Magnetic resonance imaging‐based axial‐plane analysis of lower extremity torsion with dedicated 3D planning software is reliable and efficient
Johanna Lippuner, Felix Hüttner, Felix Ferner, Dorothée Sinha, Veronika Weyer‐Elberich, Björn Vogt, Mauricio Kfuri, Mirco Herbort, Jörg HarrerAbstract
Purpose
The reliability and efficiency of a dedicated 3D planning software were evaluated for magnetic resonance imaging (MRI)‐based measurement of femoral torsion, tibial torsion, knee version and whole‐lower extremity torsion compared with manual assessment.
Methods
In this retrospective observational reliability study, torsional MRI scans of both lower extremities from 100 healthy volunteers were analysed. Measurements were performed according to Strecker and Waidelich using manual and digital assessment with mediCAD® 3D. One manual and two digital measurement rounds were performed by one experienced orthopaedic surgeon and one inexperienced orthopaedic resident. Inter and intraobserver reliability was assessed using intraclass correlation coefficients with 95% confidence intervals. Agreement was evaluated using Bland–Altman plots and equivalence testing with a predefined ±3° margin. Measurement times of the first and last 10 measurements of each series were compared.
Results
The cohort included 50 males and 50 females, with a mean age of 26.7 ± 5.8 years. All parameters demonstrated good to very good reliability. Manual interobserver reliability was highest for femoral torsion (ICC: 0.99; 95% CI: 0.98–0.99; p < 0.001). Digital interobserver reliability ranged from ICC 0.94 to 0.98 in the first digital round and from ICC 0.89 to 0.98 in the second digital round, all p < 0.001. Manual‐to‐digital agreement ranged from ICC 0.90 to 0.98 for both observers. Bland–Altman analysis showed mean digital interobserver bias within approximately 2°. All comparisons were equivalent within the predefined ±3° margin. The largest efficiency gain occurred in the inexperienced observer, with median time decreasing from 200 to 150 s, corresponding to a 25% reduction ( p < 0.001).
Conclusion
The 3D planning software demonstrated good to very good inter and intraobserver reliability for MRI‐based lower extremity torsional assessment. It improved efficiency, particularly for the inexperienced observer, and, through automated calculations and guided workflow, supports standardised preoperative planning.
Level of Evidence
Level III.