MPR realignment increases accuracy when measuring femoral neck anteversion angle.

Abstract

To compare two methods of measuring femoral neck anteversion angle (FNA): A 2D method used at Odense University Hospital until 2010, and a method labeled 3D-OUH. The latter method makes corrections to compensate for errors introduced by the individual placement of patients in the CT scanner. Twenty-six CT-examined patients were included: nine men and 17 women. The right side FNA was measured twice with each method by one observer, measuring intraobserver variability. Both methods are based on the following anatomy: femoral head center, center at the level of lesser trochanter and posterior apex of the femoral condyles. The 3D-OUH method corrects for the individual orientation of femur by realigning it prior to measurement, in accordance to Murphy et al.'s original definition of FNA. The intercondylar notch center of the knee and center at lesser trochanter was used in the realignment. The 2D method significantly overestimated FNA compared to 3D-OUH by 4.2° (95 % CI: 2.8°; 5.6°), p < 0.0001. All measurements with the 3D method needed clock-wise correction in the coronal plane, suggesting patient positioning as a consistent source of overestimation by the 2D method. The 3D-OUH method had a lower intraobserver variability with a limit of agreement (LOA) of -2.4° to 2.1° against that of the 2D method of -3.4° to 3.8° Mean anteversion was 4.2° (95 % CI: 2.8°; 5.6°) lower with the 3D-OUH method than with the 2D method. The 3D-OUH method eliminated an obvious source of error, namely the individual orientation of femur during CT-examination. Moreover, intraobserver variability was improved with the 3D-OUH method.