The study protocol was reviewed and approved by the local ethics committee, and written informed consent was obtained prior to inclusion. Ten patients with LEAD and intermittent claudication (Rutherford categories 2 and 3) were recruited in this study. Patients were considered eligible if they had one or more symptomatic de novo atherosclerotic lesions with at least 50% stenosis in the superficial femoral artery, as assessed by duplex ultrasound, and if they had been referred for peripheral artery atherectomy. All patients underwent clinical examination and physiological testing at the Department of Angiology of our institution, which included assessment of the ankle–brachial index (ABI) and color flow duplex sonography (Siemens Acuson S2000, Erlangen, Germany) of the lower limbs.

Duplex criteria subjected to predict hemodynamically relevant stenosis included a maximum peak systolic velocity >200 cm/s, loss of triphasic waveform configuration at the stenosis, damping and reduction in pulsatility in the distal arterial velocity waveform, a peak-systolic (PSV) and end-diastolic (EDV) velocity ratio >2.5 across the stenosis, and end-diastolic velocity >60 cm/s. The lesions included were regarded as appropriate for treatment by atherectomy if they were classified as TASC A or B lesions according to the TASC II classification by two independent experienced physicians (15).

Patients were excluded if they had critical limb ischemia, total occlusion of the target superficial femoral artery, severe renal insufficiency, type 1 diabetes mellitus, or contraindications to PET and/or MR imaging. All patients underwent PET/MR imaging within 24 h prior to atherectomy of the symptomatic lesions.

Patients fasted for at least 12 h before the intravenous injection of 18F-FDG and simultaneous PET/MR imaging in a whole-body scanner (Biograph mMR, Siemens Erlangen, Germany). An attachable marker was placed on the skin to indicate the area of the stenotic vascular region of interest, as assessed by color flow duplex sonography to indicate the plaque location for PET/MR imaging and atherectomy. For PET imaging, a 20-min scan was acquired at a single bed position (longitudinal FOV = 25 cm), centered at the position of the attached marker. Before each PET acquisition, an 18-s MR Dixon sequence was acquired to generate a fat-water separated map for attenuation correction (AC) of PET data (16). PET imaging parameters were as follows: 3D mode, 172 × 172 matrix, 4.17 × 4.17 mm pixel spacing, 2 mm slice thickness, and 127 slices. Images were reconstructed using the ordered-subsets expectation maximization (OP-OSEM) algorithm with a 4-mm Gaussian filter. Simultaneously, MR imaging was performed using a combination of body matrix and spine array coils to cover both upper legs. The following MRI sequences were acquired: time-of-flight (TOF), T1-w turbo spin echo (TSE), and T2-w TSE. Detailed imaging parameters for MRI sequences are listed in Supplementary Table S1. The T1-w sequence was repeated 5–7 min after injecting the Gd-based contrast agent (Magnevist, Bayer Schering Pharma AG, Berlin, Germany) at a dosage of 0.1 mmol/kg body weight). The total imaging time averaged 45–60 min.

Image analysis was performed using MATLAB (MathWorks, Natick, MA, USA). On postcontrast T1-w axial MRI images, the plaque size (maximum diameter) was measured. PET and MRI images were fused, and the location of the plaque was determined by selecting the postcontrast T1-w MRI image slice with the largest plaque (“target” slice). For MRI image analysis, a ring-shaped region of interest (ROI) was drawn around the arterial vessel wall, excluding the luminal signal, i.e., blood. A circular ROI in the adjacent musculus quadriceps femoris was used for normalization. The MR plaque enhancement ratio (ER) was calculated as previously described (17).

For PET analysis, FDG uptake was assessed on the axial image identified by MR (“target”). The ring-shaped ROI delineated on MRI images was copied to the fused PET image. Due to the low resolution of PET, the enclosed circular ROI was used (including lumen) for the artery signal. For further analysis, the ROI maximum standard uptake value (SUV_max) was used as previously described (18). The arterial PET signal was normalized to the average standard uptake value (SUV_average) in the adjacent muscle ROI. This yielded the target-to-background ratio (TBR):TBR=(SUVmax)artery(SUVmax)muscleThe analysis was also performed for the contralateral (asymptomatic) superficial femoral artery. The overall intensity of FDG uptake in femoral plaques was measured by max. SUV and the tissue-to-background ratio (ratio of max. SUV of plaque/mean SUV of blood) on PET and compared among calcified and non-calcified plaques.

All patients underwent angiography of the target lesion via an antegrade percutaneous femoral artery access using a 7-F sheeth within 24 h after PET/MR imaging. The selected symptomatic lesions of the superficial femoral artery were treated with a SilverHawk plaque excision system (Covidien ev3 Endovascular, Inc., Plymouth, MN, USA) by an experienced interventionist. The number of lesion passes and the use of an additional low-pressure balloon were left to the discretion of the interventionist.

Immediately after excision, specimens obtained by atherectomy were fixed using neutral-buffered formalin. After fixation, specimens were processed for histopathology. Decalcification was carried out after the specimens have been thoroughly fixed and prior to embedding in paraffin using ethylenediaminetetracetic acid (EDTA) at a concentration of approximately 14% as a neutralized solution. All portions were sectioned at 3–5 µm thickness. Sections were stained using hematoxylin and eosin (H&E) and Elastica van Gieson (EVG) staining methods. Immunohistochemistry included staining with antibodies targeting monocytes (CD-68) and alpha smooth muscle actin (SMA). Computerized planimetry of the stained area was performed using Image Pro Plus 6.1 (Media Cybernetics, Inc., USA) on all sections, as previously described (21). The histology slides were reviewed independently by a reviewer (J.N.) who was blinded to the imaging results and were categorized according to the AHA classification based on the histopathological classification described by Stary et al. (22, 23). A summary of advanced atherosclerotic lesions and the conventional AHA histological classification is provided in Supplementary Table S2. In this classification, the term advanced lesion is used for all lesions that disrupt intimal structure. Type V and VI lesions develop and progress by mechanisms that are superimposed on the continuing lipid accumulation that produced lesion types I–IV. In type IV lesions, the intimal structure is altered almost solely by an extensive accumulation of extracellular lipids localized in the deep intima (the lipid core). In type V lesions, the intima is thickened by substantial reparative fibrous (mainly collagenous) tissue layers (fibroatheroma). Type VI lesions are characterized by surface defects, hematoma, and thrombotic deposits. The predominant (often sheath-like) calcification of a fibro-lipid lesion corresponds to type VII (calcific lesion), while fibrous tissue layers without or with only minimal lipid accumulation (no core) and minimal or no calcium are described as type VIII (fibrotic lesion) (22, 23).

All data were expressed as mean ± standard deviation. For the comparison of MR and PET signal uptake in the left vs. right superficial femoral artery, data were pooled and assessed using the Wilcoxon rank-sum test. Correlations between imaging parameters were assessed using Spearman's correlation coefficient testing. For all analyses, p < 0.05 was considered significant.

Ten patients were enrolled in this pilot study. Two patients were excluded from the data analysis because of insufficient imaging quality. A total of eight patients were included in the data analysis. The study population had a mean age of 70 ± 6 years and consisted of 75% male patients. Patients exhibited typical cardiovascular risk factors, as shown in Table 1. The majority of patients (88%) exhibited atherosclerosis in other vascular segments, including the peripheral and coronary vasculature. Three patients (38%) had a history of prior stroke or carotid artery disease. All patients were receiving antiplatelet and statin therapy. The mean ankle-brachial index (ABI) on the symptomatic side was reduced, averaging 0.8 ± 0.07. The mean maximum peak velocity within the stenosis averaged 313 ± 92 cm/s.

Figure 1 shows the representative images of the multimodal imaging procedures performed on all study patients. Based on T2-w images, the measurements of plaque extent within the stenotic lesion of the symptomatic superficial femoral artery averaged a mean of 0.38 ± 0.15 cm, which was significantly larger compared to the contralateral control site (mean 0.23 ± 0.11 cm, p-value = 0.04), thus emphasizing positive remodeling of the target lesion (Table 2, Figure 2A). Similarly, MR contrast uptake at the level of the symptomatic stenosis assessed on T1-w images was significantly higher compared to the contralateral superficial femoral artery control area (mean 1.77 ± 0.19 vs. 1.57 ± 0.15; p-value = 0.04) (Table 2, Figure 2B). The mean degree of stenosis on MR-TOF angiography in the entire study population averaged 68 ± 11% and significantly correlated with the peak systolic velocity, as assessed by duplex sonography (ρ = 0.91, p-value < 0.01) (Table 2, Figure 2C). Contrast enhancement on T1-w imaging significantly correlated with the peak systolic velocity measurements by duplex sonography (ρ = 0.73, p-value = 0.04) (Table 2, Figure 2D).

Patients received a mean intravenous injection of 358 ± 33 MBq of 18F-FDG. After a mean circulation time of 2 h and 16 min ± 37 min, patients underwent simultaneous PET/MR imaging in a whole-body scanner (Biograph mMR, Siemens Erlangen, Germany) as shown in Figure 3.

Throughout the study population, different patterns of vascular 18F-FDG uptake were observed based on a qualitative assessment. In detail, one patient (ID 2) showed no tracer uptake. Four patients (IDs 3, 4, 6, 8) demonstrated focal FDG uptake of the symptomatic lesion, while in another three patients (IDs 1, 5, 7), tracer uptake was identified in the symptomatic lesion and the ipsilateral course of the superficial femoral artery.

Quantitative analysis showed the FDG-PET signal was enhanced in the symptomatic lesion compared to the blood signal only in the three patients with focal uptake (TBR > 1, Table 2). When comparing signal uptake in calcified vs. non-calcified plaques, we found no significant difference in FDG uptake (p-value = 0.40).

Simultaneous PET/MR imaging was successfully performed in all patients, allowing detailed postprocessing analysis of all modalities. Coregistered axial views of the atherosclerotic lesion on MRI allowed the integration of several plaque features such as positive remodeling, hyperintense signal on T2-w, and MR contrast agent uptake on T1-w images as well as 18F-FDG uptake. Representative images of simultaneous PET and MRI are shown in Figure 1. Assessing the relationship between PET and MR imaging parameters at the level of stenosis, we found no significant correlation between the PET-FDG TBR and the MR T1-w postcontrast agent enhancement (ρ = 0.29, p-value = 0.50).

All plaque specimens obtained by atherectomy consisted of fragments of fibrotic plaques and different extents of fibro-calcified components with interspersed macrophage infiltration. Histologically, we did not observe characteristic constituents of necrotic cores such as cholesterol crystals or intraplaque hemorrhage. Comparing histopathology with FDG-PET imaging showed no correlation but a weak trend toward increased TBR in fibrous plaques with low calcification and decreased TBR in fibro-calcified lesions (ρ = −0.25, p-value = 0.54). Interestingly, the patient with no 18F-FDG uptake (ID 2) had a late-stage fibro-calcified plaque with little to no macrophage deposition. Comparing histopathology with MR contrast enhancement showed no correlation. Staining for CD68 showed increased CD68-positve macrophages in plaques with no to little calcification, whereas it subsequently decreased within more fibro-calcified lesions, yielding a negative correlation (ρ = −0.73, p-value = 0.05). Comparison of CD68-staining positive macrophages with MR enhancement ratio showed a positive trend (ρ = −0.50, p-value = 0.22), whereas rather no correlation with MR plaque size measurements (ρ = 0.38, p-value = 0.36) or PET TBR (ρ = 0.24, p-value = 0.58). SMA within the plaque samples was rather low, not showing any correlation with imaging parameters.