Hybrid SPECT/CT for the assessment of a painful hip after uncemented total hip arthroplasty

Interpretation of scintigraphic uptake is dependent on the biomechanical characteristics of each prosthesis. Uncemented implants rely on osteointegration, and increased tracer uptake around a hip prosthesis is observed postoperatively. Apart from the greater trochanter and the tip of the prosthesis (Gruen [12] zones 1 and 4, Figure 2), which may show increased tracer uptake for longer than 12 months, bone turnover assessed with BS is close to normal after one year [13,14]. For reliable interpretation, we still do not recommend scintigraphic assessment of an uncemented hip prosthesis earlier than 18 months after implantation.

Uncemented cups usually have their crucial fixation zones at the outer circumference e.g. the superior (I) and inferior third (III). (Figure 2) Loosening was diagnosed when increased tracer uptake was found at the prosthesis-bone interface in more than one compartment or if increased tracer uptake corresponded with a radiolucent line or osteolysis seen on CT-scan. Increased tracer uptake in the superior or inferior third not adjacent to the bone-prosthesis interface was a common finding and rated as physiological bone-remodelling [15]. The central third of the cup does not necessarily require contact to the bone and is not crucial for fixation, sometimes osteolysis is observed in this area but does not affect stability of the cup.

In this study-group three basic stem-types of total hip arthroplasty were examined: Stems with distal (diaphyseal) fixation (Figure 2A), stems with metaphyseal/diaphyseal fixation (Figure 2B) and short stems with proximal fixation. The crucial fixation zone of the distal locking stem is the cortical contact-zone in the diaphysis [16] (Gruen zones 2,3,5,6), radiolucency or increased tracer uptake at the bone-prosthesis interface in this area indicates loosening. (Figure 3) Increased uptake at the tip of the prosthesis only in combination with signs of loosening elsewhere was interpreted as axial movement of the stem and therefore also as loosening. Proximal radiolucency medial or lateral of the stem (Gruen zones 1,7), so-called stress shielding is observed frequently and does not indicate loosening. (Figure 4) Stems with diaphyseal fixation may show signs of loosening at the proximal end of the stem, due to slight movement or swinging of the prosthesis [17].

Stems with metaphyseal/diaphyseal fixation rely on osteointegration of the mid-section on the implant, which is often coated or of a porous surface to enhance osteointegration. Gruen zones 6 and 7, as well as the distal part of Gruen zones 1 and 2 represent the crucial fixation zone of these stems. Radiolucency along with increased tracer-uptake may be observed at the distal part and tip of the prosthesis despite a stable fixation of the prosthesis [18,19]. Increased tracer uptake and signs of loosening at the proximal prosthesis-bone interface may also be observed to a certain extent without loosening of the prosthesis.

Because of the various principles of fixation we propose a close cooperation with the orthopaedic surgeon responsible in any uncertain case, in order to provide an optimized report.

CT provides 3D-information of the pelvis, femur and acetabular and femoral component of THA. Thus exact positioning of the prosthesis with cup anteversion and stem antetorsion can be assessed. The published range of 10°-15° for femoral antetorsion according to Tönnis et al. and cup anteversion of 5°-25° according to Lewinnek et al. represent a so-called safe zone [20,21]. In one case stem antetorsion of 45° and cup anteversion of 35° were measured, but the case remains unclear as in this incidence of uncertain persisting pain the patient has not yet been willing to undergo re-operation. In addition CT also allows a planar and volumetric evaluation of polyethylene wear and osteolysis, thus providing significantly more information than planar radiographs [22].

CT also shows clear benefits in the detection of hardware failure over radiographs and the characteristics of computed tomography suggest a benefit in the diagnosis of loosening, though this remains to be proven in clinical studies [23,24].

Aseptic loosening is the most frequent cause for revision, followed by dislocation, infection and fracture [2]. Sensitivity and specificity of radiographs for the detection of loosening is 82% and 81% for the femoral component and 70% and 80% for the acetabular component [25,26]. In the clinical workflow periprosthetic infection is considered unlikely if serum CRP and ESR are not elevated [9]. Dislocation and periprosthetic fractures are usually diagnosed clinically and via radiograph. With the inclusion criteria of this study, the remaining study-group is highly selected and contains only complicated and challenging cases.

In this study 5 cases of hip pain were spine related and the respective region was examined via SPECT/CT after planar imaging had shown an increased tracer uptake. It should therefore be discussed whether a routinely performed SPECT-examination of the lumbar spine should be recommended and added to the imaging protocol.

Ossifications are a known cause for hip pain and limited range of motion after THA. Conservative treatment or radiation is usually ineffective and surgical resection is the only curative treatment [27,28]. To minimize the chance of reoccurrence, surgical excision should be performed after maturation of the lesion, which can be determined with BS [29]. Activity and exact location of ossifications can be assessed with SPECT/CT and can be differentiated from component loosening if increased tracer uptake is located around the acetabular cup or the greater trochanteric region. In this cohort 4/5 patients with heterotopic ossifications were successfully treated with surgical excision after component loosening had been ruled out with SPECT/CT.

Differentiating between septic and aseptic loosening is a challenge with no exact definition for a periprosthetic infection [30]. In accordance with current guidelines a periprosthetic infection in this study-group was considered to be unlikely after a blood test with negative CRP and ESR with a negative likelihood ratio of 0 to 0.06 [9]. However, periprosthetic joint infection can never be excluded with certainty. Especially low-grade infection is a major diagnostic challenge and is methodically not adequately assessed with bone SPECT/CT. Therefore infection should remain within the spectrum of differential diagnoses as long as sufficient invasive diagnostics have been performed. This is demonstrated by one case in this study group, where the final diagnosis was septic loosening. Initial blood tests showed no abnormalities and SPECT/CT indicated loosening of the stem (F3,2). The scintigraphic uptake-pattern in this case of septic loosening did not differ from aseptic loosening, including the blood flow and blood pool phases of triple phase BS. Other imaging modalities such as white blood cell imaging, anti-granulocyte scintigraphy or FDG-PET offer an alternative for the differentiation of septic from aseptic loosening of THA [31].

Attenuation artefacts due to metallic implants can be problematic for the assessment of bone turnover at the periphery of the implant using hybrid SPECT/CT images, as they may lead to an overestimation of pixel intensity after attenuation correction. Nevertheless, a recent study showed that a reliable and robust interpretation of bone SPECT/CT images of the femoral head-neck junction is feasible even in the presence of a metal hip resurfacing implant [32]. Besides the problem of artefacts introduced by CT-based attenuation correction, the work flow for the interpretation of integrated SPECT/CT examinations, especially in terms of CT-diagnostics, needs to be improved and implemented.

Radiation exposure in hybrid examinations is a limiting factor, but new acquisition protocols and recent methodological advances like iterative CT reconstruction present promising ways to reduce exposure to a minimum (apart from the reduction of attenuation correction artefacts).

The scan protocol resulted in a volume CT dose index (CTDIvol) of 2.2 mGy and a dose-length product (DLP) of 88 mGy*cm for a single bed position. This results in an effective dose E* of 1.1 mSv for a standard patient geometry [33]. However, exposure to ionizing radiation is nevertheless an important issue and an individual risk assessment must routinely be performed prior to every examination.

Limitations of this study are the retrospective character, the low number of patients enrolled and the lack of a uniform standard of reference.

Furthermore, it is costly and time consuming, and requires exposure to ionizing radiation.