Ae

MTa

b

a

ARR2A

KP1IALA

1

vihwisailmmt

UC

0h

Computerized Medical Imaging and Graphics 37 (2013) 459– 465

Contents lists available at ScienceDirect

Computerized Medical Imaging and Graphics

jo ur nal ho me pag e: www.elsev ier .com/ locate /compmedimag

ssessment of inflammation in large arteries with 18F-FDG-PET inlderly

arlene Rossibel Montesino Orellanaa, M’hamed Bentourkiab,∗, Otman Sarrhinib,amas Fulopa, Nancy Paquetb, Éric Lavalléeb, Éric Turcotteb, Abdelouahed Khalil a

Centre du vieillissement, Faculty of Medicine and Health Sciences, Université de Sherbrooke, Québec, CanadaCentre d’imagerie moléculaire de Sherbrooke, Faculty of Medicine and Health Sciences, Université de Sherbrooke, Québec, Canada

r t i c l e i n f o

rticle history:eceived 31 August 2012eceived in revised form3 September 2013ccepted 24 September 2013

eywords:ositron emission tomography8F-Fluorodeoxyglucose

nflammationtherosclerosisarge arteries

a b s t r a c t

This paper presents repeated measurements of atherosclerosis using bimodality positron emission tomo-graphy and computed tomography (PET/CT) with 18F-fluorodeoxyglucose (18F-FDG) to assess its uptakein aorta, iliac and femoral arteries in three groups of elderly subjects classified as normals (N), hyper-cholesterolemics (H) and with stable angina (A) in a 12 months follow-up (T0 to T12). The subjects ingroup H were taking rosuvastatin (20 mg/d) for 12 months before the second scan. The calcifications inthe arteries were determined by CT imaging and the artery PET images were analyzed slice by slice. Thestandard uptake values (SUVs) for 18F-FDG uptake were classified in two main groups: calcified and non-calcified arteries and each main group comprises six sub-groups for the three subject groups N, H and A,and for the two measurements 12 months apart. Although the calcifications were present at some por-tions of the arteries in all subjects (23%, 36% and 44% of calcified sites to total sites analyzed, respectively,in groups N, H and A), the results show the most noticeable SUV changes after 12 months was in group

ging N of non-calcified arteries. In the three groups, the calcified arteries showed no significant differencesbetween T0 and T12 while significant differences were observed for the non-calcified arteries. However,there were no significant changes at T12 between groups N and H following rosuvastatin intake in groupH. In conclusion, the quantitative analysis with 18F-FDG-PET/CT could be efficient in the localization ofthe inflammation and evaluation of its progression in atherosclerosis instead of global evaluations withsystemic inflammation biomarkers.

. Introduction

Atherosclerosis is responsible for a large proportion of cardio-ascular diseases (CVDs), which are the leading causes of mortalityn the world. In 2008, out of the 17.3 million cardiovascular deaths,eart attacks were responsible for 7.3 million deaths and strokesere responsible for 6.2 million deaths [1]. Atherosclerosis is an

nflammatory process which affects medium and large blood ves-els. In the arteries, the vascular endothelium is a metabolicallyctive organ system that helps maintain vascular homeostasis andt is comprised of organized cells [2–4]. High level of low-densityipoprotein and their oxidation with free radicals raises the per-

eability of endothelium to lymphocytes and monocytes. Theonocytes engulf these attracted particles and form macrophages

hat result in the creation of atheromatous plaques and

∗ Corresponding author at: Department of Nuclear Medicine and Radiobiology,niversité de Sherbrooke, 3001, 12th Avenue, North Sherbrooke (Qc), J1H 5N4,anada. Tel.: +1 819 820 6868x11863; fax: +1 819 829 3238.

E-mail address: mhamed.bentourkia@usherbrooke.ca (M. Bentourkia).

895-6111/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.compmedimag.2013.09.006

© 2013 Elsevier Ltd. All rights reserved.

arteries stenoses. Stroke and heart attack are the consequences ofthese plaques ruptures. Several factors are known to contribute toatherosclerosis among them hypercholesterolemia, high saturatedfat diet, hyperglycemia, insulin resistance, hypertension, obesity,aging, smoking and air pollution [3,5].

Inflammation is present in all stages of the atherosclerosis pro-cess, from initiation to plaque rupture. Several biomarkers arecurrently used to measure inflammation among which C-reactiveprotein (CRP), CD40 ligand, adiponectin, interleukin 18, and matrixmetalloproteinase 9 [5]. CRP, which is a protein synthesized in theliver, is actually the most utilized biomarker for systemic inflam-mation [5–7]. It was reported that CRP concentration decreasesby 15–50% with statin therapy [2]. Since assays with CRP are per-formed in blood plasma, the detection is not appropriate to locateand quantify the site of artery inflammation.

The quantitative tomographic imaging has the potential tonon-invasively determine the site and the characteristics of

the inflammation. Several techniques, tomographic and non-tomographic, have been proposed for such imaging and mostof these are reviewed in [8,9]. The imaging modalities cur-rently reported in the literature are the X-ray angiography,

4 dical Im

inml[

suctaiaspwiwqasaMbctabatfibctias

fltItwmica

pcassapblpupiveta

60 M.R.M. Orellana et al. / Computerized Me

ntravascular ultrasound (IVUS), computed tomography (CT), mag-etic resonance imaging (MRI), and radionuclide imaging. Theodalities providing good anatomic information mostly identify

uminal diameter or stenosis, wall thickness and plaque volume8,9].

X-ray angiography, as the gold standard method and having highpatial resolution, provides measurements of the stenosis, irreg-larities in luminal surfaces, plaque disruption, thrombosis andalcification, while many lesions not detected by angiography causehe majority of acute coronary syndromes [8,10–13]. IVUS, in which

small ultrasound transceiver mounted on the end of a catheters introduced into a vessel, provides direct imaging of atheromand can provide tomographic images of the entire plaque and ves-el wall, allowing for visualization of calcified and non-calcifiedlaques with many degrees of stenosis [7,14–16]. However, itsidespread use in clinical practice may be partially limited by its

nvasive nature [9,17]. Since non-enhanced CT can image materialsith different densities, this modality is appropriate to locate and

uantify calcifications in arteries. The grade of the calcification isssessed by Agatston score which is based on the calcification inten-ity in Hounsfield units (HU) and its volume [11,18,19]. Imagingtherosclerosis with MRI uses several techniques. The advantage ofRI is its ability to provide soft tissue and functional information

y exploring proton density, perfusion, diffusion, and biochemicalontrasts in addition to the high resolution anatomical informa-ion. Contrast enhanced MRI has been successfully used to imagengiogenesis, thrombosis, and to characterize the plaque vulnera-ility [4,8]. Although the above described imaging technologies areble to quantify the extent of disease in terms of luminal obstruc-ion and visualization of some plaque elements, anatomic imagingalls of providing information of metabolic processes taking placen plaques [20]. In radionuclide imaging, many radiotracers haveeen developed in recent 20 years based on several molecules andells involved in atherogenesis to image atherosclerotic lesions oro detect the thrombus associated with these lesions [19]. Thesenclude lipoproteins (native and oxidized LDL), immunoglobulinsgainst macrophages, smooth muscle cells, and endothelial adhe-ion molecules [9].

Positron emission tomography (PET) used with 18F-uorodeoxyglucose (18F-FDG) is a molecular imaging modalityhat measures glycolytic activity in normal and diseased tissues.n a study with patients having Takayasu’s arteritis, it was shownhat 18F-FDG was taken up by the inflammatory cells in the artery,hile after successful treatments, the vessel wall showed noeasurable 18F-FDG uptake [21]. Therefore, as a functional imag-

ng modality, 18F-FDG-PET can detect and localize inflammatoryhanges in the arterial wall, which may represent early stages oftherosclerosis.

It was reported an increase of 18F-FDG uptake with age ineripheral arteries [22] and in arterial wall distinctly from the cal-ification [23,24]. Joly et al. reported that aortic inflammation asssessed with 18F-FDG-PET is associated with variation in aortictiffness [25]. Tahara et al. reported that inflammation was mea-ured by 18F-FDG-PET in 30% of patients confirmed with carotidtherosclerosis [26]. Tawakol et al. found in a study of carotidlaque inflammation with 18F-FDG-PET a significant correlationetween 18F-FDG signal and macrophage staining from histo-

ogic sections, and that 18F-FDG uptake did not correlate withlaque volume [27]. Other works confirmed the selective 18F-FDGptake in a variable degree of inflammation in the atheroscleroticlaques [28–31], while Meirelles et al. observed that inflammation

n atheroma is a waxing and waning process as assessed with the

arying 18F-FDG uptake [32]. Concerning the treatment, Taharat al. showed a decrease of 18F-FDG uptake after a three-monthreatment with simvastatin in atherosclerotic patients [33], andbout the imaging time, Menezes at al. reported that there is no

aging and Graphics 37 (2013) 459– 465

advantage to Image 18F-FDG later than one hour after injection[34].

This study was aimed to investigate the use of PET/CT to locateand quantify artery inflammation in normal, hypercholesterolemicand angina subjects in repeated measurements and to determinethe sensitivity of PET measurements to differentiate inflamma-tion between these patients. The three groups were imaged twicewith PET/CT, twelve months apart. The arterial inflammation wasassessed by measuring SUV in aorta, iliac and femoral arteries.

2. Materials and methods

2.1. Study population

All the subjects gave written informed consent as approved bythe ethics committee of the faculty of medicine and health sciences,Université de Sherbrooke. We prospectively recruited a total of10 elderly subjects, 6 males and 4 females with age 65–85 years(in fact we recruited 24 subjects but several dropped their sec-ond scan, and for some the PET data were rejected because therewere uncertainties due to abdominal movements during the scan-ning). The selection of this range of age was decided because therisk of atherosclerosis is higher in this category of population. Welimited the age to 85 to ensure of subjects’ autonomy for move-ment and for experimental consent. Three groups were assembledin this study: healthy or normals (N, n = 4), subjects with hyper-cholesterolemia (H, n = 4), and subjects with stable angina (A, n = 2).Because the progress of the atherosclerosis non-uniformly affectssome portions of the aorta, iliac and femoral arteries, we ana-lyzed 53 artery segments in normals, 45 in hypercholesterolemicsand 34 in subjects with angina. The results and the statistics arereported for these segments. The normals have normal electrocar-diography and normolipidemia and they have no clinical historyof atherosclerosis. Subjects presenting hypercholesterolemia wererecently clinically recognized with this disease at our center andthey were not under medication for hypercholesterolemia beforetheir recruitment in the study. The group with angina was clin-ically under treatment for recent (<6 months) stable angina. Allsubjects were non-smokers, not taking antioxidants, vitamin sup-plements or hormonal replacement for women, and not havingexcessive alcohol consumption. After recruitment, hypercholes-terolemic subjects were treated with rosuvastatin (20 mg/d) during12 months. Subjects with stable angina were taking their own med-ication without changes throughout the study.

2.2. Biochemical analyses

On the morning of each PET measurement, after an overnightfasting, all the subjects had blood drawn from a peripheral vein forthe measurements of glucose, total cholesterol, triglycerides, LDL,high-density lipoprotein (HDL) and CRP. Two blood samples weretaken, one at T0 (at the initial PET/CT scan) and the second one atT12 (12 months later at the second PET/CT scan) to measure plasmaglucose.

2.3. Imaging techniques

The subjects were measured with a PET/CT (Philips Gemini TF)12 months apart. The subjects fasted for at least 6 h prior to imagingto assure a normal glucose level. The CT scan was conducted firstthen the subjects were injected with an 18F-FDG bolus of 140 to

400 MBq depending on the subject weight and the PET scan wasinitiated in dynamic mode for 30 min. A 20 min additional station-ary scan was further taken. The arteries scanned were the aorta,iliac and femoral.

M.R.M. Orellana et al. / Computerized Medical Imaging and Graphics 37 (2013) 459– 465 461

T (lef

1rpid(pmafe

Fs

Fig. 1. Images of calcified aorta (arrow) of the same slices from C

The dynamic scans were divided into 26 image frames of2 × 10 s; 8 × 30 s; 6 × 240 s, and the PET and CT images were co-egistered. The images were analyzed slice by slice to target severalortions of the arteries. Using the CT images as reference, artery

mages in each transaxial slice of the PET images were manuallyelimited by regions of interest (ROI) and their time-activity curvesTAC) were produced. The CT images were also examined for theresence of calcified plaques in the walls of the same arterial seg-ents and the segments were classified in two groups as calcified

nd non-calcified. The SUV was assessed as the mean of the lastour data points in the time activity curve of each region of inter-st on the artery, divided by the injected radiotracer activity and

ig. 2. Images of non-calcified aorta. Top-left: image from 30 min. Bottom-left: image fromhowing no or very slight radiotracer activity in the ROI. Bottom-right: time-activity curv

t) and PET (right) in coronal (top) and transaxial (bottom) slices.

normalized to the subjects’ body surface area. The body surface areawas assessed as BSA =

√weight × height/3600 [35].

2.4. Statistical analyses

All statistical analyses were made with GraphPad Prism 5.Variables were expressed as mean ± SD. ANOVA statistics wereused to determine the significance of the differences betweenthe values of the biochemical parameters and between SUVs. P-

values of less than 5% were considered as statistically significantlydifferent.

early 100 s. Top-right: image from 4 min measured 26 min after 18F-FDG injectione of the aorta ROI as indicated by arrow.

462 M.R.M. Orellana et al. / Computerized Medical Imaging and Graphics 37 (2013) 459– 465

Table 1Baseline characteristics of subject groups. The columns of P-values are to compare T0 and T12 in each group of subjects.

Parameters N (3 women, 1 man) P < 0.05 H (3 men, 1 woman) P < 0.05 A (2 men) P < 0.05

T0 T12 T0 T12 T0 T12

Age (years) 69.5 ± 4.5 70.5 ± 4.5 68.75 ± 4.57 69.75 ± 4.57 69.5 ± 2.12 70.5 ± 2.12Weight (kg) 67.0 ± 4.06 67.8 ± 5.1 ns 78.3 ± 10.28 77.7 ± 11.37 ns 87.6 ± 32.3 86.9 ± 32.6 nsGlucose (mmol/l) 4.18 ± 0.19 4.73 ± 0.92 ns 4.82 ± 0.09 4.98 ± 0.29 ns 4.85 ± 0.49 5.35 ± 0.35 nsTotal cholesterol (mmol/l) 5.38 ± 0.48 5.75 ± 0.46 ns 6.56 ± 0.71 4.38 ± 0.45 P = 0.0010 3.59 ± 0.68 3.29 ± 1.13 nsLDL (mmol/l) 3.43 ± 0.61 3.64 ± 0.47 ns 4.54 ± 0.6 2.39 ± 0.46 P = 0.0007 1.91 ± 0.58 1.59 ± 0.67 nsHDL (mmol/l) 1.34 ± 0.34 1.38 ± 0.39 ns 1.33 ± 0.26 1.31 ± 0.16 ns 1.22 ± 0.16 1.18 ± 0.44 nsTriglycerides (mmol/l) 1.22 ± 0.17 1.61 ± 0.67 ns 1.52 ± 0.83 1.49 ± 0.47 ns 1.02 ± 0.27 1.16 ± 0.02 nsCRP (mg/l) 2.89 ± 0.21 3.05 ± 0.01 ns 3.00 ± 0 3.07 ± 0.15 ns 3.8 ± 1.13 11.9 ± 12.5 nsGlucose, HDL, Triglycerides and CRP differences were not significant (ns) between the three groups. Total cholesterol and LDL differences were significant

3

sbmottcp

mFHiadaia

(cTiastttaffncttbTtwba2ito(T

and lead to misalignment in the merged images that may ham-per the image assessment, adding to the subject movement duringthe dynamic scanning. By associating PET and CT imaging, we wereable to corroborate that 18F-FDG accumulates in abdominal aorta,

Fig. 3. SUVs for artery segments grouped from normal subjects (N0 and N12), fromsubjects with high levels of cholesterol (H0 and H12), and from subjects with sta-ble angina (A0 and A12). The indices 0 and 12 represent the measurements at theinitiation of the study and 12 months later, respectively. The first 6 values rep-resent calcified arteries and the last 6 values represent non-calcified arteries. Thedata were obtained from transaxial image slices. The circles in each group indicate

between the three groups.

. Results

Subjects were clinically stable and asymptomatic when bloodamples and PET/CT measurements were performed. Subjects’aseline characteristics are reported in Table 1. During the 12-onth follow-up period, there were no significant changes in

verall clinical characteristics except for total and LDL choles-erol levels for the hypercholesterolemic group. We observedhat rosuvastatin significantly decreased the levels in the hyper-holesterolemic group for total and LDL cholesterol (p = 0.001 and

= 0.0007, respectively).18F-FDG uptake was observed to be present in all artery seg-

ents of all subjects even those with calcifications as shown inig. 1. These images (Fig. 1) were taken from a subject in-group

in coronal and transaxial views of CT and corresponding PETmages summed over the entire measurement time. Fig. 2 presents

transaxial image slice from a normal subject of non-calcified aortauring whole scan (top-left image), during early 100 s (bottom-left)nd during 4 min at 26 min after 18F-FDG injection (top-right). Themage at later scanning time shows no or very slight radiotracerctivity in the aorta, indicating that possibly this segment is normal.

The analysis with SUV was assessed by grouping the aorta, iliacleft and right) and femoral (left and right) data as sub-groups foralcified and non-calcified and also for T0 and T12 for follow-up.he mean values and standard deviations of the SUVs are reportedn Fig. 3 and in Table 2. In comparing T0 and T12, the non-calcifiedrteries were found significantly different. Because SUV depends onubject’s weight and height and on the injected 18F-FDG activity,he uncertainties on these parameters contribute to the variation inhe SUV values. In order to reduce these variations, we normalizedhe artery uptake of 18F-FDG (last 4 frames) to the total artery TACnd the results are displayed in Fig. 4. Again, the only significant dif-erence was found between T0 and T12 for non-calcified arteries asor SUV. The comparisons between the groups within calcified andon-calcified arteries and for T0 and T12 are reported in Table 3. Foralcified arteries, there were no significant differences between thehree groups either at T0 or at T12. This pattern is not preserved forhe non-calcified arteries. In fact, both values of N and H increasedetween T0 and T12 while A values only slightly increased (seeable 2). This discrepancy might be explained by the low effect ofhe rosuvastatin intake by group H or because their values at T0ere already high, while the medication of group A apparently sta-

ilized the inflammation. The ratio of calcified to total analyzedrtery segments didn’t change between T0 and T12. They were3%, 36% and 44%, respectively, in groups N, H and A. By comput-

ng the mean intensities in Hounsfield units in the CT images of

he calcified arteries at T0 and T12, we observed no demarcationf the means at T12 (224.13 ± 122.01 HU) relative to those at T0230.27 ± 111.59 HU). Technically, it was not possible to compare0 and T12 for the same artery segment in the same patient due to

inherent misplacement of subjects in the same position for the twoscans.

4. Discussion

We evaluated the presence of vascular 18F-FDG uptake in10 elderly subjects: normals, hypercholesterolemics and otherswith stable angina using PET/CT as a marker of vascular inflam-mation. This imaging technique provides safe, fairly widespread,non-invasive combined morphological and functional imagingmodalities by which the entire subject may be examined in less than1 h. The correlation of 18F-FDG uptake and atherosclerotic inflam-mation has been previously demonstrated in animal and humanstudies. Studies in humans show that 18F-FDG uptake correlatesstrongly with plaque inflammation [23]. There are, however, sometechnical aspects to be considered when using 18F-FDG-PET/CTto detect atherosclerotic high-risk plaques. Subject’s movementresulting in a shift of the body position between the PET and CTregistration may introduce correlation errors in the PET images

the mean SUVs. The number above each column indicates the number of artery seg-ments included in the analysis. The outliers might be due to really higher FDG uptakeand hence higher inflammation, or, statistically speaking, the data containing out-liers might be made of two normal distributions with the same mean and differentvariances. The outliers represent a few data among the total values included.

M.R.M. Orellana et al. / Computerized Medical Imaging and Graphics 37 (2013) 459– 465 463

Table 2SUVs in non-calcified arteries were found significantly different between T0 and T12 in the three groups (ns = non significantly different; s = significantly different).

Groups SUV % Increase betweenT0 and T12

P-value at 0.05

T0 T12

Calcified N 1.80 ± 0.34 2.06 ± 0.79 14.44% ns, P = 0.285Calcified H 1.87 ± 0.26 2.07 ± 0.36 10.70% ns, P = 0.0782Calcfied A 1.83 ± 0.25 1.86 ± 0.24 1.64% ns, P = 0.691Non-calcified N 1.10 ± 0.78 2.21 ± 0.42 100.91% s, P = 1.23 10−11

Non-calcified H 1.61 ± 0.27 2.16 ± 0.52 34.16% s, P = 5.31 10−6

Non-calcified A 1.50 ± 0.23 1.65 ± 0.18 10.00% s, P = 0.0216

Table 3Statistical comparison of SUVs between the three groups for calcified and non-calcified arteries at T0 and at [T12] (ns = non significantly different; s = significantly different;calcif = calcified).

Calcif-N Calcif-H Calcif-A Non-calcif-N Non-calcif-H Non-calcif-A

Calcif-N ns [ns] ns [ns]Calcif-H ns [ns]Calcif-ANon-calcif-N s [ns] s [s]Non-calcif-H ns [s]

intg

ctitnwTSosctiot

F(i

Non-calcif-A

liac and femoral atherosclerotic plaques with higher uptake in theormal group with non-calcified arteries. Bural et al. showed thathe prevalence and intensity of 18F-FDG uptake in large arteriesenerally increases with age [22].

Even though rosuvastatin induced a significant reduction in totalholesterol and LDL levels in hypercholesterolemic group (Table 1),here was no significant difference between the SUVs for T0 and T12n calcified arteries while the values were significantly different inhe non-calcified arteries (Table 2). Global SUVs in calcified andon-calcified arteries increased from T0 to T12 by 48% in-group Hith respect to group N (mean SUV of H divided by mean SUV of N).

his reduced increase in-group H could be attributable to the higherUVs at T0 in comparison to group N and not directly to the effectf rosuvastatin. We can deduce that arterial inflammation was pre-erved unchanged throughout the 12 months of follow-up. On theontrary, the values in-group A remained low in comparison to

he other groups as these patients were under medication. Accord-ng to the Stellar study, which compared the efficacy and safetyf rosuvastatin with other well-known and used statins, rosuvas-atin is more efficient in reducing plasma LDL cholesterol [2,33]. In

ig. 4. Same as in Fig. 3 with the data calculated as the ratio of the last four frames16 min) normalized to the whole TAC amplitude. The data appear with less variationn comparison to SUV in Fig. 3, while they present the same pattern between groups.

addition, in that same study and others, rosuvastatin also producedsignificantly better reductions in total cholesterol compared withother statins [33,36,37]. Statins exert their benefits through theinhibition of the novo cholesterol synthesis, resulting in significantreductions in plasma low-density lipoprotein cholesterol (LDL-C).It remains controversial whether LDL-C lowering is the only mech-anism for the observed beneficial effects. Many LDL-C independentpleiotropic effects have been postulated [38]. One of them is attenu-ation of inflammation because statins have been shown to decreasesystemic inflammatory markers [39]. Tahara et al. suggested theanti-inflammatory effect of simvastatin on atherosclerosis plaquesmay be one of the pleiotropic effects independent of LDL-C lowering[33]. They concluded that 18F-FDG-PET visualizes plaque inflam-mation and simvastatin attenuated it. We note that the patientsstudied in the work of Tahara et al. have different health problemsand were scanned for cancer diagnoses. By comparison to the nor-mals at T12, our results do not demonstrate a decrease of 18F-FDGuptake in hypercholesterolemic subjects only treated with rosu-vastin (20 mg/d) during 12 months. More emphases should be madeon the number of subjects, the scanning duration and the analy-ses by other means than SUV in order to confirm these results. AsChen et al. demonstrated, age and hypercholesterolemia are cor-related consistently with 18F-FDG uptake in the abdominal aorta,iliac and femoral arteries [40]. The positive correlation betweenarterial 18F-FDG uptake and cardiovascular risk factors such as ageand hypercholesterolemia suggests a promising role for 18F-FDG-PET imaging in the detection of atherosclerosis and evaluation ofplaque inflammatory activity [40].

The SUV, which is commonly employed for assessing diseaseactivity with PET imaging, can provide quantitative informationabout the severity of the inflammatory process in the arterialwall even before it is clinically symptomatic or visible by struc-tural imaging techniques [40]. In the present work, we useddynamic imaging acquisitions during 30 min. We hypothesize thatthe inflammation accumulates 18F-FDG radiotracer at its first cir-culation and thus after 30 min, the contrast is well established in thearteries. Since the SUV depends only on the amplitude of the signal

in the images, there would be no necessity of imaging for longertime. Menezes et al. reported similar conclusions after studyingscan lengths [34]. The accuracy of the results depends, among otherfactors, on the sample of population used in the experimentation.

4 dical Im

Ihctsoni

amnepfattscilmibmiatt[

adt1hbbCia

5

iaeghp1utdmot

R

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

64 M.R.M. Orellana et al. / Computerized Me

f the number of subjects imaged in this work was not statisticallyigh, the high number of arteries image slices, in the same subject,onsidered as normal, having inflammation and having calcifica-ion were considered as individual samples. The state of the arteriesegments at T0 and at T12 was the major concern for the progressf inflammation and for the impact of the medication. The lowestumber of samples (13) was in the calcified arteries of normals, and

n this group we have 10 subjects (Fig. 3).Quantification of 18F-FDG uptake in arteries is more efficient

s the absence of visible calcifications on CT images does notean absence or presence of inflammation. Assuming there are

o drastic differences of partial volume effect (PVE) on the differ-nt arteries and neglecting the noise in the artery images withartial calcification, the SUVs in the abdominal aorta, iliac andemoral arteries were interpreted as inflammation for higher SUVsnd as healthy or dominated by calcification for lower SUVs. Inhe latter case, CT images can help to decide. Among other quan-ification methods, target to background ratio (TBR) was used byeveral authors and itself is based on SUVs [27,28]. Because it isalculated as the ratio of plaque SUV to venous blood SUV, TBRs more prone to uncertainties adding to uncertainties due to theow signal in the venous blood. We expect that the most opti-

al quantitative approach is the calculation of glucose metabolismn the arteries following a dynamic PET measurement. Therefore,lood samples should be withdrawn from the patients during theeasurements or mathematical tools such as factor analysis or

ndependent component analysis should be used in order to sep-rate blood component from tissue in the images and to applyhe compartmental modeling. Some such approaches include evenhe correction for partial volume effect in the kinetic modeling42].

We have shown in this work the potential of 18F-FDG-PET/CT tossess the progress of the inflammation during 12 months and toetermine the effects of the medication on the arterial inflamma-ion (stabilization in angina group). Imaging atherosclerosis with8F-FDG has been demonstrated to be feasible in many reports;owever, this technique has to undergo several validations beforeeing used in the clinic or for drug development. The relationetween arterial inflammation and several risk factors such asVDs, hypertension, oxidized lipoproteins, diabetes, obesity and

nfection has not yet been clearly defined [41]. Also the availabilitynd the high cost of the imaging techniques have to be considered.

. Conclusion

18F-FDG-PET/CT imaging helps visualizing and assessingnflammation in calcified and non-calcified aorta, iliac and femoralrteries. The quantitative analysis of 18F-FDG-PET/CT could befficient in the localization of the inflammation and its pro-ression to help guide appropriate treatments. Rosuvastatin mayave an anti-inflammatory effect on atherosclerotic plaques inde-endent of LDL-C lowering effects. However, given 20 mg/d for2 months, rosuvastatin was not observed to lower 18F-FDGptake in hypercholesterolemic subjects while their total choles-erol and LDL were reduced. In the group of elderly normals, theata showed a significant increase of 18F-FDG uptake after 12onths, suggesting the effect of inflammation. In this group, 23%

f the analyzed artery segments were found as having calcifica-ions.

eferences

[1] Mendis S, et al. World Health Organization definition of myocardial infarction:2008–09 revision. Int J Epidemiol 2011;40(1):139–46.

[2] Libby P. Inflammation and cardiovascular disease mechanisms. Am J Clin Nutr2006;83(2):456S–60S.

[

[

aging and Graphics 37 (2013) 459– 465

[3] Vanhoutte PM, et al. Endothelial dysfunction and vascular disease. Acta Physiol2009;196(2):193–222.

[4] Nighoghossian NL, Derex P, Douek. The vulnerable carotid artery plaque: cur-rent imaging methods and new perspectives. Stroke 2005;36(12):2764–72.

[5] Packard RR, Libby P. Inflammation in atherosclerosis: from vascular biology tobiomarker discovery and risk prediction. Clin Chem 2008;54(1):24–38.

[6] Farouque HM, Meredith IT. The assessment of endothelial function in humans.Coron Artery Dis 2001;12(6):445–54.

[7] Revkin JH, et al. Biomarkers in the prevention and treatment of atherosclerosis:need, validation, and future. Pharmacol Rev 2007;59(1):40–53.

[8] Naghavi M, et al. New developments in the detection of vulnerable plaque. CurrAtheroscler Rep 2001;3(2):125–35.

[9] Fayad ZA, Fuster V. Clinical imaging of the high-risk or vulnerable atheroscle-rotic plaque. Circ Res 2001;89(4):305–16.

10] Voros S, et al. Coronary atherosclerosis imaging by coronary CT angiography:current status, correlation with intravascular interrogation and meta-analysis.JACC Cardiovasc Imaging 2011;4(5):537–48.

11] Achenbach S, Raggi P. Imaging of coronary atherosclerosis by computed tomo-graphy. Eur Heart J 2010;31(12):1442–8.

12] de Feyter PJ, et al. Quantitative coronary angiography to measure progressionand regression of coronary atherosclerosis. Value, limitations, and implicationsfor clinical trials. Circulation 1991;84(1):412–23.

13] Topol EJ, Nissen SE. Our preoccupation with coronary luminology. The disso-ciation between clinical and angiographic findings in ischemic heart disease.Circulation 1995;92(8):2333–42.

14] Tardif JC. Atherosclerosis imaging. Can J Cardiol 2005;21(12):1035–9.15] Guedes A, Tardif JC. Intravascular ultrasound assessment of atherosclerosis.

Curr Atheroscler Rep 2004;6(3):219–24.16] Motoyama S, et al. Atherosclerotic plaque characterization by 0.5-mm-slice

multislice computed tomographic imaging. Circ J 2007;71(3):363–6.17] Bose D, von Birgelen C, Erbel R. Intravascular ultrasound for the eval-

uation of therapies targeting coronary atherosclerosis. J Am Coll Cardiol2007;49(9):925–32.

18] Agatston AS, et al. Quantification of coronary artery calcium using ultrafastcomputed tomography. J Am Coll Cardiol 1990;15(4):827–32.

19] Vallabhajosula S, Fuster V. Atherosclerosis: imaging techniques and the evolv-ing role of nuclear medicine. J Nucl Med 1997;38(11):1788–96.

20] Sheikine Y, Akram K. FDG-PET imaging of atherosclerosis: do we know whatwe see? Atherosclerosis 2010;211(2):371–80.

21] Andrews J, et al. Non-invasive imaging in the diagnosis and management ofTakayasu’s arteritis. Ann Rheum Dis 2004;63(8):995–1000.

22] Bural GG, et al. FDG-PET is an effective imaging modality to detect and quan-tify age-related atherosclerosis in large arteries. Eur J Nucl Med Mol Imaging2008;35(3):562–9.

23] Dunphy MP, et al. Association of vascular 18F-FDG uptake with vascular calci-fication. J Nucl Med 2005;46(8):1278–84.

24] Tatsumi M, et al. Fluorodeoxyglucose uptake in the aortic wall atPET/CT: possible finding for active atherosclerosis. Radiology 2003;229(3):831–7.

25] Joly L, et al. Aortic inflammation, as assessed by hybrid FDG-PET/CT imaging, isassociated with enhanced aortic stiffness in addition to concurrent calcification.Eur J Nucl Med Mol Imaging 2009;36(6):979–85.

26] Tahara N, et al. The prevalence of inflammation in carotid atherosclerosis:analysis with fluorodeoxyglucose-positron emission tomography. Eur HeartJ 2007;28(18):2243–8.

27] Tawakol A, et al. In vivo 18F-fluorodeoxyglucose positron emission tomographyimaging provides a noninvasive measure of carotid plaque inflammation inpatients. J Am Coll Cardiol 2006;48(9):1818–24.

28] Rudd JH, et al. Fluorodeoxyglucose positron emission tomography imagingof atherosclerotic plaque inflammation is highly reproducible: implica-tions for atherosclerosis therapy trials. J Am Coll Cardiol 2007;50(9):892–6.

29] Arauz A, et al. Carotid plaque inflammation detected by 18F-fluorodeoxyglucose-positron emission tomography. Pilot study. Clin NeurolNeurosurg 2007;109(5):409–12.

30] Davies JR, et al. Identification of culprit lesions after transient ischemic attackby combined 18F fluorodeoxyglucose positron-emission tomography and high-resolution magnetic resonance imaging. Stroke 2005;36(12):2642–7.

31] Wykrzykowska J, et al. Imaging of inflamed and vulnerable plaque in coro-nary arteries with 18F-FDG PET/CT in patients with suppression of myocardialuptake using a low-carbohydrate, high-fat preparation, Journal of nuclearmedicine: official publication. Soc Nucl Med 2009;50(4):563–8.

32] Meirelles GSM, Gonen, Strauss HW. 18F-FDG uptake and calcifications inthe thoracic aorta on positron emission tomography/computed tomogra-phy examinations: frequency and stability on serial scans. J Thorac Imaging2011;26(1):54–62.

33] Tahara N, et al. Simvastatin attenuates plaque inflammation: evaluationby fluorodeoxyglucose positron emission tomography. J Am Coll Cardiol2006;48(9):1825–31.

34] Menezes LJ, et al. Vascular inflammation imaging with 18F-FDG PET/CT: whento image? J Nucl Med 2009;50(6):854–7.

35] Verbraecken J, et al. Body surface area in normal-weight, overweight, and obeseadults. A comparison study. Metabolism 2006;55(4):515–24.

36] Jones PH, et al. Comparison of the efficacy and safety of rosuvastatin versusatorvastatin, simvastatin, and pravastatin across doses (STELLAR* Trial). Am JCardiol 2003;92(2):152–60.

dical Im

[

[

[

[

M.R.M. Orellana et al. / Computerized Me

37] Nicholls SJ, et al. Effect of two intensive statin regimens on progression ofcoronary disease. New Eng J Med 2011;365(22):2078–87.

38] Liao JK, Laufs U. Pleiotropic effects of statins. Annu Rev Pharmacool Toxicol2005;45:89–118.

39] Ridker PM, et al. Measurement of C-reactive protein for the targeting of statintherapy in the primary prevention of acute coronary events. New Eng J Med2001;344(26):1959–65.

[

[

aging and Graphics 37 (2013) 459– 465 465

40] Chen W, et al. Emerging role of FDG-PET/CT in assessing atherosclerosis in largearteries. Eur J Nucl Med Mol Imaging 2009;36(1):144–51.

41] Libby P, Ridker PM, Hansson GK. Inflammation in atherosclerosis.From pathophysiology to practice. J Am Coll Cardiol 2009;54(23):2129–38.

42] Bentourkia M. Kinetic modeling of PET data without blood sampling. IEEE TransNucl Scie 2005;52(3):697–702.