18F-FDG PET/CT and 99mTc-HMPAO-WBC SPECT/CT effectively contribute to early diagnosis of infection of arteriovenous graft for hemodialysis

Objective. An arteriovenous graft (AVG) is indicated in hemodialysis patients with failed arteriovenous access. Early treatment of AVG infection is important because an advanced prosthetic infection leads to the removal of the prosthesis. The aim of this study was to evaluate the benefits of 18F-FDG PET/CT and 99mTc-HMPAO-WBC SPECT/CT in early detection of AVG infections. Subjects and Methods. Fifty-one AVGs were evaluated. 18F-FDG PET/CT and 99mTc-HMPAO-WBC SPECT/CT studies were performed at intervals of 10, 20–30, and 40–50 weeks after AVG insertion. Agreement between the imaging methods and reference parameters (i.e. clinical presentation, C-reactive protein and microbiological findings on the hemodialysis cannula extracted after hemodialysis from AVG) was evaluated. Results. The study results showed that focal accumulation of the radiopharmaceuticals can be considered a sign of AVG infection. At 10 weeks after AVG implantation, the focal 18F-FDG findings showed the best agreement with the reference parameters (agreement coefficients AC1 clinical status: 0.693, CRP: 0.605, cannula microbiology: 0.518, respectively). At 20 to 30 weeks after AVG implantation, the diagnostic value of focal 99mTc-HMPAO-WBC accumulation increased (AC1 coefficients: 0.658, 0.658, 0.408) and was similar to that of focal 18F-FDG uptake (AC1s: 0.656, 0.570, 0.409). Between 40 and 50 weeks since AVG implantation, the diagnostic significance of focal 99mTc-HMPAO-WBC accumulation (AC1 coefficients: 0.771, 0.811, 0.611) slightly exceeded the diagnostic value of focal 18F-FDG accumulation (AC1 coefficients: 0.524, 0.456, 0.569). Conclusion. 18F-FDG PET/CT and 99mTc-HMPAO-WBC SPECT/CT can both serve as important tools contributing to early diagnosis of AVG infection.


IntroductIon
For patients on chronic hemodialysis treatment, vascular access is required.A native arteriovenous fistula is preferred as the most favorable.An arteriovenous graft (AVG) is indicated in patients with a failed arteriovenous fistula, exhausted superficial venous bed, or unsuitable vessels.Most commonly, expanded polytetrafluorethylene (ePTFE) prostheses are used.AVG is often a last option to ensure high-quality access for hemodialysis.
Infection of the AVG is a serious and life-threatening complication.It is associated with high morbidity and is the second leading cause of vascular access failure 1 .AVG infection rate ranges from 9.5% to 35% in the literature [2][3][4][5] .If a full-blown prosthetic infection occurs, the treatment options are quite limited and it mostly leads to the removal of the prosthesis; thus, access for hemodialysis is lost.
Our aim was to evaluate the benefits of fluorine-18-fluorodeoxyglucose positron emission tomography/ computed tomography ( 18 F-FDG PET/CT) and technetium-99-hexamethylpropyleneamine oxime labeled white blood cells ( 99m Tc-HMPAO-WBC SPECT/CT) in the detection of AVG infection and to compare these methods in situation when AVGs were altered by repeated punctures.We attempted to ascertain the suitability of these procedures for the early detection of AVG infection in hemodialysis programs -whether their additional use could bring desirable benefits, such as a more rapid and more precise diagnosis of AVG infection with a conservative treatment option, a reduced risk of advanced infection, and salvage of vascular access.The current graft infection rate in our group of patients was compared to a similar group of hemodialysis patients from a previous similar period before the introduction of radionuclide methods.

Patients
A prospective, open-label study included all patients in whom ePTFE prostheses for hemodialysis were implanted during a period of 3.5 years in total.The study was approved by the Ethics Committee of the University Hospital Olomouc and Faculty of Medicine and Dentistry, Palacky University Olomouc, Czech Republic and each patient signed an informed consent form.Patient data were handled anonymously.
The group was comprised of 56 patients in whom 62 AVGs for hemodialysis were created.There were 22 women and 34 men, with a mean age of 62.7 years (range 29-89 years).Eleven patients had no radionuclide examinations at all for the following reasons: death (4 cases), non-compliance (5), thrombosis (1), and early infection of AVG (1).51 AVGs were evaluated by using both radionuclide methods.
During the first year since AVG implantation, patients with an AVG for hemodialysis were monitored by followup protocol for 50 weeks in total.The monitoring period was shortened in case of infected graft removal or due to non-compliance of some patients -one-year monitoring was completed in 30 AVGs. 18F-FDG PET/CT was done 108 times and 99m Tc-HMPAO-WBC scintigraphy 101 times.

the follow-up protocol
Once a week, clinical evaluation of the patients was performed, blood samples were taken and the C-reactive protein (CRP) level was investigated as a marker of inflammation.
Microbiological testing was performed in each patient every week.A swab was sampled from the hemodialysis cannula (inserted in AVG during hemodialysis) immediately after the hemodialysis procedure, i.e. directly after its contact with the vascular prosthesis.In the case of suspected infection, the swab was taken directly from the most probable location of infection.
99m Tc-HMPAO-WBC SPECT/CT and 18 F-FDG PET/ CT were performed at intervals of 10, 20-30, and 40-50 weeks after AVG insertion.The interval between both examinations was 1 day in 88% cases, maximum delay was 16 days (due to health issues).The radiopharmaceuticals were injected intravenously into the contralateral arm.
18 F-FDG PET/CT images were acquired on the PET/ CT scanner (Biograph 16 Hi-Rez, Siemens Erlangen, Germany) allowing 3D-only PET acquisition.All patients were advised to fast for six hours prior to the PET/CT examination.5.7 MBq/kg (0.154 mCi/kg) of 18 F-FDG was injected intravenously; thereafter, patients rested for a period of 60 minutes.PET scans of the upper extremities (ranging from the neck to the hand) were acquired for 2-3 min per bed position, with a longer acquisition time in patients with a body weight above 100 kg.PET data were reconstructed into transaxial slices with a 168 × 168 matrix (pixel size 4 mm) and slice thickness of 2 mm using iterative 3D OSEM reconstruction (4 iterations, 8 subsets, 5 mm Gaussian post-reconstruction filtration).Multislice, diagnostic quality CT with intravenous contrast medium (80 ml of Ultravist® 370 mg/mL) was acquired, the CT settings were as follows: quality reference 130 mAs (CareDose4D on), 120 kV, beam width 16 × 1.5 mm, 99m Tc-HMPAO-WBC SPECT/CT scintigraphy was performed using the double detector SPECT/CT gamma camera (Infinia Hawkeye 4, GE Healthcare, Milwaukee, USA).Planar and SPECT/CT images were obtained 4 h after administration of 550 MBq (14.9 mCi) of 99m Tc-HMPAO-WBC; second planar scintigrams were acquired 24 h post injection.A pathologically increased uptake of radioactivity and its increase in time were evaluated as a criterion of inflammatory changes.

data evaluation
All scintigraphic images were evaluated independently by two experienced nuclear medicine physicians aware of the patients' clinical history and results of all the other examinations.
PET/CT and SPECT/CT studies were assessed on the basis of qualitative analysis of FDG and WBC uptake patterns.Increased accumulation of FDG and WBC in the prosthesis was classified as intense focal, diffuse, and simultaneous focal and diffuse.The finding was considered as a diffuse uptake if the majority of the graft was affected.The group of findings with any increase in the accumulation of FDG or WBC, i.e. with focal or diffuse accumulation, was also evaluated as part of the statistical analysis.

Statistics
Agreement between the methods was evaluated using the Gwet's agreement coefficient AC1 and percentage of the overall agreement.The strength of agreement between the methods was characterized using Gwet's coefficient AC1 -strength of agreement: < 0.0 Poor, 0.0 to 0.20 Slight, 0.21 to 0.40 Fair, 0.41 to 0.60 Moderate, 0.61 to 0.80 Substantial (Good), 0.81 to 1.00 Almost Perfect (Very Good).The McNemar test for symmetry was used to detect bias; in the case of statistically significant bias between the methods compared, these combinations of methods were excluded from further evaluation of agreement.

rESultS
Clinical manifestations of advanced infection developed in 7 out of 62 AVGs (11.3%), graft infection was also ascertained by radionuclide examinations as well as by increased CRP levels.All grafts had to be removed and access for hemodialysis was lost.The diagnosis of infection was verified by microbiological assays obtained at surgery.
In another 7 AVGs (11.3%), infection was suspected on the basis of positive 18 F-FDG PET/CT and a positive microbiological swab from the cannula.In these grafts, the clinical signs of graft infection were missing or very subtle.All these patients had a nonspecifically increased marker of inflammation -the CRP levels were slightly elevated (values 35-62).These patients were treated as potentially infectious using targeted antibiotic therapy.In all 7 cases, the suspicious infection was treated successfully, the possible development of clinically apparent infection was prevented, and access for hemodialysis could be rescued; the AVGs were fully functional and without signs of infection during the follow-up.Thus, no microbio-logical findings directly from a removed prosthesis were available.
Due to these facts, it was not possible to use a uniform "gold standard parameter" to demonstrate AVG infection in the study.Therefore, to assess the diagnostic significance of different findings of 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT at three time periods, an evaluation of agreement between the imaging methods and reference parameters observed (clinical presentation, CRP, and microbiological test result of the hemodialysis  It is obvious from Table 1 that the best agreement between the different reference parameters is the one between the clinical presentation and CRP.Furthermore, it is evident that, particularly in a longer time interval since AVG implantation, the positivity of the microbiological examination of the hemodialysis cannula outnumbered that of the clinical finding or the CRP in a substantial proportion of the results.In the setting of a higher positivity rate of this microbiological testing than would correspond to the clinical presentation and the CRP elevation, it is necessary to seriously consider a high probability of false positivity due to contamination of the samples taken.It could be supposed that a microbial smear from the cannula could be falsely positive due to long-lasting partially cutaneous insertion of the cannula. At 10 weeks after AVG creation (Table 2a), the best agreement between the clinical presentation and imaging was found for a focal 18 F-FDG PET/CT finding, while the McNemar test did not suggest a significantly higher rate of positive results in either of the methods.The worst agreement with the clinical presentation was shown for the finding of diffusely increased 18 F-FDG accumulation or diffusely increased 99m Tc-HMPAO-WBC accumulation.
The greatest agreement between the CRP and the imaging method was found for a simultaneous focal and diffuse 18 F-FDG PET/CT finding; however, the McNemar test (0.039) demonstrated significant bias in the resultsimaging with 18 F-FDG PET/CT yielded a positive result more often than the CRP -it could indicate an increased risk of false positive results.A good agreement between the CRP and the imaging method was also shown for a focal 18 F-FDG PET/CT finding alone; moreover, the McNemar test (0.508) showed that in this case, there was no significant bias in the results.A focal 18 F-FDG PET/ CT finding alone can thus be considered to be the best parameter in diagnosing infection when compared with the CRP.The worst agreement with the CRP was shown for diffuse 99m Tc-HMPAO-WBC hyperaccumulation.
The best agreement between a microbiological test result of the hemodialysis cannula swab and the imaging method was found for a simultaneous focal and diffuse 18 F-FDG PET/CT finding; however, the McNemar test (0.065) suggested that microbiological testing yielded a positive result insignificantly more often than 18 F-FDG PET/CT.The worst agreement with a microbiological finding was shown for diffuse 99m Tc-HMPAO-WBC hyperaccumulation.
The results show that at 10 weeks since AVG implantation, the findings with 18 F-FDG PET/CT demonstrate a better agreement with the reference parameters than 99m Tc-HMPAO-WBC SPECT/CT.Diffuse hyperaccumula- tion of 99m Tc-HMPAO-WBC is a particularly nonspecific finding.
At 20 to 30 weeks after AVG creation (Table 2b), the best agreement between the clinical presentation and imaging was found for simultaneous diffuse and focal finding with 99m Tc-HMPAO-WBC SPECT/CT while the McNemar test (1.0) did not suggest a systematically higher rate of positive results in either method.A similarly good agreement with the clinical presentation and no bias (McNemar test 1.0) was shown for a focal 99m Tc-HMPAO-WBC SPECT/CT finding.A focal 18 F-FDG PET/CT finding alone also showed a good agreement; however, the McNemar test (0.070) suggested an insignificantly higher rate of positive results with PET/CT.
The best agreement between the CRP and the imaging method was shown for a simultaneous focal and diffuse 99m Tc-HMPAO-WBC SPECT/CT finding, simultaneous focal and diffuse 18 F-FDG PET/CT finding, as well as focal 99m Tc-HMPAO-WBC SPECT/CT finding alone.McNemar test excluded asymmetric distribution of results in these cases.
The best agreement between a microbiological test result of the hemodialysis cannula swab and the imaging method with no significant bias in the results according to the McNemar test was shown for a focal 18 F-FDG PET/ CT finding.The worst agreement with a microbiological finding was shown for tests with diffuse hyperaccumulation of the radiopharmaceuticals ( 99m Tc-HMPAO-WBC and 18 F-FDG).
The results imply that at 20 to 30 weeks after AVG implantation, the diagnostic value of focal 99m Tc-HMPAO-WBC hyperaccumulation increases and is similar to that of focal 18 F-FDG PET/CT hyperaccumulation.
At 40 to 50 weeks after AVG creation (Table 2c), the greatest agreement between the clinical presentation and the imaging was found for a focal 99m Tc-HMPAO-WBC SPECT/CT finding.An identically good agreement with the clinical presentation was shown for a simultaneous focal and diffuse 99m Tc-HMPAO-WBC SPECT/CT finding.
McNemar test excluded asymmetric distribution of results in both cases.The worst agreement with a clinical finding was shown for a diffuse increase in the accumulation of both radiopharmaceuticals.
The best agreement between the CRP and the imaging method without asymmetric distribution of results was shown for focal or focal and diffuse accumulation with 99m Tc-HMPAO-WBC.The worst agreement with the CRP was shown for the findings with a diffuse increase in radiopharmaceutical accumulation with 18 F-FDG PET/ CT and 99m Tc-HMPAO-WBC SPECT/CT.
The best agreement between microbiological test result of the hemodialysis cannula swab and the imaging method was shown for a focal and diffuse 18 F-FDG PET/ CT finding (without significant asymmetry of results -McNemar 0.375).A lower agreement with a microscopy finding was shown for a diffuse increase in radiopharmaceutical accumulation with 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT.
The results show that between 40 and 50 weeks since AVG implantation, the diagnostic significance of focal 99m Tc-HMPAO-WBC hyperaccumulation increased -this finding slightly exceeds the diagnostic value of focal 18 F-FDG accumulation.
When comparing the maximum values (Gwet's agreement coefficient) in the individual periods (Table 2a-c), it is clear that the greatest diagnostic benefit is provided by imaging methods with a time distance of no less than six months since AVG creation.In this period, healing of the AVG had already occurred.

dIScuSSIon
The clinical manifestation in the early stages of infection is usually minimal and if some symptoms do appear, they are nonspecific.This is sometimes referred to as subclinical infection, when clinical signs are mild, laboratory findings may show increased markers of acute inflammation, and blood cultures are negative 6 .Common signs of inflammation are late only and carry a poor prognosis for the fate of the AVG.A lack of signs and symptoms does not exclude the possibility of a clinically silent graft infection 7 .
Only a few studies have addressed the question of early detection of infection with the rescue of hemodialysis vascular access 8,9 .Early detection and early diagnosis of AVG infection in a phase of subclinical infection is essential for the correct choice of treatment; a reliable diagnostic method is necessary in order to provide adequate treatment before the development of obvious clinical signs as well as to avoid further complications. 18F-FDG PET/CT is emerging as a promising tool in detecting prosthetic infection in vascular surgery, with a number of studies with excellent results having been published [10][11][12][13][14] .According to the literature, 99m Tc-HMPAO-WBC continues to be successfully used in the diagnosis of bacterial infections of vascular prosthesis 15 .
Currently, there are no published studies or case reports of using radiopharmaceutical methods in a patent AVG that is both functional and utilized for hemodialysis.It can be assumed that the reason for the existing underuse of radionuclide imaging techniques is the fear of false positivity of the findings due to the accumulation of radiopharmaceuticals at the venipuncture sites 16 .We tried to minimize this risk by using the longest possible time between hemodialysis and radionuclide examination and dual-time acquisition of 99m Tc-HMPAO-WBC SPECT.
It is obvious from our results that both imaging methods, i.e. 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT, exhibit significant agreement with the results of our reference methods used to detect early AVG infection (clinical presentation, microbiological test result of the hemodialysis cannula swab, and CRP).For these comparisons, the values of Gwet's agreement coef-ficient (AC1) reach the levels of "Substantial (Good)" to "Perfect (Very Good)".
These values were ascertained only if focal (or focal and diffuse) accumulation of radiopharmaceuticals was considered a sign of AVG inflammation (Fig. 1).This fact is consistent with the results of previous 18 F-FDG PET/CT studies as well as 99m Tc-HMPAO-WBC SPECT/ CT studies evaluating the importance of scintigraphy in detecting infection of vascular prostheses 11,12,15 .The nonspecific character of a diffusely increased accumulation of 18 F-FDG was described by Keidar et al. who found that this finding was observed in up to 92% of noninfected vascular prostheses 10 .
Moreover, our results are consistent with the previous publications that evaluated each method on its own and demonstrated a major benefit of both methods in detecting infection of vascular prostheses 11,12,15 .Of great importance is the fact that the diagnostic accuracy of 99m Tc-HMPAO-WBC SPET is increased significantly with the use of SPECT/CT, in which a more precise locating of increased leukocyte accumulation using CT improves the interpretation of the SPET findings 13,14 .
Our study deals with the early detection of infection of vascular prostheses using 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT.It is of particular importance that both methods were compared directly as part of a single study.
In early phases after AVG implantation (10 weeks), a higher level of agreement with our reference parameters of infection was achieved with 18 F-FDG PET/CT than with 99m Tc-HMPAO-WBC SPECT/CT imaging.In the following period of 20 to 30 weeks, the benefit of imaging with 99m Tc-HMPAO-WBC SPET/CT reached the level of that with 18 F-FDG PET/CT imaging, and in the period of 40 to 50 weeks since surgery, a focal 99m Tc-HMPAO-WBC SPECT/CT finding was the most significant one.The reason for the above-described trend cannot be clearly elicited from our cohort.Nevertheless, it can be assumed that the healing of the AVG continues with increasing time since surgery and the conditions for detecting infection using 99m Tc-HMPAO-WBC are improved.
In a comparable time period between the years 2006 and 2009, the rate of AVG infections in our department was 28.3%; all of them resulted in AVG removal and loss of access for hemodialysis 5 .The monitoring of a group of hemodialysis patients with AVG, using the above-described protocol, led to an early detection of AVG infection and allowed for a timely commencement of targeted antibiotic treatment, which resulted in an overall substantial decrease in advanced AVG infections to 11.3%.
The most significant effect of careful patient monitoring using 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT was the possibility to commence antibiotic treatment already in early stages of infection and to prevent it from developing fully.A positive effect of this active therapeutic approach was a reduction in the number of AVGs removed due to advanced infection.It is therefore obvious that in accordance with Husmann et al's conclusion that 18 F-FDG PET/CT represents a useful tool in therapy monitoring of prosthetic vascular graft infection and has an impact on patient management 15 .

concluSIon
The study results show that focal hyperaccumulation of both 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC can be considered a sign of AVG infection. 18F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT can both serve as important tools contributing to early diagnosis of AVG infection.By direct comparison of both hybrid methods, we have found that examination with 18 F-FDG PET/CT appears to be more effective in the first months after AVG creation.Both methods have the similar efficiency approximately half a year after AVG implantation, while 99m Tc-HMPAO-WBC SPECT/CT is slightly more beneficial in the later period.
Only timely administration of antibiotics is capable of preventing the development of advanced AVG infection.If the results of all individual tests are evaluated comprehensively, i.e. in collaboration of a vascular surgeon, microbiologist, nephrologist, and nuclear medicine physician, AVG infection can be detected in time and with appropriate treatment, the number of AVGs removed due to infection can be reduced significantly.
cannula swab) was used, while excluding combinations shown to have a significant bias between the two compared methods (McNemar test).

table 1 .
Evaluation of agreement between the clinical signs, microbiological test result of the hemodialysis cannula swab and CRP at 10, 21-30, and 41-50 weeks after AVG insertion.
AC1 -Gwet's agreement coefficient, Method 1 -the first method in the comparison line, Method 2 -the second method in the comparison line, McNemar -McNemar test for symmetry, Clinical -clinical signs, Micro -microbiological test result of the hemodialysis cannula swab, CRP -C-reactive protein.

table 2a .
Evaluation of agreement between reference parameters and diffuse or focal 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT findings at 10 weeks after surgery.

table 2b .
Evaluation of agreement between reference parameters and diffuse or focal 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT findings at 20 to 30 weeks after surgery.

table 2c .
Evaluation of agreement between reference parameters and diffuse or focal 18 F-FDG PET/CT and 99m Tc-HMPAO-WBC SPECT/CT findings at 40 to 50 weeks after surgery.