Can^99mTc-MDP–SPECT/CT Differentiate Loosening and Infection After Hip and Knee Replacements?

Abstract

Background

Prosthetic loosening and infection are still common complications after joint replacement. Over the past few years, single-photon emission computed tomography–computed tomography (SPECT/CT) was widely used and showed unique value based on the combination of anatomic and metabolic information of foci. However, its performance in differentiating between prosthetic loosening and periprosthetic infection after joint replacement is still the focus of clinicians and deserves further investigation.

Purpose

This retrospective study was aimed to determine whether bone scintigraphy combined with SPECT/CT still can differentiate prosthetic infection from loosening in patients after joint replacement. The differential efficacy in hip and knee prosthesis was also analyzed. Blood biomarkers for the diagnosis of periprosthetic infection were also evaluated.

Patients and methods

Data sets of 74 prosthetic joints (including knees and hips), with suspected prosthetic loosening or infection between 2015 and 2021, were evaluated. Besides the results of nuclear imaging, X-ray images and serum biomarker were also recorded. Telephone follow-up and revision surgery after SPECT/CT were used as a gold standard. The sensitivity and accuracy of different imaging modalities were calculated by Chi-square test. The diagnostic efficacy of imaging methods and serum biomarkers were then analyzed by the area under curve (receiver operating characteristic curves, ROC) in SPSS 26.

Results

In all, 47 joints (14 knees and 33 hips) were confirmed as aseptic loosening, while 25 joints (18 knees and 7 hips) were confirmed as infection. The sensitivity and accuracy of SPECT combined with SPECT/CT imaging were the highest (92.86% and 87.84%, respectively). The differential efficacy of bone scintigraphy combined with SPECT/CT imaging was also better than any other single imaging modality. In the analysis of involved prosthesis, prosthetic loosening occurred more in hip prosthesis and knee prosthesis was easily infected (P < 0.05). Finally, the sensitivity of ESR and CRP were 80% and 84%, respectively.

Conclusions

Bone scintigraphy with hybrid SPECT/CT remains encouraging in differentiating prosthetic infection from loosening after joint replacement. The diagnostic efficacy of differentiation in hip prosthesis was better than knee. Serum biomarkers cannot be used alone to differentiate prosthetic infection from loosening.

Introduction

Nowadays, it is a challenge for orthopedic surgeons to differentiate infection from loosening of prosthesis after surgery.

The major diagnostic imaging methods for prosthetic loosening or periprosthetic infection include X-ray, computed tomography (CT), magnetic resonance imaging (MRI) and single-photon emission computed tomography (SPECT), but the efficacy was still under the expectations [1, 2]. Based on SPECT technology, bone scintigraphy, leukocyte scintigraphy, bone marrow scintigraphy can be used for the therapeutic assessment after joint replacement [3]. Among them, ^99mTc-MDP SPECT was the most widely used scintigraphy, which could visualize the local bone metabolism.

As fusion technique, SPECT/CT imaging integrates functional, metabolic, and anatomical information. It sensitively reflects the active bone metabolism around the prosthesis and accurately localizes the loosening or infection site. When prosthesis loosening, the discontinuous pressure between the prosthesis and the adjacent bone lead to increased osteogenesis, blood flow and bone metabolism. In the early stage of prosthetic loosening, metabolic changes occur, while the morphology is still normal. As a result, SPECT/CT could detect prosthetic loosening earlier than X-ray, CT, and MRI findings. However, it is reported that the value of nuclear imaging in visualization of low-grade periprosthetic infection was still limited [4]. Whether 99mTc-MDP SPECT/CT remain advantageous on differentiating prosthetic loosening and periprosthetic infection is still a hot topic for experts.

This retrospective study aimed to assess whether bone scintigraphy combined with hybrid SPECT/CT can provide encouraging results in differentiating prosthetic loosening and periprosthetic infection after joint replacement. Furthermore, the differential efficacy in hip and knee prosthesis was also analyzed, respectively. Blood inflammation biomarkers were also evaluated.

Methods

Patients

The study included 74 joints (from 28 males and 42 females, 45–85 years) between 2015 and 2021 in General Hospital of Ningxia Medical University (Flow chart of methodology is shown in Fig. 1). They were referred to the nuclear medicine department as for suspicious diagnosis of prosthetic loosening or infection. Final diagnosis was confirmed by successive revision surgery after SPECT/CT, pathological specimen, puncture fluid test or clinical follow-up. Traditional plain radiography and serum biomarker (ESR and CRP) were also involved into patients’ history. The exclusion criteria included: insufficient clinical and imaging data, history of malignant or systemic diseases, normal imaging findings. Ethical approval for the study was obtained from the General hospital of Ningxia Medical University ethics committees (Ethics Number: KYLL-2022-0554). Informed consent was obtained from all individual participants enrolled in the study.

Fig. 1.

Flow chart for methodology

Imaging Protocol

All patients underwent anteroposterior and oblique radiographs before radionuclide imaging.

Bone scintigraphy and SPECT/CT fusion image were performed on SPECT/CT system (Siemens Symbia T2), equipped with a low-energy high-resolution collimator, with a peak energy of 140 keV and an energy window of 20%. The dose of injection of ^99mTc-MDP was 20 mCi (740 MBq). The bone scintigraphy was obtained by scanning 200 cm at a rate of 20–25 cm/min with a matrix of 256 × 1024, followed by SPECT and CT tomographic acquisition. The SPECT, CT transverse, sagittal, and coronal tomographic images, and fused SPECT/CT images were obtained and constructed for further analysis.

Imaging Interpretation

Images were read by two specialists with extensive imaging experience who were blinded to all clinical and radiological details. if disagreements arose, consultation and consensus were sought.

Radiological imaging for loosening often shows consistent radiolucent lines at the prosthesis-bone/bone interface or bone–cement interface and radiolucent zones up to 2 mm in width and malposition; Infection manifests as irregular luminal periprosthetic opacities with obvious bone destruction, and usually progresses rapidly and aggressively [5].

Diagnostic criteria in nuclear medicine ① Normal: the interface of bone and prosthesis shows mild nuclide concentration or no radioactivity uptake; ② Prosthetic loosening: focal radioactivity is concentrated at stress site (large and small rotor, prosthesis end, tibial plateau, etc.) or prosthesis; ③ Periprosthetic infection: the interface of the prosthesis shows intense diffuse uptake with or without soft tissue changes (fluid collections, joint distension, inflammatory soft tissue edema, and fistulous tracts).

Final Diagnosis

The final diagnosis was established by phone follow-up and revision surgery after SPECT/CT. Diagnostic criteria for prosthesis infection [6]: ① Purulent secretions flowed out; ② Joint cavity aspiration or tissue biopsy showed positive culture; ③ Histopathological sections around the prosthesis confirmed the presence of acute inflammatory changes (more than 10 polymorphonuclear leukocytes or 5 pus cells per high-power field). If none of the above criteria were met and there were still no clinical symptoms after more than 6 months of clinical follow-up, the joint pain was considered to be caused by prothesis loosening.

Statistical Analysis

Data analysis was performed using SPSS (IBM, version 26.0). The continuous variable was presented using ‾X ± SD and the categorical variable was presented median. Chi-square test was employed to calculate the true-positive rate (TPR), false-positive rate (FPR), false-negative rates (FNR). The sensitivity, accuracy and positive predict value(PPV) were then calculated. Receiver operating characteristic curves (ROC) were drawn to compare the diagnostic efficacy of several imaging methods and serum biomarkers, and the area under the curve (AUC) was concluded to evaluate the diagnostic efficacy of imaging modalities (P < 0.05).

Results

The characteristics of 74 joints (70 patients) are listed in Table 1. Of these, two patients had bilateral hip replacements and the other two had bilateral knee replacements. This means that 42 hip and 32 knee prostheses were analyzed.

Table 1.

Demography

	Loosening (final diagnosis)	Infection (final diagnosis)	Normal (final diagnosis)
Male/Female (N/n)	18/25	9/16	1/1
Age (year)	65.74 ± 10.19	66.52 ± 8.51	77 and 81
Operation interval (months)	96 (24,144)	36 (12, 84)	156 and 216
Hip/Knee (N/n)	33/14	7/18	2/0
ESR (mm/h)	17 (6, 51)	28 (15, 43)	16 and 12
CRP (mg/l)	5.39 (1.14, 21.55)	11.3 (5.59, 37.70)	2.6 and 1.03

Data are n, mean (standard deviation), or median (25th, 75th percentile)

We evaluated all prosthesis regardless of the location of prosthesis. In bone scintigraphy, 59 (79.73%) cases showed prosthetic loosening, 12 (16.22%) cases showed periprosthetic infection and 3 showed normal; In SPECT/CT imaging, 55 (74.32%) cases were regarded as loosening and 19 (25.68%) infection of the prothesis. Typical images of prosthetic loosening and infection cases are shown in Fig. 2. In contrast, false-positive and false-negative cases are shown in Fig. 3.

Fig. 2.

Typical cases of prosthetic loosening and infection. a Bone scintigraphy of prosthesis loosening in the right hip. An abnormal uptake at right femur stem was shown. a1–a4 showed the foci SPECT/CT images in coronal view (a1 and a2) and transverse view (a3 and a4). Higher uptakes were shown at stress sites. b Bone scintigraphy of periprosthetic infection in the left knee. Diffused concentration at left upper tibia was shown; b1–b4 showed diffused uptake around tibia prothesis

Fig. 3.

Typical images of false-positive and false-negative cases. a False-positive case with diffused concentration at left knee in bone scintigraphy. a1–a4 inferred the concentration aiming at stress sites. b False-negative case with higher uptake at left knee in bone scintigraphy. b1–b4 showed the tomography images of region of interest (ROI). In addition, clinical history infers that this patient took antibiotics before bone scan

Furthermore, SPECT/CT revealed other abnormal uptakes showing soft tissue swelling in 9 patients, fracture in 3 patients and joint effusion in 2 patients. All the image findings were then compared with the final diagnosis (shown in Table 2). SPECT combined SPECT/CT imaging showed highest specificity, accuracy and PPV with 92.86%, 87.84%, and 94.2%, respectively. At the same time, the diagnostic efficacy of different imaging modalities in prosthetic loosening (Fig. 4) and infection (Fig. 5) were shown in ROC curves. In bone scintigraphy combined with SPECT/CT imaging, the accuracy of diagnosing prosthetic loosening and infection were 95.74% and 74.07%, respectively (P < 0.05). Fusion imaging combined with bone scintigraphy had highest efficacy with AUC 0.849 and 0.880, respectively. The significant differences between different imaging modalities were shown with χ² test (P < 0.05).

Table 2.

Diagnostic efficacy of X-ray, bone scintigraphy, SPECT/CT and bone scintigraphy combined with SPECT/CT

	N/n				Sensitivity (%)	Accuracy (%)	PPV (%)
	TP	TN	FP	FN
X-ray	32	0	2	40	44.44	43.24	94.12
SPECT	48	0	6	20	70.59	64.86	88.89
SPECT/CT	62	0	4	8	88.57	83.78	93.94
SPECT + SPECT/CT	65	0	4	5	92.86	87.84	94.20

Fig. 4.

ROC curve of several imaging modalities for the diagnosis of prosthetic loosening

Fig. 5.

ROC curve of different imaging modalities for the diagnosis of periprosthetic infection

The position of the prosthesis (hip and knee) was also analyzed in all patients. 42 hip prosthesis and 32 knee prosthesis were included in our study (P < 0.05), as shown in Table 3. What’s more, hip prosthesis was more likely to loosen than knee prosthesis (33/14), while knee prosthesis was more likely to be infected than hip prosthesis (18/7).

Table 3.

Diagnostic efficacy of bone scintigraphy combined with SPECT/CT in hip and knee prosthesis

	N/n				Sensitivity (%)	Accuracy (%)	PPV (%)
	TP	TN	FP	FN
hip	40	0	2	0	100.00	95.24	95.24
knee	25	0	2	5	83.33	78.13	92.59

To clarify the diagnostic efficacy in periprosthetic infection, we divided the infection into three phases (shown in Table 4). The late-onset infection occupied the most art.

Table 4.

SPECT/CT results of different phases of periprosthetic infection

	SPECT/CT (joint number-N/n)
	Infection	Loosening
Early onset infection	2	0
Delayed-onset infection	6	5
Late-onset infection	12	3

The diagnostic efficacy of ESR and CRP in periprosthetic infection was also evaluated in ROC curve (Fig. 6, P > 0.05). The sensitivity and specificity of ESR vs CRP were 80% and 77.6% vs 84% and 65.3%, respectively.

Fig. 6.

ROC curve of blood markers: ESR and CRP for diagnosing periprosthetic infection

Discussion

It is difficult to differentiate periprosthetic infection and prosthetic loosening after joint replacement. Several imaging modalities have been used for differentiating them and SPECT/CT is superior to other diagnostic methods. Our study aimed to evaluate whether bone scintigraphy combined with SPECT/CT could still differentiate periprosthetic infection from prosthetic loosening. The main finding of our study was that bone scintigraphy with hybrid SPECT/CT can still have an essential value on differentiating loosening and infection after joint replacement.

Our study demonstrated the highest sensitivity, accuracy, PPV and AUC in ROC analysis of bone scintigraphy combined with hybrid SPECT/CT than any other imaging modality (as shown in Table 2). The sensitivity was reached 92.86% because of the combined information of bone metabolism and morphology. The accuracy of diagnosing prosthetic loosening was 95.74%, which was similar to Bao’s study with 97% [7]. One of the most diagnostic features was the precise localization of the higher concentration of ^99mTc-MDP in simultaneous CT image. The abnormal uptake of ^99mTc-MDP in prosthetic loosening was almost spotted or round shape and seen at the stress site of prosthesis.

However, there were still false-positive and false-negative results in some cases. As shown in Fig. 3, when the prosthetic loosening was accompanied by bone fracture, it may be led to a misdiagnosis as infection. Part of the reasons for misdiagnosing prosthetic infection as loosening was the history of antibiotic therapy for other infectious disease before the imaging. which leading to the infection alleviated and tissue metabolism decreased, as explained in Pelin Arican’s study [8]. Interestingly, we analyzed different joints (hip and knee) of prosthetic loosening and infection. We found that knee prosthesis was more likely to be infected. Possible causes were as follows: first, knees were undertaken almost the whole-body weight; second, ankle abnormalities may lead to the compensation of the knee. In addition, our study showed more hip prosthesis involved as prosthetic loosening, which was paralleled with Li’s study [9]. They reported the sensitivity of SPECT/CT imaging was 100%. We speculated that the reason may be caused by mechanical loss of fixation over time, particle-induced osteolysis surrounding the prosthesis, and so on.

We also divided periprosthetic infection into 3 phases [10]. Early onset infection occurred within 3 months after joint replacement; delayed-onset infection occurred within 3–24 months after joint replacement; late-onset infection occurred 24 months after joint replacement. The results revealed that the SPECT/CT could easily differentiate the infection within 3 months after the surgery and even higher diagnostic ability along with time. The different bacterial pathogens can explain the images in different phases of periprosthetic infection [11] Pathogens in early onset infection are active; therefore, the patients tend to be infected at this stage, especially for elderly people with weak immunity.

Serum biomarkers can also be used to diagnose prosthetic infections [12]. Our study showed the sensitivity and specificity of ESR and CRP were lower than 85%. Previous studies demonstrated that the sensitivity of ESR in diagnosing infection was 70% [13], its specificity was 86%, and those of CRP were 68% and 87% [14], respectively. The possible factors, such as trauma, autoimmune disease, and surgery will reduce their sensitivity.

Our study had several limitations. First, the sample size needs to be enlarged. Second, it is promising to combined the hybrid SPECT/CT image with three-phase bone scan.

Conclusion

Bone scintigraphy combined with integrated SPECT/CT remains advantageous in differentiating prosthetic infection from loosening. The diagnostic efficacy of differentiation in hip prosthesis was better than knee. Serum biomarkers cannot be used alone to differentiate prosthetic infection from loosening.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article.

Declarations

Conflict of interests

On behalf of all authors, the corresponding author states that there is no conflict of interest.