Clinical utility of technetium‐99m‐labelled leukocyte scan in diagnosis of vascular infections

The diagnosis of vascular graft infections and infected aneurysms via conventional methods such as computed tomography (CT) and ultrasound (US) can often be challenging. Nuclear medicine imaging techniques can detect infection in these difficult cases. We aim to evaluate the diagnostic value of technetium‐99m‐labelled leukocyte scans in diagnosing vascular graft infections in our centre.


Introduction
Vascular infections particularly when associated with a vascular prosthesis or graft are known to carry a high morbidity and mortality rate. 1 Graft infections are difficult to treat and can lead to multiple complications like amputation and are associated with high 10-50% morbidity rates and a mortality rate of around 60%. [1][2][3] Fortunately, the lifetime incidence of vascular graft infection reported is around 1-6%. 1,3,4 The clinical presentations of vascular infections are often insidious and common presentations include lowgrade pain, mild swelling and malaise. 5 Furthermore, tissue sampling of the graft and adjacent tissues is often difficult and has its own inherit risks particularly when the management involves a graft retention strategy and endoluminal procedures to exclude the lesion.
Conventional imaging such as computed tomography (CT) and ultrasound (US) can be used to diagnose vascular infections, such as mycotic aneurysm, aortic graft infections and autologous and prosthetic graft infections. These techniques are well established at detecting advanced infections. 6 The dilemma arises in patients with suspected vascular infection and equivocal findings on conventional imaging.
In 2016, the Management of Aortic Graft Infection Collaboration (MAGIC) proposed a case definition for aortic graft infection to provide a diagnostic standard to help further research into this clinical situation (Table 1). This definition uses a combination of various major and minor surgical, radiological and laboratory findings. 7 Prior to this, there was no general consensus, and the diagnosis of infection would vary among institutions. 2 Nuclear medicine imaging techniques such as technetium-99m-labelled leukocyte scan use the localization of radiolabelled white cells to areas of inflammation and can be a useful adjunct in immunocompetent patients. 8,9 However, there are certain challenges in diagnosing vascular infections via technetium-99mlabelled leukocyte scan as there is no clear distinguishing pattern between infectious and non-infectious processes. 8 There are limited studies looking at the sensitivity and specificity of technetium-99m-labelled leukocyte scans for diagnosing vascular infections and early diagnosis of infection is important for further management. 5 The aim of this study was to evaluate the diagnostic value of technetium-99m-labelled leukocyte scan in diagnosing vascular graft infections in conjunction with and adapted MAGIC case definition to include suspected cases of aortic, extra-aortic and native arterial infection.

Methods
The study was registered with the research office and the conduct of the study was approved. At our institution all technetium-99m-labelled leukocyte Scans that have been performed share the same code 'NMWBC' on the Radiology Information System (RIS) and Picture Archiving and Communication System (PACs). This code was used to identify patients who underwent a leucocyte scan over between 1 January 2004 and 1 July 2020.
A total of 164 patient-scans were identified, and an extensive review of their electronic clinical records, laboratory results and imaging was performed by two independent investigators.
The indications for scanning were split into three categories: suspected mycotic aneurysm, infected aortic graft and other infected graft. Baseline data such as age, sex, date and type of initial surgery were collected and summarized in Table 2.
There were 11 patients from 2004 to late 2009 that were excluded as the imaging could not be reviewed as PACs were not available until late 2009; only the reports were available. A further 121 patients were excluded due to a non-vascular indication or missing information, that is required to use the adapted MAGIC criterion for evaluation, summarized in Figure 1. The remaining 43 patients had their clinical and laboratory data reviewed by one reviewer who remained blinded to any radiological findings. This reviewer was a registrar working in the

Case definition
To define an infected case we adapted the criteria proposed by the Management of Aortic Graft Infection Collaboration (MAGIC) published in 2016. The study suggested using findings from three categories (clinical/surgical, radiology and laboratory) to determine whether patients had an infected aortic graft. A diagnosis of infection is made when there is a presence of a single major criterion and any other criteria (major of minor) from another category. A patient is deemed to have suspected graft infection with any isolated major criterion or minor criterion from two of the three categories. 7 It should be noted that MAGIC criteria does include the results of leucocyte scans as a minor radiological criterion. For the purposes of our study the results of the leucocyte scans were not included when applying MAGIC criteria to identify infected cases. This was to prevent any potential bias when evaluating the ability of a leucocyte scan to identify vascular infection.

Technetium-99m-labelled leukocyte scan protocol
All leucocyte scans were performed locally on either GE Infinia VC Hawkeye camera (CT-ATC) or Millennium MG (non-ATC). Patients received an injection of Technetium-99m Granulozyt, 600MBd IV, nearly all Mab fragments. Imaging was obtained after a minimum of 3 h. Whole body SPECT/CT scan and targeted views were obtained. Focal sites of peri-graft increased uptake was determined as indicative of inflammation/infection and were reported by a nuclear medicine scientist. All diagnostic reports used SPECT/CT imaging to improve diagnostic accuracy (Fig. 2).

Results
There were 43 patient-scans identified with a vascular indication and complete clinical records to allow application of the adapted MAGIC criterion to determine infection.
Indication for imaging was suspected mycotic aneurysm in 5 patients, aortic graft infection in 22 patients and other non-aortic graft infection in 16 patients. The median age was 73 years old and 31 patients were male.
There were 32 patients found to fit the case definition of infection (74%). Leucocyte scan was positive in 27 patients and negative in 16 patients.
Correlative CT imaging was done in 37 patients. Only 8 patients had another form of imaging USS or MRI, one patient had both. Of those, 7 patients had positive WCC scans, and their USS or MRI results were either indeterminate or negative.
There were 25 true positive, 9 true negative, 2 false positive and 7 false negative, summarized in Table 3. In using technetium-99m-labelled leukocyte scan for diagnosing vascular infections, the sensitivity was 78. Subgroup analysis for patients who had the scan performed for aortic graft infection for which MAGIC was intended to be used (22 patients), sensitivity was higher 85.7% and specificity was 75%. In the other vascular graft infections (16 patients), sensitivity was 84.6% and specificity was 100%. In patients with no vascular prostheses or suspected native aortic infection (5 patients), sensitivity was 40%, and specificity was 100%.
Two false-positive cases were found in men both in their 70s with aortic grafts.
The first patient had a complicated history of cancer and possible uptake around the graft due to concurrent gastrointestinal and genitourinary infections. This can make interpretation difficult when there is uptake in close proximity to the graft. This patient was subsequently treated with antibiotics. The second patient had a previous endoleak post an endovascular aortic repair with uptake around the region of the leak. This patient was also treated with long-term antibiotics.

Discussion
There are a limited number of comparable studies specific to technetium-99m-labelled leukocyte scan in the setting of vascular graft infections and infected aneurysms, with earlier studies demonstrating sensitivity and sensitivities close to 100%. 6,10 This is due in part to no clear case definition of infection at the time and possibly a selection bias with a large proportion of the scans performed in patients with clinically obvious infection. 6 There is also some variability in scintigraphy scanning techniques, with no accepted standard of radiotracer or scan protocol. 6,8,10 Our study reported two false-positive cases, both of which were advised to continue treatment with lifelong antibiotics at vascular follow-up. False positives in other cases were found in cases of pseudo-aneurysms, intraprosthetic thrombosis, haematoma and lymphocele. 11 Some of the patients in this study have received prior antibiotic therapy. This could affect the laboratory results in some of the patients but would have affected the outcome of MAGIC criterion equally.
Limitations of this study include the retrospective design, small numbers and possible referral bias. As technetium-99m-labelled leukocyte scan is not commonly performed, there is only a small study sample in each subgroup. The number of leucocyte scans performed at our centre peaked in 2010 with 21 performed that year, over twice the yearly average of 10 scans over  the 16 years we reviewed. Our nuclear medicine department is small, and, ideally, we would conduct an independently review of the nuclear medicine scans by a dual-trained nuclear medical physician. It should be noted that the MAGIC criteria were proposed for the use in diagnosis of aortic graft infections only. However, its use has been extended in other studies for grafts outside the aorta. 5 We have chosen to include patients without a graft, that is aortic infection or mycotic aneurysms which it has not been validated for. However, the principles of the criteria are used in some degree already in this clinical context. Separate analysis of each indication group is outlined above and even when non-graft patients were excluded our sensitivity, and specificity was still high.
The European Society for Vascular Surgery published a recent guideline which helps address the inconsistency across the diagnosis and management of vascular graft and endograft infections. This guideline also uses the MAGIC criteria which we adapted.
This guideline proposes the use of PET-CT as the firstline nuclear medicine test for diagnosing vascular graft infections and recommends leucocyte scanning as an alternative. 5 Some of the primary reasons are related to increased availability of PET-CT; however, this is not the case in our centre. Previous studies have also shown similar leucocyte scans to be better than PET-CT. 11 If and when PET-CT becomes readily available a future study to compare the utility of the different imaging modalities will be useful to help guide local imaging guidelines.
In conclusion, technetium-99m-labelled leukocyte scan has a high level of sensitivity and specificity and is a useful tool in diagnosing suspected vascular infections.