Diagnostic Test of TW System Bone Age of the Radius Ulna and Short of Bones in Chinese Girls With Idiopathic Precocious Puberty

Abstract

Listen

Translate article

Objective To compare the TW2, TW2CH and TW3 methods in Chinese girls with idiopathic precocious puberty (IPP) for the radius ulna and short bones (RUS) bone age. Methods 55 girls (aged from 3.2 to 9.9 years) with idiopathic precocious puberty were admitted in the Department of Pediatrics and Gynecology and Obstetrics, West China Second University Hospital from January 1995 to August 2002 as the test group. Among them, 36 cases started breast development before age 8 (aged from 1.50 to 7.90 years). 20 girls had menarche before age 10 (aged from 5.00 to 9.75 years). The control group includes: ① 14 girls started menarche after age 10 (11.45±1.05) years; ②25 girls who started menarche aged from 8 to 10 (9.20±0.67) years with lower weight; ③ 32 girls who started menarche aged from 8 to 10 (9.11±0.74) years with normal weight; ④ 12 girls who started menarche aged from 8 to 10 (9.61±0.66) years with overweight. Their left hand-wrist radiographs taken at their first visit were retrospectively analyzed. Bone age was staged single-blind using the radius ulna and short bones. TW2 London standard radius ulna and short bones (TW2), TW2 in the south of China standard radius ulna and short bones (TW2CH) and TW3 systems were used in bone age estimation. Five decision thresholds were recorded for each system. They were >97th percentile, >90th percentile, >75th percentile, >50th percentile and ≤50th percentile. Results The following four thresholds were relatively high in both sensitivity and specificity: ①>90th percentile of TW2CH (Sen=0.836, Spe=0.915), by which there were 16 cases of missed diagnosis and misdiagnosis, with a diagnosis accuracy of 88.41%; ②>90th percentile of TW3 (Sen=72.73%, Spe=86.75%), by which there were 26 cases of missed diagnosis and misdiagnosis, with a diagnosis accuracy of 81.16%; ③>75th percentile of TW2CH, by which there were 28 cases of missed diagnosis and misdiagnosis, with a diagnosis accuracy of 79.71%; ④>75th percentile of TW3, by which there were 30 cases of missed diagnosis and misdiagnosis, with a diagnosis accuracy of 78.26%. Area under the receiver operating characteristic curve (AUR) with TW2CH: 0.942±0.019 (95% CI 0.905~0.978). Area under the receiver operating characteristic curve with TW3: 0.886±0.028 (95% CI 0.831~0.941). Area under the receiver operating characteristic curve with TW2 is the smallest: 0.766±0.040 (95%CI 0.688~0.843). Conclusion To evaluate bone age and its percentile with TW-RUS system, the TW2CH has the highest value in diagnosing idiopathic precocious puberty. TW3 and TW2 systems are moderate. The TW2 Britisher standard radius ulna and short bones is the lowest. Key words: age determination by skeleton; TW2 method; TW3 method; idiopathic precocious puberty; diagnostic test; receiver operating characteristic (ROC) curve