Computer-aided estimation of skeletal age and comparison with bone age evaluations by the method of Greulich-Pyle and Tanner-Whitehouse

Bone age (BA) is usually estimated using the atlas of Greulich-Pyle (GP) or Tanner-Whitehouse (TW) and depends on the individual experience of the investigator. A computer-aided method, computer-assisted skeletal age scores (CASAS), based on the TW2-radius, ulna, short bones (RUS) method has been created to increase reliability and validity. We compared the results of the three different methods (CASAS, GP, TW2) in three groups of children with Turner's syndrome (TS), growth hormone deficiency (GHD), and familial short stature (FSS). The practicability and reliability of CASAS was investigated and the results compared with those obtained by the other methods. Each method was applied by one investigator, with up to six consecutive BA estimations per subject being carried out in 5 patients with TS, 6 with GHD and 18 with FSS. Using CASAS, individual bone evaluations had to be repeated once in 7.3% and twice in 2.7% of all probands on request of the computer system because of doubtful results. Manual interventions by the investigator were necessary in 12.3% of evaluations in TS, 8.5% in GHD and 9.0% in FSS. The frequency of a ''warning'' insert, indicating uncertainty of CASAS, was also higher in TS than in GHD and FSS (19.0% vs. 15.0% and 13.0%). The majority of external corrections for CASAS were necessary for evaluations of the fifth finger and the thumb. On three occasions with TS the progress of BA determined by CASAS demonstrated a regressive course with age. CASAS and manual TW2 BA data were comparable and generally higher than BA data obtained by GP (mean + 1.1 years). In conclusion, CASAS represents a useful method for analysing skeletal age and seems to increase reliability by rating on a continuous scale. However, difficulties with abnormally shaped banes restricts its use in some pathologic conditions.