[Comparison of different bone age assessment methods in 5 573 children aged 3-18 years]

Objective: To compare the differences among Tanner-Whitehouse 3 (TW3), the standards of skeletal maturity of hand and wrist for Chinese-China 05 (China 05), and the standards of skeletal maturity of hand and wrist for Chinese-CHN (CHN) in the practical application of bone age (BA) assessment. Methods: A cross-sectional study was conducted to investigate 6 013 children aged 3-18 years in Tongzhou District, Beijing, from September 2022 to July 2023. Basic information was collected via self-designed questionnaires; physical development indicators such as height and weight were measured on-site; and anteroposterior X-rays of the left hand were obtained. BA was calculated using the TW3 method (including Radius-Ulna-Short bones (RUS) and carpal scoring), China 05 method (including RUS and carpal scoring), and CHN method, respectively. The Wilcoxon signed-rank test was used to compare the difference between the bone age difference (BAD) and zero; the McNemar-Bowker test was employed to compare the differences among 3 BA assessment methods; and the Fleiss Kappa coefficient was used to evaluate consistency. Results: Among the 5 573 participants who met all inclusion and exclusion criteria, there were 2 856 boys and 2 717 girls. The Z-scores for height, weight, body mass index, and target height were 0.48 (-0.22, 1.19), 0.72 (-0.18, 1.78), 0.73 (-0.29, 1.86), and 0.22 (-0.19, 0.64), respectively. For TW3 RUS, BAD was below 0 in boys aged 4-8 years and girls aged 4-7 years (all P<0.05), while it was above 0 in boys aged 11-15 years and girls aged 9-14 years (all P<0.05). For TW3 carpal, BAD was below 0 in boys aged 6-9 years and girls aged 6-7 years (all P<0.001), while it was above 0 in boys aged 11-14 years and girls aged 9-12 years (all P<0.05). For China 05 RUS, BAD was below 0 in boys aged 4 years (P<0.05) and consistently above 0 in boys aged 7-14 years (all P<0.001); in girls, China 05 RUS BAD was below 0 at aged 4-6 years, and above 0 at aged 10-11 years (all P<0.05). For China 05 carpal, BAD was below 0 in boys aged 8-9 and 13 years (all P<0.05), and above 0 in boys aged 12 years ( P<0.05); in girls, BAD was below 0 at aged 6-8 and 11 years (all P<0.05). For CHN, BAD was above 0 across all age groups (all P<0.05) except for children aged 4-6 years and girls aged 17-18 years. The overall Fleiss Kappa coefficients for BA status using the TW3 RUS, China 05 RUS, and CHN standards were 0.54 for boys and 0.51 for girls (all P<0.001), with the lowest consistency observed in aged 12-15 years Fleiss (Kappa coefficients 0.45 for boys and 0.28 for girls, all P<0.001). The overall Fleiss Kappa coefficients for BA assessment using the TW3 carpal, China 05 carpal, and CHN were 0.45 for boys and 0.37 for girls (all P<0.001). At the same skeletal maturity score (SMS), the height of participating children was generally higher than that of the CHN and China 05 reference populations. Specifically, compared with the China 05 reference population, boys with SMS>400 (corresponding to BA>12 years) were 3.3-5.0 cm taller (all P<0.01), and girls with SMS>600 (corresponding to BA>11 years) were 3.1-4.6 cm taller (all P<0.01). Conclusions: The TW3 and CHN BA showed considerable deviations from chronological age (CA), particularly pronounced in pubertal children. Although the China 05 BA exhibited relatively smaller deviations from CA, the height corresponding to a given skeletal maturity level were different; thus, the impact of secular growth trends should be considered when interpreting growth potential based on skeletal maturity. In clinical practice, scientific comprehensive interpretation of BA standards requires the integration of multi-dimensional indicators and dynamic growth trends, so as to provide evidence-based references for diagnosis and treatment of growth disorders.