OBJECTIVE The purpose of this study was to evaluate the interchangeability and reliability of femoral and tibial torsion measurements in children using 3D models based on biplanar radiography compared with CT measurements. MATERIALS AND METHODS Femoral and tibial torsion were measured in 50 patients (mean age, 10.9 years; range, 4.7-14.8 years) using 3D models based on low-dose biplanar radiography by two independent readers. Measurements on transverse CT images by two independent readers served as the reference standard. Intermethod and interreader agreement was calculated using descriptive statistics, intraclass correlation coefficient (ICC), and Bland-Altman analysis. RESULTS Femoral and tibial torsion were -6°-65° and 6°-51° for 3D models based on biplanar radiography and -13°-59° and 4°-52° for CT measurements. The average difference (±SD) between the two methods was 4.9°±3.8° and 5.5°±4.1°, respectively. The intermethod ICC for biplanar radiography was 0.90 (95% CI, 0.87-0.92) for femoral torsion and 0.75 (0.68-0.80) for tibial torsion. The interreader ICC was 0.93-0.97. Mean measurement differences between the two biplanar radiography readers were 3.4° (0.0°-11.0°) for femoral torsion and 3.9° (0.0°-15.0°) for tibial torsion. Mean interreader differences at CT were 3.3° (0.0°-9.0°) for femoral and 3.0° (0.0°-10.0°) for tibial torsion. There was no trend for larger intermethod differences with decreasing age of the children. CONCLUSION Femoral and tibial torsion measurements in children using 3D models based on biplanar radiography are comparable to CT measurement results. Despite skeletal immaturity, torsion measurements in children on biplanar radiography seem to be as reliable as those on CT images.