Implant Movement Analysis (IMA), A New CT Based Technique for Diagnosis of Aseptic Loosening of Total Knee Arthroplasty

Background and purpose: Aseptic loosening is the most common indication for revision surgery after Total Knee Arthroplasty (TKA). However, the diagnosis is often difficult from planar radiography as well as other methods, making decision regarding revision surgery problematic. Implant Movement Analysis (IMA) is a newly introduced method of analysing uncertain aseptic loosening of orthopaedic implants. This is the first study where this technique is applied for analysis of potentially loose knee prosthesis. The IMA technique is based on Computer Tomography (CT) with rotational as well as varus and valgus provocation to expose any movement of the implant relatively to the surrounding bone. Our aim was to evaluate the role of IMA as a complementary method to planar radiography and clinical examination in diagnosing aseptic loosening of (TKA). Methods: 40 patients with uncertain aseptic loosening TKA were analysed. All patients had symptoms that could be suspicious for aseptic loosening. Planar radiography was used to detect radiolucent lines. IMA was performed on all patients with rotational as well as varus – valgus provocation. Results: Of the 40 knees planar radiography showed radiolucent lines in 22. Out of these 22 knees the IMA analysis showed loosening in 8. IMA showed no loosening in the17 knees without radiolucent lines. Interpretation: IMA is a dynamic investigation where CT is used during forced provocation of the knee and there by loosening between bone and implant can be more accurate diagnosed. Many suspected TKA loosening’s could be ruled out using this method and revision surgery avoided. When planar radiography shows no radiolucent lines the risk of a loose implant is very small.

Some of the methods developed to improve imaging analysis of aseptic loosening have been based on converting two-dimensional information from planar radiography into Three-Dimensional (3D) information. One of these methods is called Radiostereometric Analysis (RSA) and is the current gold standard regarding migration measurements [9,10,14,19,20]. However, the RSA technique requires the implantation of small metal (tantalum) balls during the primary operation. Therefore, RSA is manly suitable for small groups in research settings. The RSA technique is also difficult to apply on TKA since the femoral component is big and almost surrounds the distal bone of the femur making the tantalum balls hard to detect on planar radiography. To improve imaging diagnostics and find more efficient alternatives, studies have been done using direct 3D imaging diagnostic methods as Magnetic Resonance Imaging (MRI) or Computer Tomography (CT) [11][12][13]. The focus of this study has been a new CT-based method that recently moved pass the point of experimental testing and was put into clinical use.
The new CT method, Implant Movement Analysis (IMA, Sectra Inc., Linköping) is based on combining two CT volumes with torsion as well as angular loading of the affected knee in forced external and internal rotation as well as varus and valgus acquired in direct sequence. By rotating the knee as well as forcing it in varus and valgus compression this method visualises the implant's movement relative to surrounding bone structures [17,18,21,22]. IMA was developed by Olivecrona et al. [22] at Karolinska Institutet and was in collaboration introduced at Örebro University Hospital, Lindesberg in 2017. It was first used for potentially loose hip implants, but since 2018 also for TKA.
Previously most decisions regarding revision surgery have, in uncertain cases of aseptic loosening, been based on clinical examination, planar radiography and factors including patient age, comorbidity, extent of symptoms and the patient's willingness to go through surgery ( Figure 1). Since the introduction of IMA additional imaging information is provided which is believed to improve diagnosing of uncertain cases. This might reduce the number of unnecessary revision surgeries, meaning in cases where loosening was suspected on planar radiography, but the prostheses were in fact in place. The present study is the first study worldwide regarding IMA for TKA. Methods 40 patients were admitted to the X-ray department in Lindesberg, Sweden for IMA investigation due to discomfort in their knees that potentially could be because of loosening of their TKA. Radiographic examinations were done according to clinical praxis. All IMA examinations were conducted by two radiographers as described in accordance with instructions given by Henrik Olivecrona, one of the developers of IMA (Figure 2 & 3). IMA images were analyse by a specially trained radiologist (SK). IMA analysis was not blinded to results from planar radiographic examinations. The software used during IMA examinations is based on methods developed by Olivecrona et al. [21,22].  Paired examination results from planar radiography and IMA were analysed. In cases were more than one planar radiography examination had been performed previous to IMA, results from the most recent examination were used. Results regarding aseptic loosening were categorized as aseptic loosening; present or not present. These distinctions were primary made by a specially trained radiologist (SK). For study participants who following IMA-examination had surgery, operative reports were also studied. This study was part of a research project approved by the Regional Ethical Review Board in Uppsala (ref. 2019-00117).

consecutive patients were examined between January 2019
December 2020. 23 women and 17 men were included with a mean age at the time of IMA examination of 68 years (range 54 -86). 37 patients were examined after primary TKA and 3 patients after a previous revision. The mean time between the latest TKA surgery and the IMA analysis was 7.5 years (range 2 -21).
Of the 40 TKAs with uncertain loosening 22 were classified as loose on planar radiography mainly based on the presence of radiolucent lines. Out of these 22 knees 8 were detected as loose and 14 as stable with use of IMA. Of the 18 knees that were assessed as stable with planar radiography, all were also stable on the IMA analysis (Table 1 & 2).  The 8 patients that were analysed with a loose implant using IMA were operated with revision surgery and loosening of at least one component was confirmed during surgery.

Discussion
Revision surgery after TKA is difficult and with a high risk of complications [23,24]. Therefore, decision regarding revision surgery should be done with care and the preoperative evaluation is of great importance. When a patient complains about pain and discomfort some years after TKA the reasons for this can be of many different kind [5]. It is important to try to diagnose the main reason to provide the best possible treatment. One common reason for late onset problems after TKA is loosening of the implant and if the prosthesis is loose usually revision surgery is necessary. However, in many cases the diagnosis of aseptic loosening is difficult because the evaluation methods, including planar radiography, is not very accurate. It has recently been shown that young patients have a higher risk of revision [25]. In many cases the revision is of cause needed but there is also a risk that some young people are revised due to presumed loosening that is actually not true.
The aim of this study was that for the first time evaluate Implant Movement Analysis (IMA) as a complementary method to planar radiography and clinical examination in the clinical routine of diagnosing aseptic loosening of TKA. With this technique where CT scan is done during four different sets of forced provocations movement between bone and implant can be visualized on the screen (Figure 4), making the radiographically diagnosis much more accurate than by just watching planar radiographs. Our results indicate that IMA provides very valuable information in the decision-making regarding revision surgery [26,27].
Out of our 40 patients 22 had radiolucent lines on planar radiography to an extent that the radiologist assessed the implant as loose. Out of these 22 only 8 showed a detectable motion between bone and implant with use of IMA. This means that less than half of the patients with loose implants on conventional X-ray proved to be loose during forced provocation. It is well known that revision for pain and discomfort in a TKA knee where the implant is not loose has bad clinical outcome [5,6]. To avoid unnecessary revisions is thereby of great importance. Of course, the implant can get loose over time even after the IMA evaluation and follow up is needed. Our preliminary experience is however that many of the patients got satisfied with the message that their prosthesis was stable and became less worried which was beneficial for their function. This means that IMA examination can also be used for patients with high degree of anxiety because of pain and discomfort related to the knee even if the suspicion of a loose implant is low. When the patient receives the information that the TKA is well fixed and that nothing is mechanically wrong, this might lead to an improvement based on psychological reasons.
Another important finding in the present study is that out of the 18 knees without radiolucent lines no proved to be loose. This means that if there are no zones detectable on X-ray the risk of loosening of the implant is very small. To our knowledge this is the first study that has been able to verify this finding.
A limitation in our study is that this is the first study on the use of IMA for diagnosing TKA loosening. This means that there is no possibility to be absolutely sure that the results are correct. We know that of the 8 cases where IMA indicated loosening all were also loose at surgery. This means that the positive predictive value is 100 %. The problem is of course on the 32 knees where IMA showed no loosening, how can we know that this result actually is true? The answer is that we cannot. Our plan is to follow these patients over time and to perform a new evaluation 2 years after IMA evaluation with clinical and radiological examination to get an indication of the accuracy of the IMA results. We believe this is the only option to get closer to the true result. However, as the visual presentation of the IMA results are strong and the basic techniques for CT provocation has been developed over many years, we believe that the advantages of starting using IMA at this point in clinical use is great. This is also the only way to gather more information of the method and its value. We can see no disadvantages of introducing this method as a complement to the quite poor evaluation methods available today. The CT is of low-dose type and the radiation in level with a planar radiography and the procedure not to complicated.
The idea of using CT with provocation is not new. Berger et al. [23] and Reinus et al. [24] have been using this kind of technique to evaluate loosening of femoral stems in hip arthroplasty with good results. So, in line with previous findings, we believe that the future potential of IMA is very promising. In conclusion this is the first study of a new standardized technique to evaluation of potentially loose TKA implants. Our short-term experience is promising, and we can see no disadvantages in starting to use IMA as a complement to the methods used today.