Measurement of intraarticular tibiofemoral rotation in routine MRI is not reliable: A comparison of knee joints in torsional and routine MRI.

PURPOSE: Intraarticular tibiofemoral (iaTF) rotation at the axial plane of the knee is considered as one parameter of joint alignment. Its measurement is originally described on fully extended legs in torsional computed tomography/magnetic resonance imaging (CT/MRI). The aim of this study was to investigate whether the measurement of iaTF rotation is reliably possible in routine MRIs.

METHODS

Patients who underwent torsional MRI (torMRI) and routine MRI (rouMRI) of the knee were retrospectively analysed. iaTF rotation corresponds to the angle between a tangent to the dorsal femoral condyles and a tangent to the dorsal tibial plateau in the axial plane (positive values indicating external-/negative values internal rotation). In rouMRI, the flexion angle between the bony axes in the sagittal plane was measured. Spearman correlation coefficient (rS) and intraclass correlation coefficient (ICC) were calculated assessing the agreement of iaTF rotation between imaging modalities. Cases showing a deviation of 3° or less were summarised as "group-with-good-agreement". Sensitivity and specificity for assignment to the group-with-good-agreement based on the flexion were calculated using ROC-analysis.

RESULTS

90 cases (78 patients, age 25.6 ± 9.6 years) were included. iaTF rotation was 5.7° ± 6.7° in torMRI and 1.0° ± 8.0° in rouMRI. The measurements differed by an absolute value of 6.4° ± 4.6°. Strong positive correlation (r s = 0.591; p < 0.001) and moderate agreement (ICC = 0.688) were calculated between iaTF rotation measurements. Flexion in rouMRI (7.1 ± 6.1°) correlated moderately positively with the difference in iaTF rotation between torMRI and rouMRI. The group-with-good-agreement showed a significantly lower flexion (p = 0.022). Assuming that lower flexion in rouMRI predicts good agreement, the ROC analysis yielded a threshold of 4.65° flexion (model quality = 0.53). For cases below the threshold, sensitivity was 74.2% and specificity 54.5% for correct assignment to the group-with-good-agreement.

CONCLUSION

Intraarticular TF rotation is not reliable measurable on routine MRIs with no uniform protocol, as the results deviate considerably from those in the torsional MRIs.

LEVEL OF EVIDENCE: Level III.