Thailand. Concurrently, parents lack knowledge in observing signs of early puberty, leading to delayed detection of growth abnormalities. Needs: Physicians perceive a need to develop AI tailored to the Thai child context to improve assessment accuracy and reduce variability. They propose this AI be integrated into a convenient application for both physicians and parents, capable of assessing bone age, predicting final height, and continuously monitoring growth and secondary sexual development, which aligns with parental needs.

Conclusion: Physicians currently favor the GP method, but its inherent complexity and high variability affect diagnostic accuracy and growth monitoring. A potential and promising approach is the development of AI for BAA, coupled with a suitable and convenient application, to continuously support growth monitoring within the context of Thai children.

References

Cavallo F, Mohn A, Chiarelli F, Giannini C. Evaluation of bone age in children: a mini-review. Front Pediatr 2021.

Prokop-Piotrkowska M, Marszałek-Dziuba K, Moszczynska E, Szalecki M, Jurkiewicz E. Traditional and New Methods of Bone Age Assessment-An Overview. J Clin Res Pediatr Endocrinol. 2021;13(3):251-262.

Yaisilp P, Numsriskulrat N, Sahakitrungruang T. Clinical and epidemiological insights into early puberty in Thai girls: a 5-year study. Ann Pediatr Endocrinol Metab. 2025;30(1):17–24.

Michael DJ, Nelson AC. HANDX: a model-based system for automatic segmentation of bones from digital hand radiographs. IEEE transactions on medical imaging. 1989 Mar;8(1):64-9.

Visiana A/S. BoneXpert User Manual, Version 3.2. Copenhagen, Denmark; 2020. (Available from https://www.visiana.com/bonexpert)

Thodberg HH, Neuhof J, Ranke MB, Jenni OG, Martin DD. Validation of bone age methods by their ability to predict adult height. Horm Res Paediatr. 2010;74(1):15-22.

Greulich WW, Pyle SI. Radiographic atlas of skeletal development of the hand and wrist. Stanford University Press; 1959.

Tanner JM. Growth at Adolescence with a General Consideration on the Effects of Hereditary and Environmental Factors Upon Growth and Maturation from Birth to Maturity. Blackwell; 1966.

Tanner JM. Assessment of skeletal maturity and prediction of adult height: TW2 method. London: Academic Press; 1983. p. 50–106.

Tanner JM, Healy MJR, Goldstein H, Cameron N. Assessment of skeletal maturity and prediction of adult height: TW3 method. London: W.B. Saunders; 2001.

Buken B, Safak AA, Yazici B, Buken E, Mayda AS. Is the assessment of bone age by the Greulich-Pyle method reliable at forensic age estimation for Turkish children? Forensic Sci Int 2007; 173:146–153.

Ranjitkar S, Lin NH, Macdonald R, Taylor JA, Townsend GC. Stature and skeletal maturation of two cohorts of Australian children and young adults over the past two decades. Aust Orthod J 2006; 22:47–58.

Soudack M, Ben-Shlush A, Jacobson J, Raviv-Zilka L, Eshed I, Hamiel O. Bone age in the 21st century: Is Greulich and Pyle’s atlas accurate for Israeli children? Pediatr Radiol 2012; 42:343–348.

Lin NH, Ranjitkar S, Macdonald R, Hughes T, Taylor JA, Townsend GC. New growth references for assessment of stature and skeletal maturation in Australians. Aust Orthod J 2006; 22:1–10.

Zafar AM, Nadeem N, Husen Y, Ahmad MN. An appraisal of Greulich-Pyle atlas for skeletal age assessment in Pakistan. J Pak Med Assoc 2010; 60:552–555.

Goodfellow I, Bengio Y, Courville A, Bengio Y. Deep learning. Cambridge: MIT press; 2016 Nov 18.

Gross GW, Boone JM, Bishop DM. Pediatric skeletal age: determination with neural networks. Radiology 1995;195(3):689–695.

Radiological Society of North America (RSNA). (2017). RSNA Pediatric Bone Age Machine Learning Challenge 2017 [Internet]. Retrieved April 10, 2025, from https://www.rsna.org/rsnai/ai-image-challenge/rsna-pediatric-bone-age-challenge-2017

Cao F, et al. Digital hand atlas and web-based bone age assessment: System design and implementation. Comput Med Imaging Graph 2000;24(5):297–307.

Mehta C, Ayeesha B, Sotakanal A, Desai SD, Ganguly AD, Shetty V. Deep learning framework for automatic bone age assessment. In: 2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC). 2021:3093–3096.

Maratova K, Zemkova D, Sedlak P, et al. A comprehensive validation study of the latest version of BoneXpert on a large cohort of Caucasian children and adolescents. Front Endocrinol (Lausanne). 2023;14: 1130580. Published 2023 Mar 24.

Larrazabal AJ, Nieto N, Peterson V, Milone DH, Ferrante E. Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis. Proceedings of the National Academy of Sciences. 2020 Jun 9;117(23):12592-4.

Seyyed-Kalantari, L., Liu, G., McDermott, M., Chen, I. Y., & Ghassemi, M. (2021). CheXclusion: Fairness gaps in deep chest X-ray classifiers. In Biocomputing 2021: Proceedings of the Pacific Symposium (pp. 232–243).

Seyyed-Kalantari L, Zhang H, McDermott MB, Chen IY, Ghassemi M. Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations. Nature medicine. 2021 Dec;27(12):2176-82.

Bayley N, Pinneau SR. Tables for predicting adult height from skeletal age: Revised for use with the Greulich-Pyle hand standards. J Pediatr. 1952;40(4):423–441.

Marceau K, Abar CC, Jackson KM. Parental Knowledge is a Contextual Amplifier of Associations of Pubertal Maturation and Substance Use. J Youth Adolesc. 2015;44(9):1720-1734.