Tibial tuberosity to trochlear groove distance (TT-TG) has been reported to have different values according to imaging modalities, usually higher in computed tomography (CT) than in magnetic resonance imaging (MRI). This difference is thought to be caused by the degree of knee flexion during imaging, but few studies have aimed to elucidate the cause. Five hundred eight patients with knee CT or MRI performed between ages of 6 to 16 years without underlying diseases affecting the musculoskeletal system were included. This study was conducted in 2 statistical ways. (1) Propensity score matching was performed for the imaging modality, and the bony TT-TG was compared between the 2 matched groups. (2) A regression model was fitted with 484 patients with either CT or MRI (a training set), and validation of the fitted model was performed with 24 patients with both CT and MRI simultaneously taken within a week (a test set). The predicted TT-TG values were compared with the measured values. (1) Eighty-one patients were successfully matched by propensity score (all the standardized mean differences < 0.1) for each modality. In the matched patients, there was no significant difference in TT-TG according to the imaging modality (11.3 ± 3.7mm for CT, 10.4 ± 3.8mm for MRI, P = 0.126). (2) For the model fitting, different linear models were fitted before and after 10 degrees of knee flexion angle because there was a sharp change in TT-TG when the knee flexion angle was <10 degrees. The predicted TT-TG values did not significantly differ from the measured values (10.2 ± 4.3 mm vs. 9.0 ± 5.1mm, P = 0.124). This study is the first to statistically prove that the difference between TT-TG in MRI and CT originates from the different degrees of knee flexion. In addition, although more studies are needed, authors recommend imaging to be performed with the knee flexed at least 10 degrees for more reliable measurements because TT-TG changes sharply if the knee flexion angle is <10 degrees. III.