Sparman M, Wolke B, Czupalla H, Banzar D, Zink A. Positioning of total knee arthroplasty with and without navigation support: a prospective, randomised study. J Bone Joint Surg [Br] 2003;85-B:830-5.

Professor Sparmann is one of the pioneers of CAS TKR having been involved in the design, clinical application and evaluation of the technology. His clinical experience in this field is probably unsurpassed. His is an appropriate initial contribution to this series. The assumptions behind this paper are that appropriate alignment of a TKR determines, at least in part, the outcome. The hypothesis that is tested is that computer-assistance provides better alignment outcomes than conventional, mechanical, instrumentation. The methodology used is a prospective, randomised study. The result is that the computer-assisted TKR’s were better aligned in those parameters that could be measured using simple radiological techniques.

It is worth stating at this stage that the ambitions of the first generation of CAS systems were (and still are) quite limited. The technology does not even attempt to produce ‘a perfect TKR’, which currently would be defined as perfectly aligned, perfectly balanced, having a full range of movement and normal kinematics.

This paper is also not the first controlled study of CAS TKR, but it is the largest to date and it does represent the essential opening statement of the clinical application of IT in TKR. It establishes that a TKR implanted using CAS can be better aligned in the coronal plane. However, it leaves a series of questions unanswered. It does not deal with some of the alignment parameters that the CAS system claimed to control. Femoral sagittal and axial parameters, tibial rotation and soft-tissue balance were not measured.

We are left wondering whether the unmeasured parameters are also improved. So in summary this was an important first step on what is clearly a long road.

Chauhan SK, Clark GW, Lloyd S, Scott RG, Breidhal W, Sikorski JM. Computer-assisted total knee replacement: a controlled cadaver study using a multi-parameter quantitative CT assessment of alignment (the Perth CT Protocol). J Bone Joint Surg [Br] 2004;86-B:818-23.

Conceptually this paper takes over from the previous one. It recognises that to test the claims of better alignment in five parameters, all five parameters need to be measured. Conventional plain radiography cannot measure axial (rotational) parameters and has difficulty with the sagittal alignment of the femoral component. The group therefore tackled the problem by designing a CT based protocol for the measurement of all five of the alignment parameters. This protocol is easy to use both for research and as a routine outcome measure. It has already shown its values in the investigation of the unsatisfactory TKR and as a precursor to revision knee arthroplasty. A later randomised controlled study² using the Perth CT Protocol has confirmed that computer assistance provides better alignment in all of the five measured parameters and showed that the greatest contribution of CAS was in the sagittal and axial planes.

Alignment measurement in the early post-operative stage, combined with long term clinical follow-up will provide the answers to a whole range of alignment related questions.

Malalignment of which parameters causes which problems?
Is patello-femoral pain related to femoral alignment?
Which combinations of mal-alignment are the most debilitating?
What are the malalignment thresholds which limit prosthesis longevity?

Sikorski JM. Computer-assisted revision total knee replacement. J Bone Joint Surg [Br] 2004;86-B:510-14.

There are two elements to this paper which take one a little further down the computer-assisted road. Firstly, the Perth CT protocol is used to provide a single cumulative index of TKR alignment, which is easy to calculate and to use. The two part Perth Alignment Index (PAI) provides an estimate of the degree of alignment error and the number of parameters that are malaligned. A perfect outcome is 0:0. A total error of 6° spread among three parameters is 6:3. The PAI is a very useful research and audit tool.

The other feature of this paper is that it takes computer-assistance into the realm of revision knee replacement where large bone deficits exist. It shows that even with extensive bone loss, alignment can be precise and that extensive bone grafting can be done without significant loss of alignment. It is not necessary to use the medullary canals as alignment guides. The additional contribution of computer-assistance is providing an estimate of the joint line, although currently the calculations required to determine this can be complex.

Tokuhara Y, Nakagawa S, Kobayashi A, Takaoka K, Kadoya Y.The flexion gap in normal knees; an MRI study. J Bone Joint Surg [Br] 2004;86-B:1133-6.

Ligament balancing has assumed mythical status in knee replacement surgery. It is a mantra chanted by all the gurus of knee replacement that the perfect knee replacement has to be ‘balanced’ which presumably means having symmetrical soft-tissue tensions in all four quadrants of the knee. Yet there is really no evidence for the proposition. It is impossible to objectively measure ligament balance intra-operatively and there is no literature on the relationship of soft-tissue balance and outcome.

This paper moves us away from the contemporary operating theatre to the future. It also emphasises the increasing importance of MRI (a computer based technology), in understanding the structure and kinematics of the knee. It demonstrates that in the normal, flexed, knee the flexion gap (at about 87°) is asymmetrical. The gap was trapezoidal with a 4.6 mm difference between lateral and medial sides and with a 5° tilt of the femur. Clearly if the flexion gap is asymmetrical then balance between flexion and extension means that the extension gap should also be trapezoidal. This is not the case. Therefore the normal knee is not ‘balanced’. If the normal knee is not balanced does the prosthetic knee have to be?

There are two reasons why ligaments should be at an appropriate tension. The first is help stabilise the joint (actively and passively). The second is to allow uniform and minimal distribution of pressure in the polyethylene of the tibial plateau. Presumably the even distribution of a minimal amount of pressure, compatible with joint stability, will help to minimise wear. We need to know what are the pressure values we should be aiming at. We cannot simply chant ‘ligament balance’ as though it was a sacred text.

The relevance of this to computer-assisted TKR is that the concept of ‘ligament balance’ needs to be re-examined and re-defined. The relationships between gap kinematics, soft-tissue tensions and articular load distributions need to be clarified. This will only occur with accurate intra-operative measurement supplemented by appropriate outcome measurements. Computer based technologies are ideally positioned to make the required advances.

Sugano N, Ochi T, Noble PC, Kamaric E, Salama JK, Tullos HS. The morphology of the femur in developmental dysplasia of the hip. J Bone Joint Surg [Br] 1998;80-B:711-19. The final paper is also about future developments of computer-assisted systems. Perfect mechanical alignment of a TKR requires that the components are aligned along the mechanical axis and that the mechanical axis of the limb falls through the centre of the prosthesis. This is only possible if there is no extra-articular deformity. Yet in the badly deformed knee the deformity is both intra- and extra-articular. We do not know how to deal with this. We do not even define the extra-articular component before we sail blithely into the operating room.

This paper uses CT with three dimensional reconstruction models to define the morphology of the femur in femoral dysplasia. This is as a precursor to total hip replacement with the aim of allowing the surgeon to choose an implant that is appropriate to anatomy. It is a simple step to extrapolate this line of thinking to the grossly deformed knee. Once the three dimensional mechanical alignment of the knee is known then we can face the implications.

Current computer-assisted systems assume that the mechanical axes of the knee have a fixed relationship to the anatomical landmarks that are registered. It is quite possible that with major deformity this is not so. The anteroposterior axis of the femur (Whiteside’s line) certainly becomes distorted with severe osteoarthritis and it is possible that the femoral torsion may occur.

With computer-assisted systems morphological data is gathered in the process of registration. That data is almost always discarded. A bank of such data would be of great value in understanding complex deformities and how they affect surgery and surgical outcome. Once we understand how deformity should be dealt with then we have the basis of sentient computer-assisted system, which will be a major advance on the current purely reactive systems. They will analyse morphology, propose mechanical solutions and then help to achieve the outcome required.

Conclusions

These five papers outline both the progress and the future direction of computer-assisted TKR. The technology has established itself, proven its value in a limited manner and shown some of its potential in difficult revision surgery. It is however in its early stages. It has the capacity to be more than just a substitute for mechanical jigs. It has the potential to analyse deformity, measure polyethylene pressures, suggest alignment compromises and to provide information for an endless cycle of progress.

It is an exciting prospect.

References

1.Clough JFM, Oliver CW.Annotation. Orthopaedics, networks and computers. J Bone Joint Surg [Br] 2002;84-B:481-5.

2.Chauhan SK, Scott RG, Breidahl W, Beaver RJ. Computer-assisted knee arthroplasty versus a conventional jig-based technique: a randomised, prospective trial. J Bone Joint Surg [Br] 2004;86-B:372-7.

Mr Jerzy M Sikorski, Consultant Orthopaedic Surgeon
Hollywood Private Hospital, Nedlands, Western Australia