Introduction: Patellofemoral complications are one of the major causes for revision surgery. In the prosthetic knee, the main determinant within the patellofemoral mechanism is said to be the design of the groove (Kulkarni et al., 2000). Other studies characterising the native trochlear groove used indirect methods such as photography, plain radiographs and measurements using probes and micrometer. The aim of this study was to define the 3-dimensional geometry of the femoral trochlear groove. We used CT scans to describe the geometry of the trochlear groove and its relationship to the tibiofemoral joint in terms of angles and distances.

Materials and Methods: CT scans of 45 normal femurs were analysed using custom designed imaging software. This enabled us to convert the scans to 3D and measure distances and angles. The flexion axis of the tibiofemoral joint was found to be a line connecting the centres of the spheres fitted to posterior femoral condyles. These two centres and the femoral head centre form a frame of reference for reproducible femoral alignment. The trochlear geometry was defined by fitting circles to cross sectional images and spheres to 3D surfaces. Axes were constructed through these centres. The deepest points on the trochlear groove were identified using quad images and Hounsfield units. After aligning the femur using different axes, the location of the groove was examined in relation to the mid plane between the centres of flexion of the condyles.

Results: The deepest points on the trochlear groove can be fitted to a circle with a radius of 23mm (S.D. 4mm) and an R.M.S error of 0.3mm. The groove is positioned laterally (especially in its mid portion) in relation to the femoral mechanical and anatomical axes. It was also lateral to the perpendicular bisect of the transcondylar axes. After aligning the anatomical axis in screen the trochlear groove can be described on average to be linear with less than 2 mm medial/lateral translation.

In the sagital view, the centre of the circle is offset by 21mm (S.D.3mm) at an angle of 67° (S.D. 7°) from a line connecting the midpoint between the centres of the femoral condyles and the femoral head centre.

On either end of this line, the articular surface of the trochlea can be fitted to spheres of radius 30mm (S.D. 6mm) laterally and 27mm (S.D. 5mm) medially, with an rms of 0.4mm.

Discussion: The location and configuration of the inter-condylar groove of the distal femur is clinically significant in the mechanics and pathomechanics of the patellofemoral articulation. This investigation has allowed us to characterise the trochlear groove.

This can be of use in planning and performing joint reconstruction and have implications for the design of patello-femoral replacements and the rules governing their position.

Patellofemoral complications in total knee arthroplasty (TKA) are common. Patellar tracking can be adversely affected by component positioning, soft tissue imbalance and bony deformity. Lateral patellar release rates reported in the literature vary from 6– 40%. Computer assisted surgery has largely been confined to the tibio-femoral component of total knee replacement. However, with recently developed software, it can be used to visualise and quantify patellar tracking and thus guide the precise extent and site of lateral patellar release. The aim of this early study was to define the diagnostic envelope for identification and quantisation of patella maltracking using a current generation patella navigation system.

Our previous prospective analysis of 100 patients undergoing primary TKA identified pre-operative radiographic indices that correlate with maltracking of the patellofemoral joint. 20 cases were subsequently selected for computer assisted total knee replacement surgery. The navigation system (Vector Vision (BrainLab) version 1.6) was used to achieve accurate alignment and position of the femoral and tibial components. All knee replacements were performed using a posterior cruciate-retaining prosthesis. The femoral component was of a ‘patella-friendly’ design with inbuilt 3 degrees external rotation, and the patella was resurfaced in all cases with a biconvex inlay patellar prosthesis.

Patellar tracking was assessed intra-operatively using an additional patellar array and patella tracking-specific software. Real-time displays of patella shift, tilt, rotation and circle radii during multiple flexion-extension cycles were obtained. Where necessary, an ‘outside-to-in’ release of the lateral retinacular complex was performed. The navigation system was used to provide contemporaneous feedback on the effect of the soft tissue releases on the tracking characteristics of the patella component on the prosthetic trochlea. Primary outcomes included the sensitivity and specificity of the system for peri-operative patella maltracking; secondary outcomes included the definition of interventional endpoints and correlation of intra-operative tracking data with post-operative x-rays.

The demographic data for the 20 patients enrolled in this study was essentially unremarkable. As compared to standard intra-operative clinical evaluation of patella tracking, the computer navigation system is equally sensitive and specific, and it can potentially detect more subtle instances of maltracking that may elude conventional clinical evaluation. We present patterns of patellar tracking during the surgery for patient with and without pre-operative patellar maltracking. However, the significance of this is unknown without longer-term outcome data. Patella shift abnormalities that were detected by the system, but not tilt, correlated with clinical judgement of patella maltracking (p< 0.05).

Soft tissue balancing of the patella can now be performed by observing precise changes in shift and tilt. This can be as important as component alignment for optimising patellar tracking and minimising patellofemoral complications.

Patellofemoral symptoms are a prominent cause of dissatisfaction following knee arthroplasty. This may relate to difficulty in knowing where to resect the bone and in placing prosthetic components to reproduce the anatomy accurately. This study developed geometrical data to facilitate these procedures during TKR.

Thirty CT scans of patients above the age of 55 without patellofemoral disease were performed. Three dimensional images were reconstructed using computer software that enabled manipulation of these images and measurements to be taken. These models allowed the shape of the patella to be modelled, its size and the track it takes in the normal trochlea.

The anterior and proximal patellar planes could be described as flat surfaces with an rms of 0.4 and 0.3mm. The angle between these planes was 112° (stdev 5°). The median ridge of the articular surface was a straight line with an rms of 0.2mm and the average angle between the anterior plane and this line was 12° (stdev4°). The angle between the anterior plane and a line fitted to the posterior aspect of the apex of the patella was 56° (stdev 2°). Having oriented the patella with the proximal plane vertical, the distal pole of the patella was within 2mm of the same sagittal plane as the median ridge of the articular surface in all cases. The functional centre of the patella was defined as a point in the centre of 2 planes orthogonal to the sagittal plane at the midpoint between the most proximal and most distal points on the median ridge. In the transverse section this centre was always on the line separating the superficial and deep surfaces of the patella. Also the length, width and thickness of the patellae were measured at 22mm +/−4mm, 47mm +/− 3mm and 24 mm+/− 2 mm. The average ratio of the lateral facet to medial facet width was 1.3 (range 0.8–1.6). The average ratio of the patellar width to thickness was 2.0 (S.D. 0.106, 95%CI 1.96 to 2.03) with a strong correlation(r= 0.89).

From this work we have concluded that the anterior and proximal planes of the patella, which will not be affected by the disease, can be defined and used as a frame of reference for the patella, which will be helpful for navigating the patella and restoring its anatomical form in the presence of erosive changes.

The patella has a constant shape, so that its articular surface can be defined in relatively simple terms, and can be referenced off its non articular surface.