Imaging evaluation of periprosthetic loosening: A primer for the general radiologist

Abstract

In response to an ageing global population, the primary hip and knee arthroplasty rate continues to increase. Although an effective treatment, up to 25% patients may require revision arthroplasty during their lifetime, commonly due to periprosthetic loosening. Revision procedures are associated with significantly increased healthcare costs; therefore, timely and accurate diagnostics are critical for clinicians and patients. Loosening, which may be septic or aseptic, remains a challenge and requires thorough clinical examination and multimodal imaging evaluation. Plain radiographs remain an essential diagnostic tool but advanced imaging modalities such as computed tomography, magnetic resonance imaging and nuclear medicine are playing an increasingly important role. This comprehensive review, through outlining the available radiological modalities, their respective strengths and weaknesses and the pertinent imaging findings, may help radiologists and orthopaedic surgeons make more informed decisions in the management of periprosthetic loosening.

Key Words: Periprosthetic loosening; Total hip replacement; Total knee replacement; Diagnostic radiology; Aseptic loosening

Core Tip: As the rate of joint arthroplasties continue to rise, so too does the rate of revision arthroplasty. One of the primary indications for revision surgery is periprosthetic loosening; however, diagnosis of this condition is challenging. In this review, we discuss some of the major breakthroughs and advances in imaging techniques that aim to address these challenges. Improvements in metal artefact reduction techniques in computed tomography and magnetic resonance imaging have markedly improved image quality and diagnosis. In addition, nuclear medicine imaging techniques such as single-photon emission computed tomography/computed tomography are proving to be an excellent aid in diagnosis.

INTRODUCTION

Hip and knee arthroplasties are among the fastest growing surgeries worldwide with > 480000 primary total hip arthroplasties (pTHAs) and > 260000 primary total knee arthroplasties (pTKAs) performed in the USA in 2019 alone. With an ageing population, this number is expected to rise to > 700000, and 1.2 million pTHAs and pTKAs, respectively being performed annually in the USA by 2040[1]. Although the success rates and overall satisfaction rates for both pTHA and pTKA are excellent, revision surgery remains a common occurrence. Currently, the overall 10-year survival rate is 93.6% and 95.6% for pTHA and pTKA, respectively, although lifetime survival is age dependent. Indeed, younger patients (age 46–50 years) have a lifetime risk of 22.6% for revision total hip arthroplasty (rTHA) and 25.2% for revision total knee arthroplasty (rTKA)[2,3].

Not surprisingly, with the rising numbers of pTHAs and pTKAs there is a proportionate increase in the rate of rTHA and rTKA, with which comes an increased cost of care, difficulty in resource allocation and an overall stress on healthcare systems worldwide[4]. Although there are various indications for rTHA and rTKA, the most common are aseptic loosening and infection. For pTHA failures, aseptic loosening accounts for 52%, and for pTKA 21.6%[5,6]. Infection is a more common cause for failed pTKA, being estimated to be responsible for 36.3% revisions[5,6].

Serial radiographs play an important role in the diagnosis of periprosthetic loosening and various radiographic signs have been identified in the literature that indicate mechanical loosening[7]. However, limitations exist in the diagnostic capability of conventional radiographs, and they must be interpreted in conjunction with careful clinical evaluation[8]. One of the major limitations is that early loosening may not be detectable on plain radiography because evidence of bony integration of cementless implants is not visible. Therefore, an implant that appears to be firmly in place may in fact be loose[9].

As a result, recently, advanced imaging modalities such as computed tomography (CT), single-photon emission CT (SPECT), bone scintigraphy and magnetic resonance imaging (MRI) have been utilised more frequently in the diagnosis of periprosthetic loosening[10]. This review aims to provide general radiologists with a comprehensive overview of the current status of imaging in periprosthetic loosening and thus guide them in the diagnostic evaluation of these patients. Figures included within the article are accompanied by brief case descriptions with an aim of providing readers with real world examples of the utility of various imaging modalities.

LITERATURE REVIEW

During September 2024, PubMed and MEDLINE OVID were searched for relevant articles. MeSH terms included prosthesis failure, hip prosthesis, knee prosthesis and radiography. Studies published in English were included. Literature published within the last decade was focused on to consider evolving practice and technologies, however, older articles were not excluded from the literature review. Included figures were acquired from the institutional database of teaching cases and were anonymised prior to inclusion. As no identifiable patient information was collected, the clinical research ethics committee advised that formal ethical approval is not required.

PLAIN RADIOGRAPHY

Conventional radiography is the most commonly used modality in the evaluation of patients with suspected periprosthetic loosening. Typically, patients will present to the orthopaedic clinic or emergency department with a painful limb and difficulty in weightbearing. The first line imaging investigation is usually plain radiography with anteroposterior and lateral views of the affected limb. In addition to loosening, radiographs can be used to determine component alignment, polythene line wear, osteolysis, reactive bone formation as well as any evidence of periprosthetic fractures[11,12].

Numerous radiographic signs have been identified over the years that could indicate periprosthetic loosening. The most widely published being the presence of radiolucent lines > 2 mm around the prosthesis. Lucencies between 1 and 2 mm are generally considered to have indeterminate significance while lucencies < 1 mm are unlikely to be of clinical significance[13]. In addition to lucencies, loosening may be indicated by subsidence or movement/migration of the components. For example, in a THA prosthesis, vertical subsidence > 5 mm of the femoral component in indicative of loosening (Figure 1). Movement, as indicated by increased varus/valgus angulation of the femoral stem, is also indicative of loosening. Loosening may also be noted in the acetabular component and is commonly seen as superior or medial migration of the prosthetic cup.

Figure 1 Subsidence of a loose femoral component in a revision total hip arthroplasty. The above case showcases the utilisation of plain radiography in the diagnosis of aseptic loosening. This patient had presented to the outpatient’s clinic 1 year following revision total hip arthroplasty with complaints of right sided hip pain that was gradually worsening. The revision surgery utilised a diaphyseal bearing modular revision system with cerclage cables for closure of the proximal femoral osteotomy as a result of periprosthetic fracture. A–C: Immediate postoperative (A) and 1-year postoperative (B) plain radiographs are displayed with significant subsidence of the femoral component noted in the follow-up appointment. Following careful clinical evaluation and confirmation of normal biochemical markers, the patient was diagnosed with aseptic loosening and underwent a planned revision total hip arthroplasty with a proximal femoral replacement (C).

In the case of a TKA prosthesis, subsidence may be noted when the tibial component is seen sinking into the tibial plateau. This is commonly associated with varus malalignment because of high medial load when compared with previous X-rays[14,15]. However, many of these hallmark signs require the presence of previous radiographs for comparison. In the absence of these, particularly for milder cases, diagnosis based on plain radiography becomes difficult[16]. In such cases additional imaging from alternative radiological modalities may be required.

DIGITAL TOMOSYNTHESIS

Although digital tomosynthesis (DT) has been used in breast imaging for many years, it is a relatively new technology in the field of musculoskeletal imaging. DT uses standard X-ray equipment with digital flat panel detectors to detailed images from low-dose projections obtained at different angles. Simply put, instead of a plain radiograph that captures an image at one angle, multiple images at different angles are captured and computer algorithms used to reconstruct the images in 3D depending on the pathology of interest. In evaluating periprosthetic loosening, DT with advanced imaging processing techniques such as iterative reconstruction with metal artefact reduction (MAR) can provide improved image quality and permits more accurate evaluation of the bone/implant interface compared to conventional radiography[17]. In addition, the radiation dose is lower compared to that with CT with similar specificity to CT and plain radiography at detecting loosening. Although sensitivity is similar to radiography, it is significantly lower than for CT. However, there is evidence to suggest interobserver agreement is better with DT than with plain radiographs[18].

DT has shown significant advantages over plain radiographs at detecting new bone formation around prostheses and ruling out loosening. In 2021, Totsuka et al[19] found that cancellous condensation around the stem, which appears as a radiolucent line on plain radiography, resulted in loosening being overcalled. DT, however, was able to provide more detailed information and exclude loosening for many of these patients and thus significantly alter their clinical course by avoiding unnecessary revision surgery[19]. Overall, although DT is unlikely to replace either plain radiographs or CT in evaluating loosening, it allows for more detailed evaluation of the bone/implant interface compared to plain radiography without an excessive radiation dose penalty.

CT

As a result of the limitations of plain radiographs and DT, CT with its ability to capture images in multiple planes, is often utilised to aid diagnosis. While plain radiographs may show equivocal findings, CT can characterise the extent and width of the periprosthetic lucency. In conjunction with careful clinical examination, this can help clinicians reach a diagnosis. Unfortunately, despite offering excellent visualisation of bony anatomy, CT images can be significantly affected by the presence of metal artefacts[20]. The artefacts produced can vary depending on the size, shape and material of the implant and in many cases, can cause difficulty in reaching a diagnosis[21]. These artefacts include image blurring as well as dark/bright streaks as a result of the metal implant blocking or scattering the X-rays. The composition of the metal component can affect the severity of artefacts quite significantly with metals such as titanium generating more artefacts in comparison to iron and platinum, for example[22].

Fortunately, MAR techniques have improved drastically in recent times and the diagnostic capability of CT has improved for these patients (Figure 2). Modern MAR techniques focus on manipulations in data acquisition, image reconstruction and post-processing to produce an image that allows the visualisation of the interface between the metal component and osseous tissue[23]. Similar to conventional radiographs, in CT, lucency > 2 mm around a prosthesis is indicative of loosening. However, in a symptomatic patient, lucency < 2 mm cannot satisfactorily exclude loosening. When radiographs are equivocal, CT scanning provides a fast and efficient means of clarifying the diagnosis[24].

Figure 2 Computed tomography scan of loose cemented acetabular component with comparison plain radiographs. This case displays the utility of computed tomography (CT) with metal artefact reduction in diagnosing periprosthetic loosening. The gentleman in this case was referred to the outpatient’s clinic by his family physician with complaints of gradually worsening pain and discomfort in the left hip. The patient’s blood markers were normal at the time of referral and examination findings were equivocal. A and B: Acetabular loosening in a cemented total hip arthroplasty was confirmed on initial plain radiography of the pelvis (A). However, characterisation with CT was required in this case to determine the extent of loosening and appropriately plan revision surgery (B). Both images were acquired on the same day. Although acetabular loosening is visible in Figure 2A around DeLee Charnley Zones 1 & 2, the extent of loosening and bony erosion is far clearer in Figure 2B (blue arrows). An approximate 10 mm rim of lucency was demonstrated on the CT scan which was not easily appreciated on the initial plain radiographs (yellow arrow). The patient proceeded to a successful revision hip arthroplasty following review of the CT images.

An advanced type of CT technique, termed dual energy CT (DECT) was developed in the late 1970s and has allowed further improvements in image quality for cases with metal implants. In simple terms, DECT incorporates the use of two different X-ray energy spectra, which may be achieved by using two X-ray tubes operating at different energy levels, by rapidly switching the energy levels of a single X-ray tube or through the use of a dual layer detector. Comparing the attenuation characteristics at different energy levels allows for more accurate differentiation between materials thus improving image quality around metal implants.

In 2023, Foti et al[12] demonstrated the superiority of DECT over conventional CT and radiography in both diagnostic capability as well as in inter-reader agreement of the scan results in patients with TKA loosening. However, despite the improvements in CT, certain limitations exist. Firstly, DECT is a new development and the availability of these new scanners, particularly in smaller peripheral hospitals may be limited. Secondly, although advances have been made in MAR techniques, artefacts remain problematic and can cause diagnostic uncertainty in both conventional CT and DECT. Consequently, other additional imaging modalities have been utilised to aid diagnosis.

ULTRASOUND

Ultrasound (US) as an imaging modality for musculoskeletal disorders is excellent and provides visualisation of muscle, bone, tendons and ligaments. US has the advantage of offering dynamic imaging with the benefit of no radiation exposure to the patient[25]. Unfortunately, US does not allow for the direct visualisation of periprosthetic loosening. Rather, it plays a role in the identification of adverse local tissue reactions (ALTR); for example, as seen with pseudotumour formation in metal-on-metal THA prosthesis. When abrasive and/or adhesive wear of prosthetic components occurs there is a macrophage led cascade that ultimately results in the formation of ALTR, a process that will hasten aseptic loosening and subsequently accelerate the rate of failure[26]. Therefore, the ability of US to visualise ALTRs can help identify patients at risk of loosening despite not having the capability of directly making a diagnosis. The limitations of US in image definition and material differentiation limits its routine use in cases of suspected peri-prosthetic loosening and other imaging modalities are often favoured.

MRI

MRI has proved vital in the diagnosis of various metal artefact reduction pathology but its role in detecting periprosthetic loosening is less well established. Recognition of a loose component requires visualisation of the bone implant interface which is made difficult due to the presence of significant metal artefacts[27]. Recent times have seen improvements in MAR techniques applied to standard MRI sequences as well as the development of 3D multispectral imaging[27]. Two recent MRI techniques developed by Siemens and General Electric are slice encoding for metal artefact correction (SEMAC) and multi-acquisition variable resonance image combination (MAVRIC), respectively.

Conventional MRI uses a single resonant frequency to create images whereas in contrast, MAVRIC involves collecting data at various offset frequencies from the main resonance frequency. As a result, different parts of the frequency spectrum affected by the metal implant are sampled. Various sophisticated algorithms are then used to combine the multiple images obtained. By integrating the data from different frequencies, the final image aims to reduce artefacts caused by the implant and thus allow for better visualisation of the periprosthetic area region. SEMAC, rather than targeting the entire volume and relying on frequency offsets, focuses on correcting artefacts slice-by-slice. SEMAC attempts to correct distortions within each slice and subsequently combines the corrected slices to form a final image.

In cases where metal artefacts are severe, a hybrid approach combining the two techniques can be utilised for maximal artefact reduction. This technique is termed MAVRIC selective. There is evidence in the literature to show that MAVRIC selective can significantly reduce metal artefacts as compared to traditional MRI sequences and thus detailed evaluation of the periprosthetic bone–implant interface can be made possible[27,28].

It is important for clinicians to be aware of these recent developments in MAR techniques to be better able to offer the patient the greatest chance of successful diagnosis. While metal artefacts have not been completely eradicated, newer techniques have improved the quality of MRI near implants and allowed for diagnosis of more pathology and post-operative complications[29]. Signs of loosening on MRI include formation of a fibrous membrane between the implant and bone, a fluid interface, or osteolysis around the prosthesis[30]. MRI is not routinely used for investigation of periprosthetic loosening in clinical practice because of the presence of artefacts and the availability of CT, which has faster acquisition times and lower costs.

Costs and slower image acquisition times compared to other modalities can significantly limit the use of MRI. However, due to the superior performance of MRI in characterising soft tissue lesions and collections, it still has a role to play in infection-related periprosthetic loosening (Figure 3). Unfortunately, despite advances in MAR sequences, metal artefacts have not been eliminated and careful clinical evaluation as well as thorough scrutiny of other available imaging is still required.

Figure 3 Triple phase bone scintigraphy and magnetic resonance imaging demonstrating loose femoral component. The above case displays the importance of planar bone scintigraphy and magnetic resonance imaging (MRI) with multi-acquisition variable resonance image combination selective sequencing in the diagnosis of periprosthetic loosening. The patient in this case had presented 4 years following bilateral primary total hip replacements with discomfort in the right hip only. A and B: Initial plain radiographs were unremarkable, therefore higher order imaging in the form of planar bone scintigraphy (A) and MRI (B) were acquired. Marked increased uptake of the technetium-99m MDP radiotracer in the region of the right greater trochanter and shoulder of the femoral component was noted (yellow arrows). Physical examination findings of a soft tissue mass and raised inflammatory markers prompted the acquisition of MRI for further characterisation. MRI findings of a soft tissue collection (blue arrows) led to the diagnosis of septic loosening in this patient. This was later confirmed intraoperatively during revision surgery.

NUCLEAR MEDICINE

Various imaging modalities exist in the realm of nuclear medicine that have proven useful in the investigation of patients with suspected periprosthetic loosening. They include bone scintigraphy, positron emission tomography (PET) and SPECT. Each have their advantages and disadvantages and are discussed below. Bone scintigraphy has been extensively studied in joint arthroplasty and has been in use for several decades. Conventionally, technetium-99m labelled diphosphonates are used for this study. The incorporation of the radiotracer into bone is dependent on the blood supply to the bone and rate of metabolic activity. Areas of high metabolic activity result in increased uptake of the radiotracer thus aiding clinicians in the diagnosis of loosening and infection in joint arthroplasty. Unfortunately, although sensitive for loosening, studies have found bone scintigraphy to be nonspecific and differentiating between infection and aseptic loosening may be difficult (Figure 3). Therefore, correlation must be made with the patient’s clinical history, examination findings, and blood markers.

Some materials, such as hydroxyapatite-coated hip replacements, may result in high rates of false positives making a positive result in these studies hard to interpret. As with other imaging modalities discussed previously, careful clinical evaluation is essential for accurate diagnosis. Despite its low specificity, bone scintigraphy remains relatively cheap and accessible and can provide additional information to conventional plain radiographs.

SPECT is a type of cross-sectional nuclear imaging technique that uses single photon emitting radiopharmaceuticals to provide functional and physiological data on organs or pathological tissue. In the case of THA and TKA, SPECT combined with planar bone scintigraphy, can provide high sensitivity and specificity for loosening. However, this technique displays poor accuracy in the localisation of the area of increased uptake. Therefore, this limits its role clinically as the decision of which prosthetic component is to be revised, if any, can be difficult to make.

Fortunately, over the last decade, a hybrid technique combining the functional capability of SPECT with the anatomical detail of CT has been developed to address this shortcoming. This technique is proving to be increasing useful in diagnosis of loosening with high sensitivity and specificity reported[31-33] (Figure 4). In 2011, Hirschmann et al[34] described using SPECT/CT in evaluation of painful TKAs and found that SPECT/CT significantly changed the surgeons’ decision of revision with many patients who were initially planned for revision, being managed nonsurgically. In addition, malposition and patellofemoral osteoarthritis, both of which could be the cause of symptoms, were identified in TKA patients with the help of SPECT/CT[34].

Figure 4 Plain radiographs and corresponding single-photon emission computed tomography/computed tomography of tibial component loosening. A–D: This is a recent case within our institution demonstrating the capability of single-photon emission computed tomography (SPECT)/computed tomography (CT) in diagnosing periprosthetic loosening. Initially, anteroposterior (A) and lateral (B) radiographs were performed to investigate the cause of right sided knee pain 8 years after cemented primary total knee replacement. The radiographs did not display typical signs of loosening or other abnormalities that could explain the patient’s symptoms. In addition, inflammatory markers were normal and clinical examination was unremarkable. Aseptic loosening was suspected and advanced imaging in the form of SPECT/CT was acquired. Coronal (C) and sagittal (D) SPECT/CT images demonstrate increased radiotracer uptake along the inferior aspect of the tibial base plate both medially and laterally but particularly posteromedially (white arrows). On the basis of the SPECT/CT image findings in conjunction with the clinical history, exam and blood markers, a diagnosis of aseptic loosening was made. The patient opted to undergo a revision total knee replacement at an alternative private institution.

Further evidence of the capability of SPECT/CT in diagnosing loosening was seen in a systematic review published by Peng et al[35] in 2021. The authors found that SPECT/CT had an impressive sensitivity and specificity of 0.94 and 0.89 in diagnosing aseptic loosening in TKA and THA, respectively. In fact, they recommended using SPECT/CT as the primary imaging modality in diagnosing periprosthetic loosening. Unfortunately, despite its excellent performance, high costs and high radiation doses limit its use[35].

PET is another nuclear imaging technique that involves the use of radiotracers to analyse changes in metabolism. A commonly used radiotracer is F-18 fluorodeoxyglucose, which can be used to illustrate changes in glucose metabolism at the cellular level. This is particularly useful in oncology and has been used for many years for monitoring cancer activity. In assessment of periprosthetic loosening, however, sodium fluoride (NaF) conjugated to F-18 is a more useful radiotracer. Previous studies have illustrated its effectiveness in characterising bone metabolism, and its use has been validated by comparison with bone biopsies. Used in isolation, 18(F)-NaF PET can provide information on areas of increased bony metabolic activity such as infection, inflammation and tumour, however, anatomical localisation is limited.

As a result, a combination of 18(F)-NaF PET and CT is often utilised to alleviate some of these shortcomings. This hybrid imaging technique provides precise information on anatomic localisation and aids in the identification of extraosseous disease not previously seen on PET in isolation[36,37]. The effectiveness of 18(F)-NaF PET/CT was demonstrated by Ullmark[38] in 2020, who identified patients with loosening that previously had normal or equivocal plain radiograph findings. These patients identified using 18(F)-NaF PET/CT all underwent revision surgery and were confirmed to have loose components intraoperatively[38]. Although effective, PET/CT scanning is an expensive process, even more so than SPECT/CT, and availability may be even more limited as a result. Regardless, due to cost of revision arthroplasty and the potential consequences of revision for the patient, when available, PET/CT may be considered for diagnosis of suspected periprosthetic loosening.

SUMMARY

As the population in the western world continues to age, the number of primary hip and knee arthroplasties continues to rise. Consequently, the number of revision hip and knee arthroplasties being performed yearly are also rising due to complications such as aseptic loosening and infection. Accurate diagnosis of periprosthetic loosening is essential for early and effective patient management. However, diagnosis is not easy and requires careful clinical evaluation individualised to each patient.

While conventional radiography remains an excellent first-line diagnostic tool, it lacks the spatial resolution offered by CT and MRI and the functional information provided by nuclear medicine imaging. DT offers increased spatial resolution compared to plain radiography with a lower radiation dose than CT but lacks the high sensitivity of CT. Metal artefacts remain problematic and evaluation of the bone/implant interface in CT and MRI can often be obscured by artefacts. However, advances in MAR reconstruction and DECT have improved image quality, as demonstrated by the images in this review. US may be used as an adjunct to evaluate loosening, particularly in a metal-on-metal prosthesis where pseudotumour may be demonstrated; however, it cannot be used alone for diagnosis. Nuclear imaging techniques, particularly hybrid modalities such a SPECT/CT offer both functional information as well as increased spatial resolution compared to plain radiography. However, cost and accessibility remain an issue.

An integrated approach that combines careful clinical evaluation with the appropriate imaging modality is essential to accurately diagnose periprosthetic loosening and thus guide management. Through outlining the available radiological modalities, and their respective strengths and weaknesses, this article may help clinicians make informed decisions in the selection of imaging modalities for the diagnosis of periprosthetic loosening. Thus, there is the potential to minimise the number of unnecessary revision operations.

While orthopaedic surgeons should have strong involvement in the diagnostic pathway as early as possible, radiologists have a significant role in deciding on the most appropriate imaging modality for a patient. Each patient is unique, and the modality of choice may vary depending on availability, cost and clinical presentation. Thus, a summary flow diagram based on the findings of this review was developed (Figure 5). In conjunction with the cases described in each figure, the flow diagram aims to serve as a guide for clinicians in diagnosing this challenging pathology.

Figure 5 Summary flow diagram of recommended radiological diagnostic pathway. The flow diagram above aims to aid the general radiologist in the diagnosis of suspected periprosthetic loosening. As with any most pathologies, the pathway begins with careful history taking and physical examination by the referring clinician. Baseline biochemical markers and plain radiography ± digital tomosynthesis form the first step in the diagnostic pathway. The recommended further steps are based on the results of the initial step and are listed above. Discussion with the clinician who is the primary carer for the patient is crucial in every step of the pathway. FBC: Full blood count; U&E: Urea and electrolytes; CRP: C reactive protein; ESR: Erythrocyte sedimentation rate; IR: Interventional radiology; DT: Digital tomosynthesis; CT: Computed tomography; MRI: Magnetic resonance imaging; MARS: Metal artefact reduction sequence; SPECT: Single-photon emission computed tomography; PET: Positron emission tomography.

CONCLUSION

Although careful clinical history and examination remain vital in diagnosing periprosthetic loosening, appropriate use of radiological investigations can provide vital information. Clinicians should be aware of the various image modalities available to aid in the diagnosis of the challenging condition as well as the advantages and disadvantages of each.