The main finding in this study was that [18F]FDG PET/MRI and [18F]FDG PET/CT are equally well suited as a diagnostic tool in the imaging work-up of LVV while [18F]FDG PET/MRI achieved significantly lower radiation doses.

Especially in diseases that require repetitive imaging, avoiding unnecessary radiation doses has become more important in medical imaging with the aim to reduce the cumulative effective dose. [15, 16].

For the work-up of LVV and FUO in our institution, the effective radiation dose per examination was 6.56 mSv, on average in the PET/MRI group, while it was 31.69 mSv, on average in the PET/CT group. This difference is mainly based on the lack of radiation exposure of the often multiphasic CT component; however, it should be noted that there are differences between the two modalities regarding the quantitative PET component. The MR images are not correlated with tissue attenuation coefficients as CT images are. The reliable determination of the tracer uptake in a PET/MRI requires an attenuation correction of the PET data, which calculates an attenuation based on soft tissue, bone, and hardware components.

Our analysis showed that about 78% of the radiation dose could be reduced while providing the same diagnostic quality in the evaluation of LVV. This is especially of importance when multiple examinations might become necessary in the follow-up of inflammatory LVV [17, 18]. A lower applied radiation dose in every single examination to minimize stochastic radiation effects from cumulative radiation dose might reduce the carcinogenic risk of imaging in this cohort of patients. In this retrospective study, 14 [18F]FDG PET/MRI and 8 [18F]FDG PET/CT examinations were performed on patients younger than 51 years of age. The increase in the risk of developing fatal cancer in this cohort is estimated at 5% per administered Sv for the general population. While this risk increases to 15% per Sv in children, it decreases to 1% per Sv in people in their seventh decade of life [19]. Especially for young patients, PET/MRI might be preferable because it is a lower-dose variant with at least the same diagnostic confidence as PET/CT.

Due to the often unclear clinical presentation, it is assumed that other radiological diagnostics have already taken place before the specific examination using PET/MRI or PET/CT. Because the CT is a readily available imaging technique that is firmly established in the diagnosis of a variety of diseases, there has been a dramatic increase in average patient radiation exposure since its invention. Therefore, it is even more important to be able to save radiation doses in the future [15, 20, 21].

The PET/MRI combines the strength of these two single modalities. The hybrid imaging method enables the acquisition of information about the severity of the inflammation (with PET) and about the anatomical changes via MRI [22]. It is a relatively novel imaging technique whose areas of application are currently expanding [23]. In imaging of LVV, not only is the extent of the inflammation of importance but also anatomical changes are decisive for diagnosis and treatment surveillance. Thus, not every imaging method is suitable to reflect both aspects of this disease. Previous studies have already suggested that hybrid imaging methods [18F]FDG PET/CT and [18F]FDG PET/MRI provide the most reliable information on both aspects. The [18F]FDG PET/CT has already been well studied in this application [24, 25]. A study by Clemente et al. examined childhood-onset Takayasu’s arteritis patients using [18F]FDG PET/MRI and compared cohorts with positive PET and positive MRA with positive PET but negative MRA. It was shown that the hybrid method added information about the extent of the inflammation with the PET component, which is not presented in every patient in an MRA examination alone [26].

Cerne et al. showed that PET/MRI delivers more information about inflammatory markers in the diagnosis and follow-up of LVV compared to laboratory findings because this imaging method characterizes the severity and the extent of the inflammation. Inflammatory markers proved to be too unspecific and, compared to PET/MRI do not allow a statement about anatomical changes and the extent of the inflammation [27].

Laurent et al. evaluated a correlation between the clinical and laboratory presentation of patients with Takayasu arteritis and giant cell arteritis, and the results of imaging using [18F]FDG PET/MRI. The image analysis was graded based on the tracer uptake in the vessels, which was compared to the tracer uptake in the liver, using grading from 0 to 3, and by measuring the thickening and enhancement of the aortic wall on delayed enhancement MR images. This study showed that the tracer uptake in the vessels correlated with the quantity of the inflammation, proven by the clinical and laboratory results. In conclusion, [18F]FDG PET/MRI might even be suitable to replace invasive methods such as biopsies of the temporal artery, or as a diagnostic tool for giant cell arteritis [28], and appears to be a suitable method for detection and follow-up examinations in LVV; however, that study did not contain any comparison to other imaging methods.

In another study, Einspieler et al. drew comparisons regarding visual and quantitative parameters between both hybrid-imaging methods, [18F]FDG PET/MRI and [18F]FDG PET/CT, in patients with known LVV. The image scoring revealed the intensity of the arterial uptake of the tracer in contrast to the uptake into the liver. The statistical analysis was made with the measurements SUVmax and the target to background ratio for each measured vascular segment. The calculation of the target background ratio SUVmax was divided by SUVmin using a region of interest in the blood pool. This study proved that both modalities are equivalent with respect to the visual scores compared to the measured inflammatory parameters [29].

In contrast to the studies listed above, initial diagnoses of LVV were included in this study. The data presented in our paper is in line with these publications showing that imaging-based diagnosis using [18F]FDG PET/MR appears to be possible at a low radiation dose in patients suspected of having LVV; however, previous studies did not compare [18F]FDG PET/CT vs. [18F]FDG PET/MR in terms of effective radiation dose of these two modalities. The present study was the first to show that [18F]FDG PET/MRI offers equivalent diagnostic image quality in the work-up of suspected inflammatory vascular disease when compared to [18F]FDG PET/CT with a relatively high sample size.

Current studies are researching new tracers that are intended to further improve the diagnosis and follow-up of LVV. As glucose acts in metabolically active processes, which occur physiologically to an increased extent in the brain and myocardium, the diagnosis of, for example, temporal arteritis or the assessment of the coronary arteries in LVV can be misleading. In addition, with the use of [18F]FDG it is not possible to distinguish between processes of active inflammation, chronic vascular remodeling, and residual disease activity after treatment. According to a study by Ćorović et al., the tracer [68Ga]-dotatate, which binds somatostatin receptor 2 (SSTR2), already used in neuroendocrine tumor diagnostics, seems promising. The SSTR2 is expressed by macrophages during inflammation and localizes around CD68+ macrophages, such as those found in inflamed carotid atherosclerotic plaques [30]. In addition, other promising tracers are currently being investigated with a view to expanding their areas of application. Some examples include [68Ga]-pentixafor, a PET tracer that was originally developed for cancer imaging by binding to CXCR4, which is expressed on several proinflammatory immune cell types, or 68Ga-FAPI, which binds to the transmembrane serine protease fibroblast activation protein‑α (FAP), whose expression is high in activated stromal fibroblasts at sites of tissue remodeling and is found primarily in pathological conditions such as fibrosis, scar/granulation tissue, cancer, and arthritis [31].

However, current limitations of PET/MRI include its limited availability worldwide and its often predominant use in the work-up and staging of oncologic disease. Furthermore, access to this modality is sometimes hindered by contraindications for MRI such as claustrophobia and non-MR-compatible implants or non-MR pacemaker systems; however, currently, even pacemakers are often deemed MR-compatible, which might improve access to PET/MRI in the future [32]. There are relative contraindications for contrast medium such as terminal kidney failure.

With respect to the investment costs and the ongoing operational expenses, PET/MRI is far more expensive than PET/CT. The costs are related not only to the price of a PET/MR hybrid device itself, but to the availability of the radiological staff that can operate this new device. There is the longer duration required for an MRI compared to a CT scan. For cancer patients, the costs of the routine staging or follow-up per examination is about 50% higher for PET/MRI than for PET/CT [33].

Although the cumulative radiation dose is a strong argument for establishing PET/MRI in the diagnosis of LVV, there are applications where PET/CT is more effective. PET/CT with CT angiography is advantageous for planning interventions in cases of complications like stenosis or aneurysm formation as a result of inflammation.

In terms of the radiation exposure derived from PET/CT, it should be noted that currently, a new generation of modern PET/CT scanners are entering the market. These scanners can provide whole-body scans with ultra-low dose (e.g., 0.37MBq/kg) protocols that show no significant difference in quantitative and qualitative image analysis [34]. These long axial field of view (LAFOV) PET/CT scanners show an increased signal-to-noise ratio and can achieve higher sensitivity based on a higher count density than the scanner used in our analysis. In addition, modern postprocessing methods, such as ultrahigh sensitivity (UHS) reconstruction modes on these new scanners, may cut the PET acquisition time in half, which would reduce the examination time [35]; however, to generate whole-body parametric PET images, a minimum of 45–60 min of scanning is required due to the slow kinetics of 18F-FDG.

The main limitation of this study is the small patient population and its retrospective nature in a single center. Different CT protocols were included in the PET/CT cohort, resulting in an elevated CT radiation dose when compared to CT scans without contrast alone. It should be noted that the PET/CT scanner used in this study is already outdated when compared with recent scanner technology. Additionally, a non-significantly higher 18F-FDG dose had to be applied in PET/CT when compared to PET/MRI as the PET component of the currently available PET/MRI system has a higher sensitivity.

In the risk-benefit balance, the radiation dose savings outweigh the limitations of PET/MRI. The disadvantages of PET/MRI, as with MRI, are its contraindications for use in individuals with claustrophobia, metallic implants, and non-MR-compatible pacemakers, Furthermore, the low availability due to the high costs currently limits the further spread of PET/MR [33].

Although comparable cohorts were considered in this study, both examinations were not performed on a single patient, which precludes exact comparability.