Keywords:
Performed at one institution, Cross-sectional study, Retrospective, Quality assurance, Technology assessment, Conventional radiography, Paediatric, Bones
Authors:
L. Sobak1, M. Lintrop2; 1TALLINN/EE, 2Tartu/EE
DOI:
10.26044/ecr2020/C-12097
Methods and materials
215 radiographs of the left hand from Caucasian children between the ages from 2 to 16,6 years with various diagnoses were retrospectively reassessed using BoneXpert version 2.4.5.1 (Figure 1). The results from automated method were then compared with the pediatric radiologists reports, who used Tanner-Whitehouse 3 (TW3) [2] method. 20 films were excluded from analysis, because the evaluation was not possible - in 15 cases the age was too low (the bone age was lower than 2 years); in 1 case the age was too high (the bone age was higher that 16 years); in 1 case there was wrong hand pose; in 1 case bones had abnormal shape (the suspected diagnose was Turner syndrome or pseudohypoparathyreoidism); in 1 case there was a foreign object (a cannula) on the film; in 1 case the radiologists report was incomplete. 195 radiographs from children between the ages from 2 to 16,6 years were included in the final analysis. Results were evaluated using Intraclass correlation and paired t-test analysis.