**2. Conventional diagnostic investigation of disorders after THR**

### **2.1 Plain radiography**

Plain radiography is an imaging modality, which is less cost-intensive than other imaging techniques and therefore the first method used to determine the osseointegration of a THR (Nilsson et al., 1994). Diagnosis of THR loosening typically requires a plain radiograph - in the optimal case it can be compared with an immediate post operative radiograph - and clinical symptomatic like hip pain. In order to achieve an accurate interpretation of the radiography, the observer needs adequate clinical expertise regarding the surgical techniques applied in THR (Cuckler, 2010). Two different views are usually taken into account: the anteroposterior view of the hip and pelvis and the lateral view (supine position) to investigate the affected femoral stem (Temmerman et al., 2005).

The radiographic criteria vary between the type and operative technique of the THR (Ostlere & Soin, 2003). Different radiographic criteria for the identification of loosening were reported (DeLee & Charnley, 1976; Dihlmann et al., 1991; Gruen et al., 1979, Johnston et al., 1990). The appearance of radiolucent lines around the endoprosthetic components as distinctive periprosthetic membrane is one criteria (Paterson et al. 1986), when the width of the radiolucency is more than 2 mm (Fig. 2) (Böhler et al. 1994).

In this article, a review of diagnostic possibilities to detect the loosening status of a THR is given. In general, diagnostic methods can be divided into imaging and implant integrated sensors (Fig. 1). The most important research activities were conducted in the field of vibrometry with the use of accelerometers (Marschner et al., 2009). Moreover, new research topics which link sensors to THR are described. Finally, we propose a new excitation

Fig. 1. Overview of the modalities which can be used to diagnose loosening of the THR

Plain radiography is an imaging modality, which is less cost-intensive than other imaging techniques and therefore the first method used to determine the osseointegration of a THR (Nilsson et al., 1994). Diagnosis of THR loosening typically requires a plain radiograph - in the optimal case it can be compared with an immediate post operative radiograph - and clinical symptomatic like hip pain. In order to achieve an accurate interpretation of the radiography, the observer needs adequate clinical expertise regarding the surgical techniques applied in THR (Cuckler, 2010). Two different views are usually taken into account: the anteroposterior view of the hip and pelvis and the lateral view (supine

The radiographic criteria vary between the type and operative technique of the THR (Ostlere & Soin, 2003). Different radiographic criteria for the identification of loosening were reported (DeLee & Charnley, 1976; Dihlmann et al., 1991; Gruen et al., 1979, Johnston et al., 1990). The appearance of radiolucent lines around the endoprosthetic components as distinctive periprosthetic membrane is one criteria (Paterson et al. 1986), when the width of

**2. Conventional diagnostic investigation of disorders after THR** 

position) to investigate the affected femoral stem (Temmerman et al., 2005).

the radiolucency is more than 2 mm (Fig. 2) (Böhler et al. 1994).

**2.1 Plain radiography** 

method for detection of total hip stem loosening using vibrometry.

Fig. 2. Radiograph of an uncemented THR with a clearly visible radiolucent line at the interface of the femoral component

Cemented components can be identified by radiolucency between bone and cement (Sarmiento et al., 1990), while in the uncemented case, radiolucent lines occur between the implant and bone. However, some radiolucency can be observed e.g. in the proximal part of the uncemented Zweymueller design hip stem, though the THR is well osseointegrated (Suckel et al., 2009). This example impedes a precise interpretation of the THR with plain radiographs, because these findings can be associated with loose as well as well fixed THR. The ambiguity in diagnosing loosening can be compensated by comparing the current radiograph with a previous post-operative reference (Wroblewski, 1991). Additional radiographic criteria refer to the tilt of any component. Another feature indicating impeding failure in both cemented and uncemented prostheses is focal osteolysis, due to the particle response of the tissue. Osteolysis is usually seen as decrease of osseointegration at the femoral site or bone destruction beneath the acetabular component (Keogh et al., 2003). Migration of the THR, particularly the femoral component, is the most significant and certain criteria to differentiate between a loosened and a fixed implant. If the location of the implant in the bone stock determined in the radiographs has changed over a distance of more than 4 mm in the postoperative follow-up the implant can be identified as loosened.

Plain radiography is the imaging of choice to assess osseointegration of THRs since in the standard clinical follow-up this is mainly the only imaging required (Ostlere & Soin, 2003). Radiography has been consistently reported to be an accurate technique to assess cemented and uncemented THR loosening. However, in some cases, especially in the observation of infectious THR other methods like e.g. scintigraphy have to be adducted since differentiation between aseptic and septic loosening using radiographs may not be possible. Many studies evaluating the sensitivity and specificity of radiographs in diagnosing loosening of the THR have been published (Fig. 3). Despite radiography and radiographic criteria being highly developed, an accuracy of 100 % in diagnosing loosening can not yet be achieved (Miniaci et al., 1990; Ovesen et al., 2003; Pfahler et al., 1998).

Current Possibilities for Detection of Loosening of Total Hip

component was confirmed (Ovesen et al. 2003)

the implant has to be considered (DeLee & Charnley, 1976).

Replacements and How Intelligent Implants Could Improve Diagnostic Accuracy 367

introduced, in which subtracted images can be made during injection of the contrast agents. Joint motion can reduce the resolution of the arthrogram and for this reason a pixel shift is

Fig. 4. (A) Cemented painful total hip arthroplasty in 60-year-old man with no signs of loosening on plain radiographs. The patient also had a history of spinal stenosis. (B) interoposterior and (C) 45° oblique DSA clearly show a thin contrast leakage at the bonecement interface in zones 1 and 2. White arrow indicates contrast in the interface; black arrow indicates filling of the lymphatic system. At repeat surgery, loosening of the femoral

Increasing the injected amount of contrast media into the hip joint can be helpful to improve the expressiveness of the achieved arthrogram (Palestro, 2003). One arthrographic criterion for loosening is the contrast leakage in the interface distal to the intertrochanteric line (Ovesen et al., 2003). For the acetabular component, contrast leakage in three zones around

Digital subtraction techniques can be used to increase the sensitivity. Temmerman et al. found an increased sensitivity of 15 % compared to conventional nuclear arthrography for the examination of the acetabular component (Temmerman et al., 2004). In contrast to this optimization, arthrography is identified to be non specific for the acetabular component (Köster et al., 1993). This result is based on the overall sensitivity of nuclear arthrography for acetabular components, which was determined to 57 %, whereas the value of specificity lies at 67 % (Temmerman et al., 2005). However, there is a discrepancy of the benefit of arthrography. Several researchers present the arthrogram as the only picture revealing contrast leakage and therefore loosening of the THR (Hardy et al., 1988; O`Neill & Harris, 1986). Other studies conclude that the selection of patients for revision was influenced by the results of the arthrography (Phillips & Kattapuram, 1982). Ovesen et al. reported a sensitivity and specificity for the femoral component of 93 % and 92 % respectively (Ovesen et al. 2003). They concluded that digital subtraction arthrography is a useful technique for

applied to reduce motion artefacts (Apple et al., 1986; Ginai et al., 1996).

Fig. 3. Literature comparison of selected results determining sensitivity and specificity of loosening diagnosis of the femoral component (FC) and the acetabular component (AC) using plain radiography

The highest accuracy in diagnosing loosening of femoral (98 %) and acetabular components (97 %) was documented by Lieberman et al. (Lieberman et al., 1993). These data show the lowest accuracy of 76 % for femoral components (Köster et al., 1993) and 66 % for acetabular components (Ovesen et al., 2003) and demonstrate that the loosening status of the THR cannot be verified precisely (Lyons et al., 1985). In this context it is important to mention that not every author differentiates between investigating the acetabular and the femoral component (Gelman et al., 1976; Barentsz et al., 1986; Tehranzadeh et al., 1981).

### **2.2 Arthrography**

Arthrography is an imaging method which is mainly used in addition to plain radiography in case of unresolved pain of the artificial joint (Newberg & Wetzner 1985). Contrast agents are injected into the joint for the visualization of the periprosthetic membrane around the THR and the cement-bone interface, in order to increase the accuracy and therefore specificity and sensitivity in loosening diagnosis (Hendrix et al. 1983). Several methods to perform arthrography of the hip joint are applied: Conventional arthrography is performed using x-ray examination under fluoroscopic guidance, in which anterolateral puncture of the hip joint is often applied (Hendrix et al., 1983). Anteroposterior and lateral views are normally the views which are obtained during the examination of the THR. Additionally, arthrography can be executed by using magnetic resonance imaging (MRI) or computed tomography (CT).

Arthrographic criteria for loosening are based on the infiltration of the injected contrast agents into the periprosthetic membrane or into the cement-bone interface (Ovesen et al., 2003) (Fig. 4). The THR and the contrast medium have similar radiographic densities. For that reason, the small width of the periprosthetic membrane cannot be clearly visualized by conventional arthrography. Therefore, digital subtraction arthrography (DSA) has been

Fig. 3. Literature comparison of selected results determining sensitivity and specificity of loosening diagnosis of the femoral component (FC) and the acetabular component (AC)

The highest accuracy in diagnosing loosening of femoral (98 %) and acetabular components (97 %) was documented by Lieberman et al. (Lieberman et al., 1993). These data show the lowest accuracy of 76 % for femoral components (Köster et al., 1993) and 66 % for acetabular components (Ovesen et al., 2003) and demonstrate that the loosening status of the THR cannot be verified precisely (Lyons et al., 1985). In this context it is important to mention that not every author differentiates between investigating the acetabular and the femoral

Arthrography is an imaging method which is mainly used in addition to plain radiography in case of unresolved pain of the artificial joint (Newberg & Wetzner 1985). Contrast agents are injected into the joint for the visualization of the periprosthetic membrane around the THR and the cement-bone interface, in order to increase the accuracy and therefore specificity and sensitivity in loosening diagnosis (Hendrix et al. 1983). Several methods to perform arthrography of the hip joint are applied: Conventional arthrography is performed using x-ray examination under fluoroscopic guidance, in which anterolateral puncture of the hip joint is often applied (Hendrix et al., 1983). Anteroposterior and lateral views are normally the views which are obtained during the examination of the THR. Additionally, arthrography can be executed by using magnetic resonance imaging (MRI) or computed

Arthrographic criteria for loosening are based on the infiltration of the injected contrast agents into the periprosthetic membrane or into the cement-bone interface (Ovesen et al., 2003) (Fig. 4). The THR and the contrast medium have similar radiographic densities. For that reason, the small width of the periprosthetic membrane cannot be clearly visualized by conventional arthrography. Therefore, digital subtraction arthrography (DSA) has been

component (Gelman et al., 1976; Barentsz et al., 1986; Tehranzadeh et al., 1981).

using plain radiography

**2.2 Arthrography** 

tomography (CT).

introduced, in which subtracted images can be made during injection of the contrast agents. Joint motion can reduce the resolution of the arthrogram and for this reason a pixel shift is applied to reduce motion artefacts (Apple et al., 1986; Ginai et al., 1996).

Fig. 4. (A) Cemented painful total hip arthroplasty in 60-year-old man with no signs of loosening on plain radiographs. The patient also had a history of spinal stenosis. (B) interoposterior and (C) 45° oblique DSA clearly show a thin contrast leakage at the bonecement interface in zones 1 and 2. White arrow indicates contrast in the interface; black arrow indicates filling of the lymphatic system. At repeat surgery, loosening of the femoral component was confirmed (Ovesen et al. 2003)

Increasing the injected amount of contrast media into the hip joint can be helpful to improve the expressiveness of the achieved arthrogram (Palestro, 2003). One arthrographic criterion for loosening is the contrast leakage in the interface distal to the intertrochanteric line (Ovesen et al., 2003). For the acetabular component, contrast leakage in three zones around the implant has to be considered (DeLee & Charnley, 1976).

Digital subtraction techniques can be used to increase the sensitivity. Temmerman et al. found an increased sensitivity of 15 % compared to conventional nuclear arthrography for the examination of the acetabular component (Temmerman et al., 2004). In contrast to this optimization, arthrography is identified to be non specific for the acetabular component (Köster et al., 1993). This result is based on the overall sensitivity of nuclear arthrography for acetabular components, which was determined to 57 %, whereas the value of specificity lies at 67 % (Temmerman et al., 2005). However, there is a discrepancy of the benefit of arthrography. Several researchers present the arthrogram as the only picture revealing contrast leakage and therefore loosening of the THR (Hardy et al., 1988; O`Neill & Harris, 1986). Other studies conclude that the selection of patients for revision was influenced by the results of the arthrography (Phillips & Kattapuram, 1982). Ovesen et al. reported a sensitivity and specificity for the femoral component of 93 % and 92 % respectively (Ovesen et al. 2003). They concluded that digital subtraction arthrography is a useful technique for

Current Possibilities for Detection of Loosening of Total Hip

signs of loosening or infection (Kisielinski et al., 2003)

loosening diagnosis of THR using bone scintigraphy

loosening and supports the first presumption.

Replacements and How Intelligent Implants Could Improve Diagnostic Accuracy 369

uptake. A second substantial lesion in the region of the lesser trochanter is a further sign for

Fig. 6. 99mTc bone scan with normal anterior and posterior image of the prosthesis, without

A continuous pathological uptake at the cup-bone-interface is a sign of loosening of the acetabular component. Finally, for diagnosing infection the blood pool images need to show a substantial lesion in the region of the lesser trochanter. A loosening of the acetabular cup is diagnosed in cases with a continuous pathological uptake in the cup-bone interface. In the case that a pathological uptake around the THR can be found in the additional blood pool

Fig. 7. Literature comparison of the results determining sensitivity and specificity of

further diagnosis of loosened THR (Fig. 5). The minimum sensitivity of nuclear arthrography was found by Miniaci et al. (Miniaci et al., 1990), whereas the minimum specificity of nuclear arthrography was determined to be 75 % (Herzwurm et al. 1991). Studies using arthrography often do not discriminate between cemented or uncemented components (Maus et al., 1987). The mentioned results are summarized in Fig. 5.

Fig. 5. Literature comparison of the results determining sensitivity and specificity of loosening diagnosis of the femoral component of the THR using nuclear arthrography

Temmermann et al. evaluated the mean sensitivity/specificity of several studies with nuclear arthrography (85 %/83 %) and subtraction arthrography (86 %/85 %) (Temmerman et al., 2005). Hence, these results show only a 4 % higher sensitivity/specificity than plain radiographs. For this reason arthrography does not add a significant benefit to the plain radiography results for the THR. Arthrography gets more and more abandoned, because of the risk of infection and the absent additional benefit, in case of aseptic loosening.

### **2.3 Scintigraphy**

Examination of an infected THR can be performed by using scintigraphy, which is often combined with white blood cell imaging (Larikka et al., 2001). Scintigraphy uses radioisotopes as body radiation source by producing a two-dimensional picture using a gamma-camera for the detection of the radiation (Katz et al., 1986; Zilkens et al. 1988) (Fig. 6). Usually, a two-phase or triple-phase bone scan is generated (Segura et al., 2004). The triple-phase bone scan is an enhancement with a reputation of high sensitivity (Reinartz 2009) and uses gallium scans to improve its specificity (Kraemer et al., 1993). However the data in the literature about its diagnostic efficiency show a considerable variability. This inconsistency is caused by the use of highly differing scan interpretation criteria.

Scintigraphic criteria are based on the suggestion of Wilson et al. (Wilson et al., 1997). Mechanical loosening of the femoral component should show a significant pathological uptake of radioisotopes in the distal part of the THR at the tip and increased periprosthetic

further diagnosis of loosened THR (Fig. 5). The minimum sensitivity of nuclear arthrography was found by Miniaci et al. (Miniaci et al., 1990), whereas the minimum specificity of nuclear arthrography was determined to be 75 % (Herzwurm et al. 1991). Studies using arthrography often do not discriminate between cemented or uncemented

components (Maus et al., 1987). The mentioned results are summarized in Fig. 5.

Fig. 5. Literature comparison of the results determining sensitivity and specificity of loosening diagnosis of the femoral component of the THR using nuclear arthrography

the risk of infection and the absent additional benefit, in case of aseptic loosening.

inconsistency is caused by the use of highly differing scan interpretation criteria.

**2.3 Scintigraphy** 

Temmermann et al. evaluated the mean sensitivity/specificity of several studies with nuclear arthrography (85 %/83 %) and subtraction arthrography (86 %/85 %) (Temmerman et al., 2005). Hence, these results show only a 4 % higher sensitivity/specificity than plain radiographs. For this reason arthrography does not add a significant benefit to the plain radiography results for the THR. Arthrography gets more and more abandoned, because of

Examination of an infected THR can be performed by using scintigraphy, which is often combined with white blood cell imaging (Larikka et al., 2001). Scintigraphy uses radioisotopes as body radiation source by producing a two-dimensional picture using a gamma-camera for the detection of the radiation (Katz et al., 1986; Zilkens et al. 1988) (Fig. 6). Usually, a two-phase or triple-phase bone scan is generated (Segura et al., 2004). The triple-phase bone scan is an enhancement with a reputation of high sensitivity (Reinartz 2009) and uses gallium scans to improve its specificity (Kraemer et al., 1993). However the data in the literature about its diagnostic efficiency show a considerable variability. This

Scintigraphic criteria are based on the suggestion of Wilson et al. (Wilson et al., 1997). Mechanical loosening of the femoral component should show a significant pathological uptake of radioisotopes in the distal part of the THR at the tip and increased periprosthetic uptake. A second substantial lesion in the region of the lesser trochanter is a further sign for loosening and supports the first presumption.

Fig. 6. 99mTc bone scan with normal anterior and posterior image of the prosthesis, without signs of loosening or infection (Kisielinski et al., 2003)

A continuous pathological uptake at the cup-bone-interface is a sign of loosening of the acetabular component. Finally, for diagnosing infection the blood pool images need to show a substantial lesion in the region of the lesser trochanter. A loosening of the acetabular cup is diagnosed in cases with a continuous pathological uptake in the cup-bone interface. In the case that a pathological uptake around the THR can be found in the additional blood pool

Fig. 7. Literature comparison of the results determining sensitivity and specificity of loosening diagnosis of THR using bone scintigraphy

Current Possibilities for Detection of Loosening of Total Hip

(DeHeer et al., 2001; Kadoya et al., 1998).

loosening diagnosis of THR using FDG-PET

not reached the status of a standard examination.

Replacements and How Intelligent Implants Could Improve Diagnostic Accuracy 371

glucose by emitting positrons for the generation of cross-sections (Fig. 8). FDG-PET is one of the most expensive imaging modalities and has already been documented to be a precise method to diagnose infection particularly (Chryssikos et al., 2008). Active cells like leukocytes and macrophages show a higher energy demand compared to other cells and therefore preferred ingest the FDG (Reinartz, 2009). Due to this fact, infected tissue with a high number of active cells revealed an increased uptake of FDG, which leads to a positive finding in the PET scan (Stumpe & Strobel, 2006). Furthermore, higher FDG uptake in patients with aseptic loosening of the THR can be explained by wear debris and the development of granulomatous tissue in the periprosthetic membrane around the THR

The criteria for the interpretation of a PET scan differ between different studies (Reinartz et al., 2005). In the area of the femoral component an increased uptake of FDG can be interpreted as loosening or infection, while an increased uptake at the femoral head and neck cannot be determined as loosening or unspecific (Zhuang et al., 2007). In contrast the increased uptake of FDG at the distal tip of the stem is unspecific. Furthermore, a high uptake at the implant-bone interface of the acetabular component can be identified as a

Based on the great variety of interpretation criteria to analyse PET scans, sensitivities of PET to implant loosening differ from 33 % (Stumpe et al., 2004) to 100 % (Zhuang et al., 2002), while specificity varies from 78 % (Vanquickenborne et al., 2003) to 100 % (Zhuang et al.,

2002) with an average of 85 %/90 % and an accuracy of 89 % (Reinartz, 2009) (Fig. 9).

Fig. 9. Literature comparison of the results determining sensitivity and specificity of

Figure 10 shows the comparison of the mean sensitivities and specificities of the presented imaging methods to diagnose loosening of the THR. FDG-PET has both, the highest sensitivity and specificity and therefore is the most accurate method to diagnose loosening exactly. Although the high accuracy, FDG-PET is highly cost-intensive, for that reason it has

**2.5 Summary of actual loosening diagnostics using imaging techniques** 

pathologic process or a sign of loosening (Mayer-Wagner et al., 2009).

image the diagnosis of infection and septic loosening is positive. Due to a normal bone scan, the existence of a sepsis can be excluded and an additional white blood cell imaging should be taken into account.

Mean sensitivity and specificity of scintigraphy are at 85 % and 72 % (Temmermann et al. 2005) (Fig. 7). Reinartz found a mean sensitivity/specificity of 78 % /84 % for diagnosing THR loosening with triple-phase bone scans (Reinartz 2009). It was also concluded, that bone scans of THR's are highly sensitive but not specific (Segura et al., 2004). The lowest sensitivity was found by Zilkens et al. with a value of 50 % (Zilkens et al., 1988), while the lowest specificity was 38 %(Ovesen et al., 2003).
