Introduction

Correct postoperative leg length restoration is among the most important goals of hip arthroplasty. Therefore, we developed, validated and clinically applied a novel software algorithm based on surgical navigation, which allows the surgeon to restore a defined femur position without establishing a femoral coordinate system or the hip joint center and measure the leg length accurately and simply.

There is a complex interaction among acetabular component position and the orientation of the femoral component in determining the maximum, impingement-free prosthetic range of motion (ROM) in total hip arthroplasty (THA). Regarding restrictions in ROM, femoral antetorsion is one of the most important parameters. But, ROM is also influenced by parameters like the deviation between the femoral shaft and the mechanical axis in a sagittal projection. This deviation is best described as “Femoral Tilt” (FT). This study analysis the incidence of FT in clinical practice and its consequences on post-operative ROM. Based on these results, the effects of changes in FT on ROM-based cup optimisation are assessed by a using a virtual ROM analysis.

For studying the incidence of FT, 40 (16 male, 24 female) postoperative computerised tomography (CT) scans were analysed using a 3D CT planning software. The implant models were superimposed onto the image data to determine their exact position. The anatomical orientations were determined by planning anatomical landmarks and coordinate directions (i.e. mechanical axis, posterior condyle axis). Descriptive statistics were calculated for FT. Effects of changes in FT and CCD on ROM were analysed by calculating zones of compliance. FT was varied between 2.1° and 9.3° for 135°.

The overall range of post-operative values for femoral tilt was 5.7° ± 1.8° (mean ± standard deviation, minimum 1.7°, maximum 10.2°). The zone of compliance significantly depended on FT (difference more than 200%). The optimum cup position changed from 35° radiographic inclination/30° anteversion to 39°/30° when FT was increased from 2.1° to 9.3°.

Within this study, it was demonstrated that FT has a significant effect on postoperative ROM in THAs. First of all, it was shown that clinically FT values lie in a range between 2.1° and 9.3° (95% CI), where we used a long-shaft stem type with a relatively low possibility to influence sagittal tilt angles. FT may significantly change zones of compliance up to 200% as well as optimised cup positions. Thus, standard combined anteversion formulas, which were proposed in the literature to implement femur first approaches for THA, do only particularly address an optimisation of post-operative ROM. Instead, a sophisticated virtual ROM analysis based on a navigated femur-first approach would enable accurate ROM estimations as parameters like FT are hard to be assessed intra-operatively.

Incorrect restoration of leg length (LL) and offset is a major source of patient dissatisfaction and dysfunction after total hip arthroplasties (THAs). Evaluations on anterior-posterior x-ray images are state-of-the-art to assess the accuracy of such techniques. However, x-ray based measurements of LL and offset are challenging and limited in terms of accuracy. Within this study, we evaluated the accuracy of such measurements by analysing a series of clinical data. We evaluated the results on the non-treated side, since we know that there should be no significant difference between pre- and postoperative measurements on this side.

A series of 44 consecutive patients was analysed regarding changes in the difference between pre- and post-operative LL and offset measurements. Anterior-posterior x-rays were taken pre- (pre-OP) and post-operatively (post-OP) with a calibration by a scaling ruler (pre-OP) or implant size (post-OP). The LL and offset measurements were performed with a digital planning software based on the teardrop and transischial line. The distance between the teardrop/transischial line and the trochanter minor was measured to assess LL differences. Femoral offset (FO) was calculated as the orthogonal distance between the centre of the femoral head and the proximal shaft axis. Global offset (GO) was calculated as the distance between the inferior aspect of the teardrop figure and the shaft axis along the teardrop line. Descriptive statistics (mean value ± standard deviation) were calculated for the different types of measurements. Statistically significant differences were checked according to a student's t-test (α = 0.05).

The differences between the pre-and post-operative situation was 0.8±3.2 mm for LL, 0.2±3.5 mm for GO, and −0.5±2.5 mm for FO when referencing to the teardrop line and 0.9±4.0 mm (LL) and −0.3±2.7 mm (FO) for the transischial line. The error distributions did not show statistically significant differences when referencing to the teardrop or transischial line. But high differences (0.1±6.6 mm) were found when comparing the LL values (teardrop vs. transischial) case-by-case.

Within this study we demonstrated that x-ray based offset and LL measurements show reasonable inaccuracies. X-ray based evaluations of navigation-based techniques to assist LL and offset restoration cannot produce significantly better results than these analysed limits. That is, even if the navigation technique would be perfectly accurate, the evaluation would not achieve better accuracies than approximately ±3.5 mm for LL, ±3.5 mm for GO, and ±2.5 mm for FO.

Limited postoperative range-of-motion (ROM) can lead to patient dissatisfaction and dislocation in total hip arthroplasties (THAs). To avoid this, femur first approaches have been developed which optimise particular aspects of ROM by using a virtual analysis of ROM. This study analysis whether it is possible to accurately assess ROM based on an intra-operative acquisition of anatomical structures by using an image-free navigation system. It compares the outcome of a collision detection algorithm when using 3d models from computerised tomography (CT) scans on the one side and intra-operatively acquired 3D models on the other side within a cadaver study. It focuses on peri-acetabular impingements.

During the cadaver session 14 hips (7 cadavers) were treated surgically by using press-fit implants. 3D models of the pelvis and femora were generated based on segmented pre-operative CT data sets. Intra-operative data acquisition was performed by using a CT-free navigation software. Beside standard landmarks, points at the acetabular rim and femoral resection plane were acquired. For assessing ROM, a 3D model of the pelvis was generated. The information about the femoral resection plane was directly entered into the collision detection algorithm. 3D Computer Aided Design (CAD) models provided by the implant manufacturer were used for the implants. Based on this setup, the ROM values for flexion (FLEX), external rotation at 0° flexion (EXT), and internal rotation at 90° flexion (INTROT90) were compared. Differences within intended ROM were considered relevant, since the goal was to enable the prevention of clinically relevant ROM limitations.

The average difference between the CT based and navigation data based ROM analysis was 2.13° ± 3.11° for FLEX, 3.33° ± 5.51° for EXT, and 1.6° ± 3.66° INTROT90. The values reduce to 1.58° ± 2.78° (FLEX) and 0.91° ± 3.77° (INTROT90) when only ROM values within the intended ROM are considered. For EXT all ROM values lied above the threshold for intended ROM. Thus, no relevant differences were found for this motion direction.

In this study, a real-time collision detection based approach was developed and evaluated, which allows to virtually detect prosthetic and bony impingements. It was shown that ROM can be assessed accurately based on an image-free navigation technique. This information can be used intra-operatively to adjust the position of the implants and thus avoid postoperative ROM limitations. In particular, it enables a comprehensive femur first approach which allows us to optimise the post-operative results regarding functional parameters like ROM.