Skip main navigation

18F-Sodium Fluoride Uptake Is a Marker of Active Calcification and Disease Progression in Patients With Aortic Stenosis

Originally publishedhttps://doi.org/10.1161/CIRCIMAGING.113.001508Circulation: Cardiovascular Imaging. 2014;7:371–378

Abstract

Background—

18F-Sodium fluoride (18F-NaF) and 18F-fluorodeoxyglucose (18F-FDG) are promising novel biomarkers of disease activity in aortic stenosis. We compared 18F-NaF and 18F-FDG uptake with histological characterization of the aortic valve and assessed whether they predicted disease progression.

Methods and Results—

Thirty patients with aortic stenosis underwent combined positron emission and computed tomography using 18F-NaF and 18F-FDG radiotracers. In 12 patients undergoing aortic valve replacement surgery (10 for each tracer), radiotracer uptake (mean tissue/background ratio) was compared with CD68 (inflammation), alkaline phosphatase, and osteocalcin (calcification) immunohistochemistry of the excised valve. In 18 patients (6 aortic sclerosis, 5 mild, and 7 moderate), aortic valve computed tomography calcium scoring was performed at baseline and after 1 year. Aortic valve 18F-NaF uptake correlated with both alkaline phosphatase (r=0.65; P=0.04) and osteocalcin (r=0.68; P=0.03) immunohistochemistry. There was no significant correlation between 18F-FDG uptake and CD68 staining (r=−0.43; P=0.22). After 1 year, aortic valve calcification increased from 314 (193–540) to 365 (207–934) AU (P<0.01). Baseline 18F-NaF uptake correlated closely with the change in calcium score (r=0.66; P<0.01), and this improved further (r=0.75; P<0.01) when 18F-NaF uptake overlying computed tomography–defined macrocalcification was excluded. No significant correlation was noted between valvular 18F-FDG uptake and change in calcium score (r=−0.11; P=0.66).

Conclusions—

18F-NaF uptake identifies active tissue calcification and predicts disease progression in patients with calcific aortic stenosis.

Clinical Trial Registration—

URL: http://www.clinicaltrials.gov. Unique identifier: NCT01358513.

Introduction

The mechanisms underlying aortic stenosis (AS) remain incompletely understood, and the accurate prediction of disease progression remains a challenge.1 Calcification and inflammation are thought to play key pathophysiological roles. Indeed, the amount of established calcium in the valve correlates with disease severity and predicts future adverse cardiovascular events.2,3 Although computed tomography (CT) and echocardiography can provide measures of established valvular calcification, they cannot directly assess ongoing calcification activity, which is considered to be the main driver of disease progression.

Clinical Perspective on p 378

Recent reports have investigated 2 positron emission tomography (PET) radiotracers, 18F-sodium fluoride (18F-NaF) and 18F-fluorodeoxyglucose (18F-FDG), as measures of calcification activity and inflammation, respectively, in the aortic valve,4 coronary arteries,5,6 and major vessels.7 18F-FDG PET has become a widely used tool for the assessment of inflammation in the aorta and carotid arteries, with uptake correlating with macrophage burden.8 Several studies have investigated its uptake in AS, although histological validation is lacking.4,9 18F-NaF has been used as a bone tracer for >40 years, displaying increased activity in conditions associated with increased bone metabolism such as Paget disease. In bone, it is thought to bind and then incorporate into exposed hydroxyapatite crystals via an exchange mechanism with hydroxyl groups to form fluoroapatite. Given that hydroxyapatite is also a key structural component of calcification in the aortic valve and vascular atheroma, it is presumed that similar mechanisms explain its accumulation in these tissues. However, this remains hypothetical. The principal aims of the present study were, therefore, to validate the use of 18F-NaF and 18F-FDG in AS by comparing in vivo radiotracer uptake with immunohistochemistry of calcification and inflammation in excised valvular tissue and to investigate whether either of these agents predicts disease progression at 1 year.

Methods

Patient Populations

Two cohorts of patients with AS were recruited into this study: (1) 12 patients undergoing valve replacement surgery, and (2) 18 patients with asymptomatic disease under surveillance at the Edinburgh Heart Center. The latter cohort was randomly selected for repeat scanning from a larger, previously described population who underwent baseline PET imaging.4,5,10

All patients were >50 years of age, and exclusion criteria included a normal aortic valve, insulin-dependent diabetes mellitus, end-stage renal failure, life expectancy of <2 years, and metastatic malignancy. Patients with severe AS were excluded from the cohort of patients under surveillance because of the potential for disease progression and symptom development before the follow-up 1-year scan. The study was performed in accordance with the Declaration of Helsinki and after local research ethics committee approval. All patients provided written informed consent before participating.

Baseline Assessment

All patients underwent full clinical assessment at baseline, and AS severity was assessed using Doppler and 2-dimensional echocardiography by means of the peak transvalvular velocity, mean gradient, and aortic valve area according to American Heart Association/American College of Cardiology guidelines.11 Aortic sclerosis was defined as thickening of the aortic valve cusps in the absence of accelerated flow (<2 m/s) through the valve. Combined PET and CT scans of the aortic valve were performed using a hybrid scanner (Biograph mCT; Siemens Medical Systems, Erlangen, Germany) 60 minutes after administration of 125 MBq of 18F-NaF. Subsequently, a second PET/CT scan was performed using the same hybrid scanner 90 minutes after administration of 200 MBq of 18F-FDG. Glucose is a major energy source of the myocardium, so that intense 18F-FDG uptake frequently occurs, spilling over and contaminating the signal in the valve. We attempted to reduce myocardial uptake by asking patients to avoid carbohydrates for 24 hours before their 18F-FDG scan, thereby switching the myocardium from glucose to free fatty-acid metabolism. Myocardial 18F-FDG uptake was assessed within regions of interest (ROIs) placed in the basal septum of the left ventricle and classified as being adequately suppressed if mean standard uptake values were <5.0.4 An ECG-gated breath-hold CT scan (noncontrast enhanced, 40 mA/rot [CareDose]; 100 kV) was performed for calculation of the aortic valve calcium score using dedicated analysis software (VScore; Vital Images, Minnetonka, MN) on axial scans.12 Particular care was taken to differentiate valvular calcium from that in the aortic root and mitral valve annulus.6 At 1-year follow-up, patients in the surveillance cohort underwent repeat clinical assessment and CT calcium scoring using the same protocol.

Quantification of Aortic Valve PET Activity

18F-NaF and 18F-FDG uptake in the aortic valve was quantified using an Osirix workstation (OsiriX version 3.5.1 64-bit; OsiriX Imaging Software, Geneva, Switzerland) as reported previously.4,5 Briefly, fused PET-CT images were reoriented into the plane of the valve and circular ROIs drawn on adjacent 3-mm slices until the entire valve had been examined. For 18F-NaF, ROIs were placed around the perimeter of the valve while excluding the aortic root (whole-valve technique). To reduce the potential for myocardial 18F-FDG activity contaminating the aortic valve signal, ROIs for this tracer were drawn in the center of the valve as previously described (center-valve technique).4,9 Within these ROIs, mean standard uptake values were calculated for each slice, averaged, and corrected for blood pool activity to provide mean tissue/background ratios. Mean tissue/background ratios were selected prospectively for subsequent comparisons with histology and disease progression because this measure was felt to best represent tracer uptake across the valve as a whole.

Distribution of 18F-NaF in the Aortic Valve Relative to Calcium Scoring

We undertook a voxel-by-voxel analysis comparing the distribution of calcium on CT with 18F-NaF uptake. ROIs were drawn around the valve, and each voxel was assessed for the presence of calcium (>130 HU) and increased 18F-NaF uptake (tissue/background ratio max >1.97 based on the highest uptake in the control cohort of our previous study)4 using dedicated software MATLAB (Mathworks Inc, Natick, MA). We hypothesized that regions of completely novel calcium development might have an even more important effect on disease progression, and therefore we calculated the percentage of the valve with increased radiotracer uptake in the absence of underlying calcium on CT (% of PET-positive but CT-negative pixels).

Histological Assessment

In the patients undergoing aortic valve replacement, the aortic valve was removed at the time of operation, with care taken to preserve the integrity of the valve architecture. Samples were then fixed in 4% paraformaldehyde for 24 hours. Plaques were decalcified in EDTA for 10 days and embedded in paraffin, and 5-μm sections were prepared. Immunohistochemical staining for osteocalcin (antihuman mouse mAb ab13418; Abcam), CD68 (antihuman mouse clone PG-M1 m0876, DAKO), and tissue nonspecific alkaline phosphatase (TNAP; antihuman rabbit pAb CAT#LF PA50004; Abfrontier) was then undertaken after heat-induced epitope retrieval using a Citrate Buffer pH 6 (Novocastra Leica microsystems) in a decloaking chamber. Osteocalcin staining required no heat-induced epitope retrieval. Sections were stained using a Leica Vision Biosystems Bond×immunostaining robot. After blocking in peroxide for 10 minutes, sections were incubated with the specific antihuman antibodies for 2 hours at room temperature at the following dilutions: osteocalcin 1:200, TNAP 1:100, and CD68 1:100. All incubation steps were followed by washing in TBS/Tween. Sections for osteocalcin and CD68 were incubated for 15 minutes with prepolymer/postprimary followed by 15 minutes with polymer (HRP) for all antibodies before 3,3′-diaminobenzidine visualization and hematoxylin counterstain. Sections were dehydrated in graded ethanol and cleared in xylene before cover slipping in Pertex.

Images were taken on a Zeiss Axioskop2 fitted with an Axiocam MRc digital camera using Axiovision software. Tissue cross-sectional area on each section was manually delineated using Image Pro Plus 5 (Rockville, MD). Immunohistochemical staining for osteocalcin and TNAP was identified by visual assessment and quantified using automated color-based segmentation by a trained observer blinded to the PET data. Staining was expressed as a percentage of the total valve area. Macrophage infiltration using CD68 was assessed using a similar approach but with an object size set threshold applied at 20×10 pixels to limit counting to cell-sized objects. The density of cell staining in the valve tissue was expressed as cells per square millimeter. This technique was also used to identify cellular staining for TNAP and osteocalcin.

Reproducibility Studies

Interobserver reproducibility of the immunohistochemical data was investigated. Tissue staining with alkaline phosphatase, CD68, and osteocalcin was quantified in 5 valves independently by 2 trained observers (W.S.A.J. and A.T.V.).

Autoradiography

Clinical PET systems have limited resolution. To gain further information about the precise localization of the 18F-NaF signal in aortic valve tissue, we undertook autoradiography. Nondecalcified valvular tissue was rapidly cooled in dry ice and then sectioned at 7-μm thickness using a cryostat (CM1520; Wetzlar, Germany). Sections for autoradiography were mounted on Superfrost slides (Gerhard Menzel, Braunschweig, Germany) before treatment with spray fixative. Sections were bathed in a solution of 18F-NaF at a concentration close to in vivo imaging concentrations (1 kBq/mL) for 60 minutes and then rinsed with PBS. A freshly blanked phosphor screen was then placed over the slides and an overnight exposure undertaken. The screen was then read using a FujiFilm FLA-5100 Fluorescent Image Analyser (Raytek Scientific Limited, Sheffield, UK). Sections adjacent to those used for autoradiography were stained for elemental phosphate (ie, calcium orthophosphate) using Von Kossa stain, and after surface decalcification in situ with Von Ebner solution, for TNAP and osteocalcin. Sections were then manually registered and examined for colocalization with 18F-NaF signal.

Statistical Methods

Continuous variables were assessed for normality both visually and using the D’Agostino-Pearson test. Variables were expressed as either mean±SD or median with interquartile ranges (IQRs) subject to whether they approximated a normal distribution. Categorical data were presented as n (%). The 95% normal range for differences between sets of immunohistochemical measurements (the limits of agreement) were estimated using Bland–Altman analysis by multiplying the SD of the mean difference by 1.96.13 Intraclass correlation coefficients with 95% confidence intervals were calculated for interobserver variation. Baseline and follow-up calcium scores approximated a normal distribution and were compared using a paired t test. However, despite attempts at data transformation, the changes in calcium scores were not normally distributed, and correlations with CT progression data were assessed using Spearman correlation and linear regression analysis. We acknowledge the limitations in using linear regression in the context of a non-normal distribution. A 2-sided P<0.05 was regarded as statistically significant. Statistical analysis was performed with the use of Graph Pad Prism version 6.0 (GraphPad Software Inc, San Diego, CA).

Results

Histology Cohort

Twelve patients with symptomatic AS were recruited into the histology cohort (8 men; 76±6 years of age; peak aortic valve velocity, 4.6±0.9 m/s). Patients underwent PET scanning a median of 92 days before surgical aortic valve replacement. Eight patients received both 18F-NaF and 18F-FDG PET scans. In addition, 2 had a single 18F-NaF scan, whereas 2 more had a single 18F-FDG scan. Thus, 10 valves were available for the histological validation of each tracer. No patient had a significant perioperative complication (Table 1). Effective myocardial suppression of 18F-FDG activity was achieved in 40% (median myocardial standard uptake value, 5.4; IQR, 1.9–10.4).

Table 1. Baseline Characteristics of Progression Cohort

TotalAortic SclerosisMild Aortic StenosisModerate Aortic Stenosis
No.186 (33)7 (39)5 (28)
Age, y75 (71–79)74 (70–78)74 (69–80)79 (70–83)
Men15 (83)4 (66)7 (100)4 (80)
Hypertension13 (72)3 (50)5 (71)5 (100)
Hyperlipidemia12 (67)4 (66)5 (71)3 (60)
Diabetes mellitus5 (28)1 (16)4 (57)0 (0)
Ischemic heart disease13 (72)4 (66)6 (86)3 (60)
Serum creatinine, μmol/L92±2985 (72–92)91 (67–125)84 (69–133)
Cigarette smoking0 (0)0 (0)0 (0)0 (0)
Peak aortic valve velocity, m/s2.6 (1.8–3.1)1.7 (1.6–1.8)2.4 (2.1–2.6)3.4 (3.2–3.6)
Aortic valve area, cm21.68 (1.26–2.28)1.92 (1.8–2.1)1.63 (1.42–1.87)1.03 (0.78–1.18)
Mean gradient, mm Hg10.8 (7.0–16.5)6.2 (4.8–7.0)11.0 (9.3–14.0)22.0 (18.8–27.2)
Aortic valve calcium score, AU314 (193–540)106 (13–204)355 (211–536)1167 (436–1472)
Time between CT scans, d386 (377–409)390 (375–408)394 (376–426)183 (360–399)
18F-FDG dose injected, MBq193 (188–196)193 (185–205)191(185–194)194 (190–206)
18F-NaF dose injected, MBq123 (120–126)124 (117–127)123 (117–128)123 (120–126)

Categorical displayed as total number (%). Median (interquartile range). CT indicates computed tomography; 18F-FDG, 18F-fluorodeoxyglucose; and 18F-NaF, 18F-sodium fluoride.

Immunohistochemistry and Autoradiography

All valve samples displayed positive cellular staining for TNAP (225 cells/mm2 valve tissue; IQR, 143–328), osteocalcin (130 cells/mm2 valve tissue; IQR, 85–274), and CD68 (172 cells/mm2 valve tissue; IQR, 73–271; Figure 1). Extensive TNAP and osteocalcin staining was also observed in the extracellular matrix, occupying approximately a sixth of the valve area sampled (17±5% and 17±7%, respectively).

Figure 1.

Figure 1. Histology and 18F-sodium fluoride (NaF) autoradiography of excised aortic valve tissue from patients with aortic stenosis. AF, Fixed, decalcified, and paraffin-embedded aortic valve tissue after exposure to tissue nonspecific alkaline phosphatase (TNAP), osteocalcin, and CD68 antibodies. AC, Widespread positive staining for TNAP, osteocalcin, and CD68 (magnification ×4) in the extracellular matrix, which is also observed on an individual cellular level (DF, magnification ×20), respectively. G-I, Three adjacent and consecutive aortic valve leaflet sections displaying positive immunohistochemical staining for osteocalcin (I, magnification ×4) that colocalizes to areas of maximal 18F-NaF uptake on autoradiography (H). These likely represent areas of ongoing calcification activity, which extend beyond the areas of established calcium identified in black by Von Kossa stain (G, magnification ×4).

On autoradiography, 18F-NaF uptake was observed to colocalize closely with staining for structural calcium phosphate, TNAP, and osteocalcin (Figure 1). However, signal was also clearly apparent in areas free of macroscopically visible calcium, thus highlighting the sensitivity of 18F-NaF in the detection of newly evolving calcification.

Reproducibility of Immunohistochemistry

Interobserver reproducibility was good for the quantification of osteocalcin and TNAP staining, as well as CD68 cell counting. All observations were characterized by an absence of fixed or proportional biases, narrow limits of agreement (−13.4% to 9.3%, −8.0% to 5.0%, and −7.9% to 9.6%, respectively), and interclass correlation coefficient values of 0.90 (0.35–0.99), 0.88 (0.60–0.97), and 0.99 (0.99–1.00), respectively (Tables 2 and 3).

Table 2. Histology Cohort Data

Baseline characteristics
 No.12
 Age, y76±6
 Men9 (75)
 Hypertension8 (66)
 Hyperlipidemia5 (42)
 Ischemic heart disease4 (33)
 Cigarette smoking1 (8)
 Diabetes mellitus0 (0)
 Serum creatinine, μmol/L87±26
 Peak aortic valve velocity (m/s)4.6±0.9
 Aortic valve area, cm20.70 (0.53–0.97)
 Mean gradient, mm Hg48 (44–65)
 Aortic valve calcium score, AU5343 (3114–6292)
 Aortic sclerosis0 (0%)
 Mild aortic stenosis0 (0%)
 Moderate aortic stenosis3 (25%)
 Severe aortic stenosis9 (75%)
 Time between 18F-NaF scan and AVR, d92 (24–345)
 Time between 18F-FDG scan and AVR, d96 (23–331)
 18F-FDG dose injected, MBq200 (193–209)
 18F-NaF dose injected, MBq129 (119–132)
In vivo aortic valve PET data
 18F-NaF uptake (mean TBR)2.15 (1.98–2.48)
 18F-FDG uptake (mean TBR)1.40 (1.31–1.76)

Categorical data are displayed as n (%). Normally distributed data are displayed as mean±SD. Non-normally distributed data are distributed as median (interquartile range). AVR indicates aortic valve replacement; FDG, fluorodeoxyglucose; NaF, sodium fluoride; PET, positron emission tomography; and TBR, tissue/background ratio.

Table 3. Immunohistochemical Analysis

HistologyOsteocalcinTNAPCD68
No.101010
Mean valve area analyzed, mm2234±152253±116190±86
% staining of the valve17±717±5n/a
Positive cellular staining, cells/mm2130 (85–274)225 (143–328)172 (73–271)
Interobserver reproducibility
 Mean difference−2.1%−1.5%0.8%
 Limits of agreement−13.4% to 9.3%−8.0% to 5.0%−7.9% to 9.6%
 ICC0.88 (0.60–0.97)0.90 (0.35–0.99)0.99 (0.99–1.00)

Categorical data are displayed as n (%). Normally distributed data are displayed as mean±SD. Non-normally distributed data are distributed as median (interquartile range). ICC as value (95% confidence interval). ICC indicates interclass correlation coefficient; n/a, not applicable; and TNAP, tissue nonspecific alkaline phosphatase.

Correlation With Radiotracer Uptake

There was a good correlation between in vivo valvular 18F-NaF uptake and both alkaline phosphatase (r=0.65 [95% confidence interval, 0.03–0.90]; P=0.04) and osteocalcin (r=0.68 [0.10–0.91]; P=0.03; Figure 2) staining of the excised tissue. In comparison, there was no association between 18F-FDG uptake and CD68 staining in the valve (r=−0.43; P=0.22).

Figure 2.

Figure 2. Correlations between in vivo aortic valve positron emission tomography (PET) activity and histological markers of calcification and inflammation. A, 18F-Sodium fluoride (NaF) vs tissue nonspecific alkaline phosphatase (TNAP). A good correlation was observed between the percentage aortic valve tissue staining for TNAP and the valvular 18F-NaF activity (mean tissue to background ratio [TBR]); r=0.65, P=0.04. B, 18F-NaF vs osteocalcin. Again a strong correlation was observed between the percentage surface area of the valve stained with osteocalcin and the aortic valve 18F-NaF PET activity (mean TBR); r=0.68, P=0.03. C, 18F-fluorodeoxyglucose (FDG) vs CD68. A poor correlation was observed between CD68 staining on immunohistochemistry and 18F-FDG PET activity in the aortic valve (mean TBR); r=−0.43, P=0.22.

Disease Progression

Of the 18 patients (75±6 years of age; 17 men; peak aortic-jet velocity, 2.6±0.9 m/s) reassessed at a median interval of 386 days (Table 1), 6 had aortic sclerosis, 7 had mild AS, and 5 had moderate AS. Effective myocardial suppression of 18F-FDG uptake was achieved in 66% (median myocardial standard uptake value, 3.6; IQR, 2.0–5.4).

A correlation was observed between baseline aortic valve calcium scores on CT and 18F-NaF activity on PET (r=0.74 [0.42–0.90]; P≤0.001). However, as described previously, the pattern of 18F-NaF uptake was distinct from the distribution of established calcium.4,5,14 Indeed 18F-NaF uptake in the absence of underlying calcium occupied a median of 8.3% (IQR 1.6–23.4) of the total valve area, emphasizing that 18F-NaF provides distinct and complementary information to CT calcium scoring (Figure 1).

At 1 year, aortic valve calcium scores increased from 314 (193–540) to 365 (207–934) AU (P<0.01). Interestingly, these regions of novel calcium developed in much the same distribution as the observed baseline 18F-NaF uptake (Figure 3A and 3B). Indeed, we observed an excellent correlation between baseline valvular 18F-NaF PET uptake and the change in calcium score after 1 year (r=0.66 [0.27–0.86]; P=0.003; Figure 3C). This was similar to that observed for the current gold standard method of prediction: the baseline calcium score (r=0.58 [0.15–0.82]; P=0.01; Figure 3D) improved further when only increased 18F-NaF uptake in the absence of underlying CT macrocalcification was considered (r=0.75 [0.42–0.90]; P=0.01). No statistically significant correlation was observed between 18F-FDG uptake and the subsequent change in CT calcium score (r=−0.11 [−0.56 to 0.39]; P=0.66; Figure 3E).

Figure 3.

Figure 3. Change in aortic valve computed tomography (CT) calcium score and 18F-sodium fluoride (NaF) positron emission tomography (PET) activity after 1 year. A and B, Coaxial short axis views of the aortic valve from 2 patients with mild aortic stenosis (top and bottom). On baseline CT scans (left) established regions of macrocalcification appear white. Baseline fused 18F-NaF PET and CT scans (middle) show intense 18F-NaF uptake (red, yellow regions) both overlying and adjacent to existing calcium deposits on the CT. One-year follow-up CT scans (right) demonstrate increased calcium accumulation in much the same distribution as the baseline PET activity. CE, Predictors of progression in aortic valve calcium score. An excellent correlation was observed between baseline 18F-NaF activity in the aortic valve and the subsequent change in calcium score at 1 year r=0.66, P<0.01 (A). This matched the current gold standard predictor of disease progression the baseline calcium score r=0.58, P=0.01 (B). By contrast, there was a poor correlation with 18F-fluorodeoxyglucose (FDG) activity in the valve r=−0.11, P=0.66 (C).

Discussion

We provide the first preliminary evidence that valvular 18F-NaF uptake acts as a marker of calcification activity in patients with AS. Not only did uptake values demonstrate a correlation with histological markers of active calcification (TNAP and osteocalcin), but they were also a good predictor of the subsequent progression in aortic valve CT calcium scores at 1 year. In contrast, 18F-FDG uptake did not correlate with CD68 staining on histology or the progression in calcium scores. Our data indicate that 18F-NaF holds promise as a biomarker of disease activity in patients with AS.

The pathophysiology of AS is incompletely understood, delaying the development of biomarkers and effective medical therapies. Calcification and inflammation are thought to play a key pathological role,1 so that noninvasive markers of their activity are of interest in better understanding the cause of this condition and in predicting disease progression.

Recent studies have investigated 18F-NaF PET as a marker of vascular calcification in AS3 and atherosclerosis affecting the aorta,7 coronary,5,6 and carotid arteries.15 However, this is the first study to provide histological validation of 18F-NaF uptake in vascular tissue. In bone, 18F-NaF is thought to incorporate onto the surface of hydroxyapatite crystal.14 Given that hydroxyapatite is also a key component of vascular calcification, it too has been the presumed radiotracer target in AS and atherosclerosis. This hypothesis is supported by our autoradiography and immunohistochemical data, demonstrating a good correlation between 18F-NaF activity and osteocalcin staining: a well-recognized osteogenic protein that itself binds to hydroxyapatite.

Given that 18F-NaF binds to a structural component of vascular calcification, why then does it not simply label all regions of macrocalcification identified by CT? Indeed it is common for regions of dense calcium on CT to show no 18F-NaF uptake. This phenomenon is likely related to the available surface area of exposed hydroxyapatite crystal to which the 18F-fluoride ion can adsorb and the inactivity of established areas of calcification. 18F-NaF uptake is much greater at sites of evolving powdery microcalcification than established regions of field calcification in which the core of hydroxyapatite is internalized and, therefore, hidden from the 18F-NaF tracer. Thus, 18F-NaF binds more readily to regions of developing calcium and acts as a marker of calcification activity, providing distinct information to calcium scoring. In contrast, the latter quantifies regions of established macroscopic calcium in the valve but cannot inform whether the process of calcification is quiescent or active. Again this hypothesis is supported by our data. We have demonstrated a strong correlation between in vivo 18F-NaF uptake and staining for 1 of the key enzymes regulating mineralization: TNAP. This enzyme is expressed in the early stages of new calcium formation and is known to work by breaking down pyrophosphate: a potent inhibitor of mineralization.16 Furthermore, as one would expect from a measure of activity, baseline 18F-NaF uptake closely correlated with the subsequent change in calcium score at 1 year. Indeed, 18F-NaF uptake performed as well as the current gold standard method of prediction, the degree of established calcium in the valve at baseline.2,3 However, larger studies are now required to compare these 2 techniques, whereas calcium scoring may be easier to obtain; changes in the 18F-NaF PET signal are likely to occur more quickly, making it a more attractive technique with which to assess the early and more immediate effects of novel treatment strategies.

Interestingly, the pattern of 18F-NaF uptake may be important, with 18F-NaF uptake remote from established macrocalcification on CT offering the best prediction of calcium score progression in our cohort. The spatial resolution of PET/CT is ≈4 mm, and we acknowledge that the voxel-by-voxel analysis used to establish this observation is at the limit of resolution for PET imaging. Nevertheless, the strong correlation with progression is of interest and indicates that further investigation of the spatial distribution of 18F-NaF uptake is warranted.

The results of valvular 18F-FDG imaging were somewhat disappointing and surprising given previous data suggesting an important role for inflammation in AS.17 Although correlations between 18F-FDG uptake and macrophage burden have previously been demonstrated in regions of aortic and carotid atheroma,8 we were unable to replicate this with respect to the valve. There are several explanations for this discrepancy. The first is the close proximity of the valve to the myocardium. As discussed, avid uptake of 18F-FDG by the left ventricular myocardium can spill over into the aortic valve contaminating its signal. Unfortunately, despite the stringent dietary restrictions and center-valve analysis technique, it remains possible that myocardial contamination occurred, confounding the correlation with CD68 immunohistochemistry. Indeed poor myocardial suppression was achieved in the histology group, perhaps reflecting their advanced disease and symptomatic status. Alternative methods have been used to reduce further this myocardial uptake, including administration of heparin18 and a high-fat drink before scanning.19 However, these make the practicalities of scanning more difficult and have yet to show a clear advantage over dietary restrictions. An alternative explanation for the poor correlation with histology is that the aortic valve 18F-FDG signal relates to uptake by nonmacrophage cell types within the valve, such as osteoblasts, or is governed by external factors, such as hypoxia.20 In this scenario, one might still expect 18F-FDG to predict disease progression, but once again this was not evident in our cohort. It would, therefore, seem that 18F-FDG holds less potential as a predictor of disease progression than 18F-NaF does, although it remains possible that longer periods of follow-up are required to detect such an association. Indeed on occasion, we also observed 18F-NaF activity that did not translate into a detectable change in calcium score at 1 year. AS is a slowly developing condition, so it is likely to take time for relatively low levels of 18F-NaF or 18F-FDG uptake to translate into new areas of macrocalcification detectable on CT imaging. Larger studies with longer follow-up are, therefore, required to address this issue, confirm our preliminary data, and assess whether 18F-NaF PET can predict disease progression with respect to echocardiographic parameters of valvular stenosis.

Conclusions

We provide the first preliminary data to support 18F-NaF as a marker of valve calcification activity in AS and as a potential method for predicting disease progression.

Acknowledgments

We acknowledge the support of staff at the Edinburgh Heart Centre at the Royal Infirmary of Edinburgh, the radiography and radiochemistry staff of the Clinical Research Imaging Centre, and the histology staff at the Queens Medical Research Institute.

Footnotes

*Drs Dweck and Jenkins contributed equally to this work.

Correspondence to Marc R. Dweck, MD, PhD, Centre for Cardiovascular Science, University of Edinburgh, 47 Little France Crescent, Edinburgh, UK, EH16 4TJ. E-mail

References

  • 1. Quarto C, Dweck MR, Murigu T, Joshi S, Melina G, Angeloni E, Prasad SK, Pepper JR. Late gadolinium enhancement as a potential marker of increased perioperative risk in aortic valve replacement.Interact Cardiovasc Thorac Surg. 2012; 15:45–50.CrossrefMedlineGoogle Scholar
  • 2. Messika-Zeitoun D, Bielak LF, Peyser PA, Sheedy PF, Turner ST, Nkomo VT, Breen JF, Maalouf J, Scott C, Tajik AJ, Enriquez-Sarano M. Aortic valve calcification: determinants and progression in the population.Arterioscler Thromb Vasc Biol. 2007; 27:642–648.LinkGoogle Scholar
  • 3. Rosenhek R, Binder T, Porenta G, Lang I, Christ G, Schemper M, Maurer G, Baumgartner H. Predictors of outcome in severe, asymptomatic aortic stenosis.N Engl J Med. 2000; 343:611–617.CrossrefMedlineGoogle Scholar
  • 4. Dweck MR, Jones C, Joshi NV, Fletcher AM, Richardson H, White A, Marsden M, Pessotto R, Clark JC, Wallace WA, Salter DM, McKillop G, van Beek EJ, Boon NA, Rudd JH, Newby DE. Assessment of valvular calcification and inflammation by positron emission tomography in patients with aortic stenosis.Circulation. 2012; 125:76–86.LinkGoogle Scholar
  • 5. Dweck MR, Chow MW, Joshi NV, Williams MC, Jones C, Fletcher AM, Richardson H, White A, McKillop G, van Beek EJ, Boon NA, Rudd JH, Newby DE. Coronary arterial 18F-sodium fluoride uptake: a novel marker of plaque biology.J Am Coll Cardiol. 2012; 59:1539–1548.CrossrefMedlineGoogle Scholar
  • 6. Joshi NVVA, Williams MC, Shah ASV, Calvert PA, Craighead FHM, Yeoh SE, Wallace WA, Salter D, Fletcher AM, van Beek EJR, Flapan AD, Uren NG, Behan MWH, Cruden NLM, Mills NL, Fox KAA, Rudd JHF, Dweck MR, Newby DE. 18F-Fluoride positron emission tomography identifies ruptured and high-risk coronary atherosclerotic plaques. Lancet. 2013; 13:61754–61757.Google Scholar
  • 7. Derlin T, Tóth Z, Papp L, Wisotzki C, Apostolova I, Habermann CR, Mester J, Klutmann S. Correlation of inflammation assessed by 18F-FDG PET, active mineral deposition assessed by 18F-fluoride PET, and vascular calcification in atherosclerotic plaque: a dual-tracer PET/CT study.J Nucl Med. 2011; 52:1020–1027.CrossrefMedlineGoogle Scholar
  • 8. Tawakol A, Migrino RQ, Bashian GG, Bedri S, Vermylen D, Cury RC, Yates D, LaMuraglia GM, Furie K, Houser S, Gewirtz H, Muller JE, Brady TJ, Fischman AJ. In vivo 18F-fluorodeoxyglucose positron emission tomography imaging provides a noninvasive measure of carotid plaque inflammation in patients.J Am Coll Cardiol. 2006; 48:1818–1824.CrossrefMedlineGoogle Scholar
  • 9. Marincheva-Savcheva G, Subramanian S, Qadir S, Figueroa A, Truong Q, Vijayakumar J, Brady TJ, Hoffmann U, Tawakol A. Imaging of the aortic valve using fluorodeoxyglucose positron emission tomography increased valvular fluorodeoxyglucose uptake in aortic stenosis.J Am Coll Cardiol. 2011; 57:2507–2515.CrossrefMedlineGoogle Scholar
  • 10. Dweck MR, Khaw HJ, Sng GK, Luo EL, Baird A, Williams MC, Makiello P, Mirsadraee S, Joshi NV, van Beek EJ, Boon NA, Rudd JH, Newby DE. Aortic stenosis, atherosclerosis, and skeletal bone: is there a common link with calcification and inflammation?Eur Heart J. 2013; 34:1567–1574.CrossrefMedlineGoogle Scholar
  • 11. Bonow RO, Carabello BA, Chatterjee K, de Leon AC, Faxon DP, Freed MD, Gaasch WH, Lytle BW, Nishimura RA, O’Gara PT, O’Rourke RA, Otto CM, Shah PM, Shanewise JS, Smith SC, Jacobs AK, Adams CD, Anderson JL, Antman EM, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Nishimura R, Page RL, Riegel B. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing Committee to Revise the 1998 guidelines for the management of patients with valvular heart disease) developed in collaboration with the Society of Cardiovascular Anesthesiologists endorsed by the Society for Cardiovascular Angiography and Interventions and the Society of Thoracic Surgeons.J Am Coll Cardiol. 2006; 48:e1–e148.CrossrefMedlineGoogle Scholar
  • 12. Cowell SJ, Newby DE, Burton J, White A, Northridge DB, Boon NA, Reid J. Aortic valve calcification on computed tomography predicts the severity of aortic stenosis.Clin Radiol. 2003; 58:712–716.CrossrefMedlineGoogle Scholar
  • 13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement.Lancet. 1986; 1:307–310.CrossrefMedlineGoogle Scholar
  • 14. Derlin T, Richter U, Bannas P, Begemann P, Buchert R, Mester J, Klutmann S. Feasibility of 18F-sodium fluoride PET/CT for imaging of atherosclerotic plaque.J Nucl Med. 2010; 51:862–865.CrossrefMedlineGoogle Scholar
  • 15. Derlin T, Wisotzki C, Richter U, Apostolova I, Bannas P, Weber C, Mester J, Klutmann S. In vivo imaging of mineral deposition in carotid plaque using 18F-sodium fluoride PET/CT: correlation with atherogenic risk factors.J Nucl Med. 2011; 52:362–368.CrossrefMedlineGoogle Scholar
  • 16. Narisawa S, Harmey D, Yadav MC, O’Neill WC, Hoylaerts MF, Millán JL. Novel inhibitors of alkaline phosphatase suppress vascular smooth muscle cell calcification.J Bone Miner Res. 2007; 22:1700–1710.CrossrefMedlineGoogle Scholar
  • 17. Toutouzas K, Drakopoulou M, Synetos A, Tsiamis E, Agrogiannis G, Kavantzas N, Patsouris E, Iliopoulos D, Theodoropoulos S, Yacoub M, Stefanadis C. In vivo aortic valve thermal heterogeneity in patients with nonrheumatic aortic valve stenosis the: first in vivo experience in humans.J Am Coll Cardiol. 2008; 52:758–763.CrossrefMedlineGoogle Scholar
  • 18. Ishimaru S, Tsujino I, Takei T, Tsukamoto E, Sakaue S, Kamigaki M, Ito N, Ohira H, Ikeda D, Tamaki N, Nishimura M. Focal uptake on 18F-fluoro-2-deoxyglucose positron emission tomography images indicates cardiac involvement of sarcoidosis.Eur Heart J. 2005; 26:1538–1543.CrossrefMedlineGoogle Scholar
  • 19. Cheng VY SP, Ahlen M, Thomson LEJ, Waxman AD, Berman DS. Impact of carbohydrate restriction with and without fatty acid loading on myocardial 18F-FDG uptake during PET: a randomized controlled trial.J Nucl Cardiol. 2010; 17:286–291.CrossrefMedlineGoogle Scholar
  • 20. Folco EJ, Sheikine Y, Rocha VZ, Christen T, Shvartz E, Sukhova GK, Di Carli MF, Libby P. Hypoxia but not inflammation augments glucose uptake in human macrophages: implications for imaging atherosclerosis with 18fluorine-labeled 2-deoxy-D-glucose positron emission tomography.J Am Coll Cardiol. 2011; 58:603–614.CrossrefMedlineGoogle Scholar

CLINICAL PERSPECTIVE

Aortic stenosis is the most common form of valvular heart disease in the developed world; however, its rate of progression can vary greatly between patients, and as yet we lack effective medical therapies. Progressive valvular calcification is a key pathophysiological process driving aortic stenosis. Recent studies have suggested that the positron emission tomography (PET) tracer 18F-sodium fluoride (18F-NaF) can measure vascular calcification activity, although validation of this technique has previously been lacking. In this study, we sought to provide such validation, comparing aortic valve 18F-NaF uptake with both histological measures of calcification activity and the progression of aortic valve calcium scores on computed tomography. Ten patients scheduled for aortic valve replacement underwent 18F-NaF PET before their operation with valvular PET uptake, demonstrating a significant correlation with histological markers of calcification activity on the excised aortic valve tissue. In a separate cohort, we assessed 18 patients with calcific aortic valve disease who underwent 18F-NaF PET and computed tomography calcium scoring of the aortic valve both at baseline and after 1 year. Once again a good correlation was observed between the baseline PET uptake and change in calcium score. Our results are promising, providing the first preliminary validation of 18F-NaF as a marker of calcification activity in the aortic valve. More studies are now required to investigate 18F-NaF PET as a method for predicting disease progression in aortic stenosis and assessing the early efficacy of novel treatments aimed at reducing calcification activity in this condition.