Variability of [¹⁸F]FDG-PET/LDCT reporting in vascular graft and endograft infection

Vascular graft and endograft infection (VGEI) is a major complication of vascular surgery and is associated with high morbidity and mortality [1, 2]. The incidence of VGEI is difficult to assess, because the aetiology of this complication is complex and multifactorial including patient-related risk factors and pre-, intra-, and post-operative factors [3]. To improve early diagnosis and clinical outcomes, adequate treatment is important. However, diagnosis can be complicated due to the inability to take microbiological cultures because of a complex anatomical location or due to a subtle and non-specific clinical presentation [4, 5]. VGEI can present early (within 4 months) or late (> 4 months). Especially late VGEI can be challenging to diagnose due to lack of systemic signs of infection or elevated white blood cell count [6].

In patients with a suspected VGEI, computed tomography angiography (CTA) is usually the preferred imaging modality to be performed [6]. However, nuclear medicine modalities may be needed to confirm the diagnosis and to analyse the extent and possible spread of the infection [7]. Literature has shown that ¹⁸F-fluoro-D-deoxyglucose positron emission tomography with low-dose and/or contrast-enhanced computed tomography ([¹⁸F]FDG-PET/CT) scan reveals a high sensitivity for the diagnosis of VGEI, but it should be performed preferably at least 4 months post-operative to avoid false positive findings [3, 5, 7]. False positive results can be caused by physiologic FDG uptake due to a sterile inflammatory response after surgery [5, 7].

Reporting [¹⁸F]FDG-PET/CT scans of suspected VGEI is challenging, reader dependent, and report standards or interpretation criteria are still lacking. In contrast to VGEI, reporting standards on other specialities (e.g. oncology) are already available for a decade, are widely used, and are known to improve clinical outcomes [8‐10]. Interpretation of [¹⁸F]FDG-PET/CT scans in VGEI patients can be performed in many ways: (1) visually, by uptake pattern (focal vs diffuse), uptake intensity, uptake outside vessel boundaries, uptake in regional lymph nodes; and/or (2) semi-quantitatively by SUV measurements, by comparison (ratios) with for example blood pool or liver. Different interpretation criteria exist, but no standardisation of these criteria is accepted yet [11, 12]. Therefore, the aim of this study was to evaluate variability of [¹⁸F]FDG-PET/low dose CT (LDCT) reporting of suspected VGEI using a proposed standard reporting format based on findings in current literature and to evaluate if incompleteness of reports influenced the diagnostic accuracy.

In this retrospective study, we assessed the completeness of [¹⁸F]FDG-PET/LDCT reports of suspected VGEI patients based on ten predefined criteria as reported in current literature. Less than half of all [¹⁸F]FDG-PET/CT reports contained all criteria. The least frequently assessed criterion was the pattern of [¹⁸F]FDG uptake, despite its critical significance to determine the diagnosis of VGEI [12]. A sensitivity of 91% and specificity of 72% were found in the overall cohort, which is comparable with the existing literature [3, 5, 7, 16]. Furthermore, a higher sensitivity and specificity in fully evaluated reports compared to reports using fewer evaluation criteria were observed. This is an important finding since it demonstrates the additional value of striving for standardised reporting including all predefined criteria to increase the diagnostic accuracy. Accordingly, standardised and complete reports should be recommended in the new guidelines for patients with suspected VGEI.

The importance of the uptake pattern has been already addressed when [¹⁸F]FDG-PET/LDCT or [¹⁸F]FDG-PET/CTA are used in other clinical situations, such as in the diagnosis of infectious endocarditis and in patients with suspected infections after a Bentall procedure [17, 18]. In these settings careful assessment of the presence of persistent host versus biomaterial coating reaction, the sewing ring of the valve, chronic tension, or friction exerted on anchor points as well as of all the factors affecting the intensity of [¹⁸F]FDG uptake (i.e. time elapse from surgery, surgical and post-surgical complications, ongoing antimicrobial treatment, specific strains) has been demonstrated of fundamental to maintain high specificity when using [¹⁸F]FDG [19]. If the proper protocol for patients’ preparation and imaging acquisition are followed and specific imaging interpretation criteria are used, sensitivity and specificity can reach 91% in case of infective endocarditis and 97% and 73% in case of Bentall procedures [17, 18].

The presence of para-physiologic, [¹⁸F]FDG uptake along the wall of the vascular grafts representing reactive granulomatosis is often visible and validated semi-quantitative SUV cut-off points are lacking. Therefore, describing the uptake pattern remains of utmost importance [7, 16, 20]. The uptake pattern has a comparable sensitivity, but a significant higher specificity compared to the common description of the intensity of [¹⁸F]FDG-uptake against the SUVmax or tissue-to-background ratio (TBR) [16, 21]. Indeed, focal or heterogeneous uptake along the vessel is a hallmark of VGEI as compared to linear, diffuse, and homogenous uptake which does in general not represent infection [11, 22, 23]. Therefore, to increase the diagnostic accuracy of [¹⁸F]FDG-PET/LDCT in VGEI, it is necessary to provide a combination of visual uptake pattern with [¹⁸F]FDG-uptake intensity [7]. The some less frequently observed criterion of [¹⁸F]FDG-uptake pattern in the current study is maybe due to dated reports with less attention to uptake patterns in VGEI. The [¹⁸F]FDG-uptake intensity in our cohort was scored using a four-point scale which has been validated in our centre, used for several years, and has been recommended in previously published literature [7]. Recently, in 2015, Sah et al. proposed a new scoring method that consists of a five-point scale [24]. In the future, researchers in the field of [¹⁸F]FDG-PET should be aware of this and a comparison should be made between these two scoring methods.

A recently published study has shown that the presence of positive (defined as follows: visual uptake of grade two or four and/or a short axis diameter > 10 mm on LDCT) locoregional lymph nodes on [¹⁸F]FDG-PET/CT imaging has a high specificity (96%) and positive predictive value (95%) for VGEI [25]. However, these findings were accompanied by a low sensitivity. Therefore, the positive locoregional lymph nodes could have a positive influence on the specificity of new interpretation criteria. The current study corroborates to the conclusion that further research is needed to evaluate the diagnostic accuracy of lymph nodes for detecting and diagnosing VGEI, as positive lymph nodes in the area surrounding the vascular graft were observed in only 14.6% of the patients with a diagnosed VGEI.

In the first half of the study period, an increase in both numbers of reports and percentages of completely evaluated reports was noted. The decrease that was observed from 2020 was probably due to the COVID-19 pandemic which resulted in lower patient admissions. The increase of number of reports might be due to the fact that over time there was more knowledge about the value of [¹⁸F]FDG-PET/LDCT in the diagnosis of infection. In 2011, the first report that met all criteria was observed. In this year, the Department of Nuclear Medicine at the UMCG implemented a systematic method of reporting according to body compartments. The increased percentage of completely evaluated reports over the years could be caused by developments in imaging modalities and/or due to the introduction of multidisciplinary consultation (including a vascular surgeon, a microbiologist, an infectiologist, a radiologist, and a nuclear medicine physician) of VGEI patients. Another explanation could be an increased knowledge on patterns that can be observed on [¹⁸F]FDG-PET/LDCT. Despite the fact that there is a slight upward trend in reporting VGEI, < 50% of the reports contained all criteria, while the report is often the only way of communication between the nuclear medicine physician and the clinician (in this case, the vascular surgeon) [26]. This indicates that there is a strong need to improve the reporting, including a more systematic reporting approach with standardised interpretation criteria. As described earlier by the European Association of Nuclear Medicine and the European Association of Cardiovascular imaging, it is crucial that the referring clinician understands the report as intended by the nuclear medicine physician, as this approach is already more common in conventional nuclear cardiology [27], and in PET/CT applications in oncology [8‐10, 28]. Reporting of several [¹⁸F]FDG-PET/CT applications in cardiovascular diseases are less well addressed, as in the current situation with VGEI, but also in other infections and inflammatory diseases, such as (infective) endocarditis, infection of cardiac implantable electronic devices, large vessel vasculitis, and polymyalgia rheumatica [29, 30]. Although interpretation for several inflammatory-, infective-, infiltrative-, and device-related diseases is described, specific and user-friendly recommendations on reporting are often incomplete [31]. As highlighted now for VGEI, recommendations on reporting with standardised interpretation criteria should be compiled for these missing parts of [¹⁸F]FDG-PET/CT applications in cardiovascular diseases. An example of user-friendly, standardised reporting standards for nuclear imaging on cardiac amyloidosis are published by Dorbala et al. [32, 33]. It is recommended to write the report clearly and as simple as possible, with a limited number of abbreviations, with quantified data instead of qualitative (e.g. small, large, slightly) descriptions (if possible), and with less as possible defensive expressions (e.g. cannot be excluded) [27].

One hundred twenty-three (80.9%) out of 152 patients were diagnosed with VGEI. This high proportion is due to the fact that CTA is still the gold standard imaging modality in suspected VGEI [3]. An [¹⁸F]FDG-PET/LDCT scan was performed subsequently to confirm the diagnosis and/or to evaluate the extent of the infection. As a consequence, there is a selection bias with a high prevalence of VGEI in this suspected VGEI cohort. Almost three-quarter of the patients in our cohort got a graft infection after an open surgical procedure, most likely explained by an overall higher incidence after open surgical repair compared to endovascular repair. For open aortic repair the incidence is up to 4.5% versus 0.3–1.0% for endovascular aortic repair [34]. One of the reasons for this difference is the large, longer lasting surgical wound in open procedures.

Based on the results of the current study and according to available literature (VGEI specific and [¹⁸F]FDG-PET/CT broad), we provide a first recommendation for a concise and complete [¹⁸F]FDG-PET/LDCT report for VGEI (Fig. 2) [7‐9, 12, 31‐33]. Standardisation of nuclear medicine reporting in the wide field of cardiovascular diseases should follow as well and this proposed reporting standard can serve as format for new reporting standards on other cardiovascular diseases, such as infective native aortic aneurysm, where there is a potential role of and value in performing a [¹⁸F]FDG-PET/LDCT, but a lack of studies in the field [35].

In this study, < 50% of the [¹⁸F]FDG-PET/LDCT reports of patients with a suspected VGEI met the predefined criteria for being complete. This led to a lower sensitivity and specificity in comparison with complete reports. Implementing a specific recommendation for VGEI reporting is therefore needed in a next VGEI guideline update [7]. Based on the results of the current study and accompanying literature, we provided a first recommendation for a concise and complete [¹⁸F]FDG-PET/LDCT report for VGEI. Standardisation of [¹⁸F]FDG-PET/LDCT reporting is warranted to improve accuracy and to reduce heterogeneity between different medical centres and to allow comparison between studies.