Rotation matters: CT and MRI yield different tibiofemoral rotation angles in patellofemoral patients. An investigational plateau‐anchored MRI method is descriptively associated with a smaller mean CT–MRI difference

Abstract Purpose To compare tibiofemoral rotation (TFR) measured on magnetic resonance imaging (MRI) and computed tomography (CT) in patellofemoral (PF) patients and descriptively report mean differences using an alternative plateau‐anchored MRI measurement method. Methods Surgical candidates for tibial derotational osteotomy with both CT and MRI were retrospectively reviewed. Demographics/surgical indication/tibial torsion were recorded. TFR was measured on CT/MRI using standard posterior condylar–posterior tibial axis technique. A plateau‐anchored MRI method utilizes the angle between the distal femoral posterior condylar axis and the axis connecting the proximal tibial plateau's most medial and lateral aspects. Paired t tests compared TFR between modalities and MRI methods; subgroup analyses evaluated diagnosis. Pearson correlation assessed association between CT‐based TFR and tibial torsion. Results Fifty‐eight knees/46 patients (50 female; mean age 21 ± 7 years; body mass index [BMI] 25 ± 6 kg/m 2 ) were analysed. Indications: PF instability ± pain: 20 knees, pain without instability: 38 knees. Mean MRI TFR: 2.11°; CT TFR 8.28°; plateau‐anchored MRI method 9.80°. Mean difference between standard MRI and CT was −6.18° (95% confidence interval [CI] −7.48 to −4.88; p < 0.001); the difference between plateau‐anchored MRI and CT was 1.51° (95% CI 0.01–3.01; p = 0.048). Bland‐Altman plots showed good inter‐ and intra‐observer agreement. In exploratory subgroup analyses (PF instability ± pain vs. pain without instability), no statistically significant differences were observed in the CT–MRI mean TFR differences. Tibial torsion (38.9° ± 7.0°) did not correlate significantly with CT‐based TFR ( r = 0.25; p = 0.06). Conclusion CT and MRI yield systematically different TFR values in the same PF patients; standard MRI yielding lower values than CT. A plateau‐anchored MRI method yielded a smaller mean difference relative to CT than the standard MRI method in this cohort. These results should not be interpreted as demonstrating interchangeability, accuracy, or clinical readiness of the plateau‐anchored method. Level of Evidence Level II.