This study aimed to assess the effect of flexion and external rotation on measurement of femoral offset (FO), greater trochanter to femoral head centre (GT-FHC) distance, and neck shaft angle (NSA). Three-dimensional femoral shapes (n=100) were generated by statistical shape modelling from 47 CT-segmented right femora. Combined rotations in the range of 0–50° external and 0–50° flexion (in 10° increments) were applied to each femur after they were neutralised (defined as neck and proximal shaft axis parallel with detector plane). Each shape was projected to create 2D images representing radiographs of the proximal femora.

As already known, external rotation resulted in a significant error in measuring FO but flexion alone had no impact. Individually, neither flexion nor external rotation had any impact on GT-FHC but, for example, 30° of flexion combined with 50°of external rotation resulted in an 18.6mm change in height. NSA averaged 125° in neutral with external rotation resulting in a moderate increase and flexion on its own a moderate decrease. However, 50° degrees of both produced an almost 30 degree increase in NSA.

In conclusion, although the relationship between external rotation and FO is appreciated, the impact of flexion with external rotation is not. This combination results in apparent reduced FO, a high femoral head centre and an increased NSA. Femoral components with NSAs of 130° or 135° may historically have been based on X-ray misinterpretation. This work demonstrates that 2D to 3D reconstruction of the proximal femur in pre-op planning is a challenge.

Unknown femur orientation during X-ray imaging may cause inaccurate radiographic measurements. The aim of this study was to assess the effect of 3D femur orientation during radiographic imaging on the measurement of greater trochanter to femoral head centre (GT-FHC) distance.

Three-dimensional femoral shapes (n=100) of unknown gender were generated using a statistical shape model based on a training data of 47 CT segmented femora. Rotations in the range of 0° internal to 50° external and 50° of flexion to 0° of extension (at 10 degree increments) were applied to each femur. A ray tracing algorithm was then used to create 2D images representing radiographs of the femora in known 3D orientations. The GT-FHC distance was then measured automatically by identifying the femoral head, shaft axis and tip of greater trochanter.

Uniaxial rotations had little impact on the measurement with mean absolute error of 0.6 mm and 3.1 mm for 50° for pure external rotation and 50° pure flexion, respectively. Combined flexion and external rotation yielded more significant errors with 10° around each axis introducing a mean error of 3.6 mm and 20° showing an average error of 8.8 mm (Figure 1.). In the cohort we studied, when the femur was in neutral orientation, the tip of greater trochanter was never below the femoral head centre.

Greater trochanter to femoral head centre measurement was insensitive to rotations around a single axis (i.e. flexion or external rotation). Modest combined rotations caused the tip of greater trochanter to appear more distal in 2D and led to deviation from the true value. This study suggests that a radiograph with the greater trochanter appearing below femoral head centre may have been acquired with 3D rotation of the femur.

For any figures or tables, please contact the authors directly by clicking on ‘Info & Metrics’ above to access author contact details.

Anterior-posterior (AP) x-rays are routinely taken following total hip replacement to assess placement and orientation of implanted components. Pelvic orientation at the time of an AP x-ray can influence projected implant orientation.¹However, the extent of pelvic orientation varies between patients.²Without compensation for patient specific pelvic orientation, misleading measurements for implant orientation may be obtained. These measurements are used as indicators for post-operative dislocation stability and range of motion. Errors in which could result in differences between expectations and the true outcome achieved. The aim of this research was to develop a tool that could be utilised to determine pelvic orientation from an AP x-ray.

An algorithm based on comparing projections of a statistical shape model of the pelvis (n=20) with the target X-ray was developed in MATLAB. For each iteration, the average shape was adjusted, rotated (to account for patient-specific pelvic orientation), projected onto a 2D plane, and the simulated outline determined. With respect to rotation, the pelvis was allowed to rotate about its transverse axis (pelvic flexion/extension) and anterior-posterior axis (pelvic adduction/abduction). Minimum root mean square error between the outline of the pelvis from the X-ray and the projected shape model outline was used to select final values for flexion and adduction. To test the algorithm, virtual X-rays (n=6) of different pelvis in known orientations were created using the algorithm described by Freud et al.³The true pelvic orientation for each case was randomly generated. Angular error was defined as the difference between the true pelvic orientation and that selected by the algorithm.

Initial testing has exhibited similar accuracy in determining true pelvic flexion (x̄_error = 2.74°, σ_error=±2.21°) and true pelvic adduction (x̄_error = 2.38°, σ_error=±1.76°). For both pelvic flexion and adduction the maximum angular error observed was 5.62°. The minimum angular error for pelvic flexion was 0.37°, whilst for pelvic adduction it was 1.08°.

Although the algorithm is still under development, the low mean, maximum, and standard deviations of error from initial testing indicate the approach is promising. Ongoing work will involve the use of additional landmarks for registration and training shapes to improve the shape model. This tool will allow surgeons to more accurately determine true acetabular orientation relative to the pelvis without the use of additional x-ray views or CT scans. In turn, this will help improve diagnoses of post-operative range of motion and dislocation stability.

The main mode of failure of the acetabular component in total hip arthroplasty is aseptic loosening. Successive generations of cementation techniques have evolved to alleviate this problem.

This paper evaluates one such method, Negative Pressure Intrusion cementation. Two groups of machined bovine cancellous bone samples were created; experimental (n = 26) and control (n = 26). The experimental group was cemented using the negative pressure technique and control group was cemented in the absence of negative pressure. The relative cement intrusion depths were then assessed for each group using MicroCT. These samples were then further machined and tested to failure in torsion to estimate their mechanical properties.

Results show mean cement intrusion depth for the negative pressure group to be 8676μm and 6042 μm for the control group (p = 0.078). Mechanical testing also revealed a greater mean torque in the negative pressure group (1.6223Nm vs 1.2063Nm) (p = 0.095).

This work quantifies the effect of negative intra-osseous pressure on cement intrusion depth in cancellous bone and for the first time relates this to increased mechanical strength.

The cement-in-cement femoral revision is a possible method of reducing complications. During recent research on this revision it was observed that a number of the inner cement contained macropores. It was hypothesized that porosity of the mantle influenced the subsidence and inducible displacement of the revision stems. The aim was to calculate the porosity and assess its relationship to the above factors.

Primary cement mantles were formed by cementing a stem into sections of tubular steel. At this stage, the specimen was chosen to be in a test or a control group. If in the test group, it underwent a fatigue of 1 million cycles. This was carried out in a fatigue machine mounted with a specifically designed rig. If in the control group, no such fatigue was undertaken. Into these fatigued and unfatigued mantles, the cement-in-cement procedure was performed. Both groups underwent a fatigue of again 1 million cycles. Subsidence and inducible displacement was recorded. The composites were then sectioned and photographed. The images underwent image analysis to calculate the porosity.

Multiple regression and a general linear model showed subsidence was inversely correlated to the porosity of the “fresh cement” in Gruen zones 3 and 5 (p = 0.021, R2 = 0.36). This relationship was not expected. The reason could be related to the fact that the migration of the stems in each separate direction was not monitored. Inducible displacement was inversely correlated to porosity of the inner cement, again in Gruen zones 3 and 5 (p = 0.001, R2 = 0.61). A possible explanation is that the stem was able to subside more due to the higher porosity and find a more stable position.

The subsidence and inducible displacement of these stems is influenced by porosity, specifically by the porosity of the distal inner cement.

One method of reducing intra-operative complications in revision hip surgery is the cement-in-cement technique. Some concern exists regarding the retention of the existing fatigued cement mantle. It was hypothesised that leaving the existing fatigued cement mantle does not degrade the mechanical properties of the cement in cement revision construct. The aim of this research was to test this hypothesis using in vitro fatigue testing of analogue cement in cement constructs.

Primary cement mantles were formed by cementing a large polished stem into sections of tubular stainless steel using polymethylmethacrylate with Gentamicin. At this stage, the specimen was chosen to be in the test group or the control group. If in the test group, it underwent a fatigue of 1 million cycles. This was carried out in a specifically designed rig and a fatigue testing machine. Into these fatigued and unfatigued primary mantles, the cement in cement procedure was carried out. Both groups underwent a fatigue of again 1 million cycles. Subsidence of the stems and their inducible displacement was recorded. A power calculation preceded testing.

Completion of a Mann Whitney test on the endpoints of the subsidence curves revealed that there is no statistical difference between the data sets (means 0.51, 0.46, n=10 + 10, p = 0.496). This data was also calculated for the inducible displacement. Again, there was no statistical difference in the separate groups for this parameter (means 0.38, 0.36, p = 0.96). This methodology produces a complex 3 dimensional reconstruction of the cement in cement revision which replicates the in vivo structure. This reconstruction has undergone fatigue testing. Neither of these two aspects has been produced for the study of cement in cement revision before.

A fatigued primary cement mantle does not appear to degrade the mechanical properties of the cement in cement revision construct

Computational modelling has the potential of becoming a useful tool in assessing revision risk on a patient-specific basis. However, there are many difficulties encountered in generating subject-specific computational models that have unknown influences on such predictions, e.g. accuracy of the anatomical geometry and material properties of the patient. This study compares the influence of these two patient-specific parameters on predictions of revision risk due to aseptic loosening.

First, X-rays from seventeen patients were processed using previously developed technique utilising rigid scaling of a generic femur to match selected dimensions from each patient’s post-operative X-ray and, then, the same set of 3D models was obtained by using an automated technique that generates 3D extra-cortical geometries from planar X-rays using a combination of 2D contour extraction and 3D warping of a generic model to match the extracted contour.

A cement and cement-metal interfacial damage accumulation algorithm developed previously was used. For each geometric set two types of simulations were performed. First, constant cortical and cancellous bone apparent Young’s moduli were assumed. A second set of simulations used age-dependent Young’s moduli for each bone type. Walking and stair-climbing activities were simulated. Resultant migration of the prostheses was used to indicate revision risk.

Factorial analysis has shown that the geometry has a larger influence on resultant migration magnitude for each case; however, unexpectedly, using more realistic geometry weakened the strength of predictions. This is most likely to be due ongoing mesh-induced contact problems.

Aseptic loosening can be considered as a combination of both mechanical and biological failure scenarios. This study investigated the influence of including bone remodelling in the simulation of aseptic loosening of cemented hip prostheses.

A combined strain and damage stimulated bone adaptation algorithm (Mulvihill et al., Proc. ESB Summer Workshop, p.114–115, 2007) was modified for use on an apparent tissue level. Constant rate resorption or deposition occurs if local strain falls outside a quiescent reference strain range. Furthermore, damage accumulates as a function of tensile stress. Resorption and simultaneous repair is activated above a critical damage level. Model parameters are related to specific surface area expressed as a function of apparent tissue density. Elastic modulus was also a function of accumulated damage. This algorithm was applied in conjunction with a bone cement and cement-metal interfacial damage accumulation algorithm to simulate aseptic loosening for a retrospective dataset of early revision and long-term-unrevised patients (Lennon et al. JOR, 779-88, 2007). One year of walking activity was simulated and resultant migrations of the prostheses were used to indicate revision risk.

The current implementation demonstrated increased migration for simulations with bone remodelling (p= 0.01). Variability was increased but mean predicted migration for early revisions was significantly higher than for the unrevised group (p= 0.03). Bulk bone remodelling was predicted primarily in the proximal regions. Interfacial bone remodelling demonstrated oscillation in damage at the interface due to alternate resorption-repair and deposition cycles. Interfacial bone density changes were more prominent in proximal regions but some models did show small amounts of resorption in more distal Gruen zones.

We conclude that bone remodelling has potential to predict more realistic migration patterns but further development and assessment is needed to identify the correct parameters for the bone adaptation algorithm.