Assessment of Inflammation and Calcification in Pseudoxanthoma Elasticum Arteries and Skin with 18F-FluroDeoxyGlucose and 18F-Sodium Fluoride Positron Emission Tomography/Computed Tomography Imaging: The GOCAPXE Trial

Background: Pseudoxanthoma elasticum (PXE) is an inherited metabolic disease characterized by elastic fiber fragmentation and ectopic calcification. There is growing evidence that vascular calcification is associated with inflammatory status and is enhanced by inflammatory cytokines. Since PXE has never been considered as an inflammatory condition, no incidence of chronic inflammation leading to calcification in PXE has been reported and should be investigated. In atherosclerosis and aortic stenosis, positron emission tomography combined with computed tomographic (PET-CT) imaging has demonstrated a correlation between inflammation and calcification. The purpose of this study was to assess skin/artery inflammation and calcification in PXE patients. Methods: 18F-FluroDeoxyGlucose (18F-FDG) and 18F-Sodium Fluoride (18F-NaF) PET-CT, CT-imaging and Pulse wave velocity (PWV) were used to determine skin/vascular inflammation, tissue calcification, arterial calcium score (CS) and stiffness, respectively. In addition, inorganic pyrophosphate, high-sensitive C-reactive protein and cytokines plasma levels were monitored. Results: In 23 PXE patients, assessment of inflammation revealed significant 18F-FDG uptake in diseased skin areas contrary to normal regions, and exclusively in the proximal aorta contrary to the popliteal arteries. There was no correlation between 18F-FDG uptake and PWV in the aortic wall. Assessment of calcification demonstrated significant 18F-NaF uptake in diseased skin regions and in the proximal aorta and femoral arteries. 18F-NaF wall uptake correlated with CS in the femoral arteries, and aortic wall PWV. Multivariate analysis indicated that aortic wall 18F-NaF uptake is associated with diastolic blood pressure. There was no significant correlation between 18F-FDG and 18F-NaF uptake in any of the artery walls. Conclusion: In the present cross-sectional study, inflammation and calcification were not correlated. PXE would appear to more closely resemble a chronic disease model of ectopic calcification than an inflammatory condition. To assess early ectopic calcification in PXE patients, 18F-NaF-PET-CT may be more relevant than CT imaging. It potentially constitutes a biomarker for disease-modifying anti-calcifying drug assessment in PXE.

Abstract: Background: Pseudoxanthoma elasticum (PXE) is an inherited metabolic disease characterized by elastic fiber fragmentation and ectopic calcification. There is growing evidence that vascular calcification is associated with inflammatory status and is enhanced by inflammatory cytokines. Since PXE has never been considered as an inflammatory condition, no incidence of chronic inflammation leading to calcification in PXE has been reported and should be investigated. In atherosclerosis and aortic stenosis, positron emission tomography combined with computed tomographic (PET-CT) imaging has demonstrated a correlation between inflammation and calcification. The purpose of this study was to assess skin/artery inflammation and calcification in PXE patients. Methods: 18F-FluroDeoxyGlucose (18F-FDG) and 18F-Sodium Fluoride (18F-NaF) PET-CT,

Introduction
Pseudoxanthoma elasticum (PXE, OMIM 264800) is a rare disorder characterized by fragmentation and progressive calcification of elastic fibers in connective tissue of the skin, vascular system and Bruch's membrane of the retina [1]. PXE is caused by mutations in the ABCC6 gene, encoding a transmembrane ATP-binding cassette (ABC) transporter primarily expressed in the liver and kidney [2,3]. ABCC6 endogenous substrates are as yet unknown. It was recently discovered that absence of ABCC6-mediated adenosine triphosphate release from the liver, causing reduced plasma inorganic pyrophosphate (PPi) levels, underlies calcification-induced PXE [4,5]. Peripheral artery disease (PAD) resulting from calcification in the internal elastic lamina of the medial layer in muscular and elastic arteries is highly prevalent in PXE patients and mimics vascular calcification observed in other acquired metabolic diseases such as diabetes mellitus (DM) and chronic kidney disease (CKD) [6,7]. Vascular calcification is the result of a regulated process that is orchestrated by vascular smooth muscle cells (VSMCs) and develops similarly to the physiological mineralization process [6,7]. In normal vessels, VSMCs are protected from calcification by mineralization inhibitors such as PPi and matrix Gla protein [4,5,7]. However, VSMCs can die by apoptosis in response to an attack, releasing apoptotic bodies that contain calcium phosphate crystals as identified in atherosclerotic lesions and medial vascular calcification (MVC) [7]. Calcium phosphate crystals have been shown in vitro to induce a pro-inflammatory response involving cytokine release (IL-1β, IL-6, TNFα) in cultures of differentiated human macrophages that potentially leads to a vicious cycle of pro-inflammatory macrophage infiltration, extracellular matrix breakdown and VSMC apoptosis [8]. There is growing evidence that vascular calcification is connected to inflammatory status and is enhanced by inflammatory cytokines [7]. Studies have suggested that in DM-related PAD, atherosclerosis or CKD, low grade chronic inflammation (LGCI) may occur prior to vascular calcification (VC) due to immune-cell recruitment and pro/anti-inflammatory cytokine imbalance [6,7]. Since PXE has never been considered as an inflammatory condition, no incidence of chronic inflammation leading to calcification in PXE has been reported and as such requires investigation. Non-invasive techniques such as positron emission tomography combined with computed tomographic imaging (PET-CT) imaging have demonstrated a correlation between inflammation and calcification in atherosclerosis [9,10] and aortic stenosis [11]. 18F-FluroDeoxyGlucose (18F-FDG) and 18F-sodium fluoride (18F-NaF), PET tracers for inflammation and active mineral deposition respectively, have been used in several dual tracer PET-CT studies designed to investigate vascular conditions where LGCI and VC, also called inflammaging [12], are deemed key pathogenic factors [13]. The purpose of the present study was to investigate skin/artery LGCI using 18F-FDG-PET-CT, and tissue calcification using 18F-NaF-PET-CT. Additionally, artery wall 18F-NaF-PET-CT activity and arterial calcium scores obtained from CT-scans were compared. We also investigated any correlation between: 18F-FDG-PET-CT/18F-NaF-PET-CT and PWV; 18F-FDG-PET-CT and hsCRP; 18F-NaF-PET-CT and PPi plasma levels. An attempt was made to determine evidence of a distinctive circulating blood factor such as cytokines in PXE patients.

Ethical Standards
The data supporting the findings herein are available upon reasonable request from the corresponding author. The said author accepts responsibility for the reliability of all study data to which full access was provided, including data analysis.
The trial protocol was approved by the local research ethics committee (CPP Ouest II, Angers, France; EudraCT identification number: 2014-A01614-43 and CPP identification number 2014/35) and was implemented as per the most recent amendments to the Declaration of Helsinki and good clinical practice guidelines. Written informed consent was obtained from all patients prior to enrolment. The GOCAPXE study was registered with Clinicaltrials.gov on 6 March 2017 (NCT 03070860) and has been overseen by an independent data safety and monitoring committee. No control group of patients undergoing PET-CT imaging was included in this trial for obvious ethical reasons related to radiation risks. The lead author wrote the first manuscript draft, and each co-author contributed to and validated subsequent revised versions.

Patient Population
From 2017-2018, PXE patients and healthy volunteers (HVs) were enrolled prospectively in the present trial at the National Reference Center for PXE at Angers University Hospital. Following written informed consent, each participant was examined for screening purposes. During this examination detailed medical history was obtained including drug use, smoking habits, and family medical background. Each patient received a comprehensive baseline clinical examination including evaluation of each cardiovascular risk factor profile.

PXE Patients
PXE diagnosis was established genetically and/or clinically in accordance with international diagnostic criteria for clear-cut PXE: (i) evidence of angioid streaks on eye funduscopy; (ii) typical skin lesions featuring yellowish papules or large coalescent plaques in the neck/flexural region; (iii) skin biopsy demonstrating dermal elastorrhexis and calcification by positive von Kossa staining [14]. ABCC6 mutations were identified through genotyping (Supplementary Material Table S1).
Exclusion criteria were: women of childbearing age using no contraception; pregnant or breastfeeding women; diabetic patients; patients with osteopenia, inflammatory or autoimmune systemic disease; patients with high blood glucose (>11 mmol/L) due to potential competition between glucose and 18F-FDG. Each patient received a comprehensive baseline clinical examination, including evaluation of each cardiovascular risk factor profile.

Healthy Volunteers
HVs were recruited prospectively by the Clinical Investigation Center of Angers University Hospital and matched to PXE patients by age and sex.
Blood testing involved: hsCRP mg/L using standard laboratory techniques; PPi (µmol/L) [5] assay; several immunological tests. Each blood sample was drawn following an 8-hour overnight fasting period.

Multiplex Immunoassays Using Luminex ® Technology
Associations of change in functional assessment were investigated using biological assays of a set of circulating blood factors. Plasma samples from 23 PXE patients and 23 HVs were obtained from the Angers University Hospital BRC. 46 chemokines, cytokines, growth factors, lectin adhesion molecules, osteogenic factors, matrix metalloproteinase and fibrogenic factors were quantified in PXE patient and HV sera using Luminex assay kits following manufacturer instructions (R&D Systems).

PPi Assay
For PXE patients only, plasma PPi was measured by enzymatic reaction using ATP sulfurylase to convert PPi into ATP in the presence of excess adenosine-5 -phosphosulfate (Sigma-Aldrich, St. Louis, MO, USA), as previously described [5].

Carotid-Femoral PWV Measurement
Carotid-femoral PWV resulting from aortic stiffness was recorded tonometrically. Transcutaneous carotid-femoral PWV in the right common carotid and femoral arteries was recorded consecutively on a high-resolution tonometer (PulsePen, DiaTecne, Milan, Italy) [16]. ECG signals were used as a time reference. Distance (in mm) between the two recording sites was measured with a ruler and calculated as direct carotid-femoral distance corrected by a factor equal to 0.8 as recommended by the European Society of Hypertension [17]. Carotid-femoral PWV was determined by an intersecting tangent algorithm and expressed in mm/s.

PET-CT and CT-Scan Imaging Techniques
PXE patients underwent all-body 18F-FDG/18F-NaF-PET-CT and CT-scans. Spared regions such as the popliteal arteries (vascular investigation) [1] or the lumbar region (skin investigation) were regarded as negative controls in these patients.
Whole-body PET-CT images were obtained within 60 min of 18F-FDG injection, and 90 min of 18F-NaF injection.

Image Analysis: 18F-FDG/18F-NaF-PET-CT
Each image dataset was analyzed by a nuclear medicine physician on an Imagys ® workstation (Keosys ® , Saint-Herblain, France). Maximum intensity projection PET-CT images were assessed visually for evidence of radiotracer accumulation in the femoral or popliteal artery walls, as previously described [18]. In semi-quantitative analysis, maximum SUV was determined by manually drawing an individual ROI 1 cm 3 around the most fixed arterial segment ( Figure 1) that included: the left and right carotid arteries, aorta (ascending aorta, aortic arch, descending aorta, abdominal aorta), left and right iliac arteries, damaged left and right femoral arteries and spared left and right popliteal arteries, as shown on coregistered transaxial PET-CT findings. The ROI was adjusted to the vascular wall using coronal and sagittal PET-CT images. Blood pool SUVmax/mean was expressed as the SUVmax/mean of a 1 cm fixed diameter ROI drawn mid lumen in the superior and inferior vena cava, as previously described [19]. The SUVmax of each artery lesion was divided by blood pool SUVmax to ascertain TBRmax [13,20]. For the purposes of analysis, mean TBR in the left and right carotid, iliac, femoral, popliteal arteries and the proximal (ascending aorta and aortic arch) and distal aorta (descending and abdominal aorta) was calculated [18] (Figure 1). TBR in the left and right carotids and the proximal aorta was calculated by dividing the SUVmax by SUVmax derived from the superior vena cava [21]. TBR in the distal aorta, iliac, femoral and popliteal arteries was calculated by dividing the SUVmax by SUVmax derived from the inferior vena cava [21]. With respect to skin analysis, SUVmax and SUVmean were measured after a circular ROI had been drawn around three regions (spared lumbar/damaged neck/axillary folds) [22]. Mean linear SUV was calculated in the left and right axillary folds [22].

Figure 1. Whole-body 18F-FDG/18F-NaF-PET/CT to assess subclinical arterial inflammation and active mineral deposition.
Skin/arterial inflammation quantified as 18F-FDG SUVmax. Active mineral deposition quantified as 18F-NaF SUVmax. SUVmax was determined by manually drawing an individual ROI 1cm 3 around the most fixed arterial segment. TBRmax in vascular system obtained by dividing artery SUVmax by vena cava (blood pool) SUVmax. SUVmax measured from ascending to abdominal aorta as mean total aorta (meanTBR) in addition to neck, axillary fold and lumbar skin regions.

Image Analysis: CS in Lower Limb Arteries
Each PXE patient was submitted to a non-contrast-enhanced-64-row-multidetector-CT scan (Brillance 64, Philips HealthCare, Dest, The Netherlands) of the lower limbs, from the iliac crest to the tips of the toes, without injection of contrast medium [23].
Investigators blinded to patient clinical status calculated CS using automated 3D image-analysis software (Synapse 3D, Fujifilm Medical Systems, Greenwood, SC, USA). ROIs were divided into 3 segments for each individual leg: (1) the femoral segment (common and femoral arteries), extending from the iliac crest to the adductor magnus opening; (2) the popliteal segment (from adductor magnus opening to origin of anterior tibial artery); (3) the sub-popliteal segment (anterior and posterior tibialis and fibular arteries from their origin to malleolar region) [23]. Each segment length was measured using 3D scans and expressed in mm. Calcified regions with a cross-sectional area ≥0.7 mm 2 and density ranging between 150-400 HU (Hounsfield Unit) were automatically identified on cross-sectional lower-extremity images. The CS was ascertained and expressed as ALCS in each segment of both legs [24]. The ALCS of each segment was normalized to its length (arbitrary units). Arteries with CS = 0HU were regarded as non-calcified.

Statistical Analysis
Continuous variables were expressed as mean ± SD/median and IQR values. Categorical variables were expressed as counts/percentages. The Student's t-test (or Mann-Whitney Wilcoxon exact test where appropriate) was used to compare continuous variables and the chi-squared test (or Fisher's exact test where appropriate) to compare categorical variables.
The Wilcoxon signed-rank test was used to analyze LGCI on 18F-FDG-PET-CT and ectopic calcification on 18F-NaF-PET-CT.

Figure 1. Whole-body 18F-FDG/18F-NaF-PET/CT to assess subclinical arterial inflammation and active mineral deposition.
Skin/arterial inflammation quantified as 18F-FDG SUVmax. Active mineral deposition quantified as 18F-NaF SUVmax. SUVmax was determined by manually drawing an individual ROI 1 cm 3 around the most fixed arterial segment. TBRmax in vascular system obtained by dividing artery SUVmax by vena cava (blood pool) SUVmax. SUVmax measured from ascending to abdominal aorta as mean total aorta (meanTBR) in addition to neck, axillary fold and lumbar skin regions.

Image Analysis: CS in Lower Limb Arteries
Each PXE patient was submitted to a non-contrast-enhanced-64-row-multidetector-CT scan (Brillance 64, Philips HealthCare, Dest, The Netherlands) of the lower limbs, from the iliac crest to the tips of the toes, without injection of contrast medium [23].
Investigators blinded to patient clinical status calculated CS using automated 3D image-analysis software (Synapse 3D, Fujifilm Medical Systems, Greenwood, SC, USA). ROIs were divided into 3 segments for each individual leg: (1) the femoral segment (common and femoral arteries), extending from the iliac crest to the adductor magnus opening; (2) the popliteal segment (from adductor magnus opening to origin of anterior tibial artery); (3) the sub-popliteal segment (anterior and posterior tibialis and fibular arteries from their origin to malleolar region) [23]. Each segment length was measured using 3D scans and expressed in mm. Calcified regions with a cross-sectional area ≥0.7 mm 2 and density ranging between 150-400 HU (Hounsfield Unit) were automatically identified on cross-sectional lower-extremity images. The CS was ascertained and expressed as ALCS in each segment of both legs [24]. The ALCS of each segment was normalized to its length (arbitrary units). Arteries with CS = 0HU were regarded as non-calcified.

Statistical Analysis
Continuous variables were expressed as mean ± SD/median and IQR values. Categorical variables were expressed as counts/percentages. The Student's t-test (or Mann-Whitney Wilcoxon exact test where appropriate) was used to compare continuous variables and the chi-squared test (or Fisher's exact test where appropriate) to compare categorical variables.
The Wilcoxon signed-rank test was used to analyze LGCI on 18F-FDG-PET-CT and ectopic calcification on 18F-NaF-PET-CT.
Univariate analysis using linear regression or variance analysis was applied to investigate factors associated with artery wall 18F-FDG/18F-NaF uptake. Variables with p < 0.20 were then selected for multivariate analysis. The dependent variable was artery wall 18F-FDG/18F-NaF uptake (mean TBR), and independent covariates were selected on the assumption they were linked to 18F-FDG/18F-NaF uptake and CVR factors (PWV, SBP, DBP, right/left ABI, age, sex, BMI, HbA1c, smoking, total cholesterol, LDL, hsCRP, PPi).
Backward stepwise analysis was applied. Normal distribution of the measured variables was verified.
The Spearman coefficient was used to analyze correlation. A statistical significance threshold of 0.05 was adopted for all tests. SAS ® 9.4 software (SAS Institute, Cary, NC, USA) was used for statistical analysis.
18F-FDG uptake in all arteries walls (see Figure 1) was not correlated with hsCRP levels (Supplementary Material, Figure S1c).
No correlation was established between 18F-NaF uptake in any of the artery walls (see Figure 2) and plasma PPi levels (Supplementary Material, Figure S1d).

18F-FDG/18F-NaF Uptake Correlation in the Vascular Network
No significant correlation was detected between 18F-FDG and 18F-NaF uptake in any of the artery walls ( Figure 5).

18F-FDG/18F-NaF Uptake Correlation in the Vascular Network
No significant correlation was detected between 18F-FDG and 18F-NaF uptake in any of the artery walls ( Figure 5).

Assessment of Blood Circulating Factors
A significant difference was observed in MMP-2 and MMP-3 plasma levels between PXE patients and HVs ( Figure 6A,B).
No difference was observed in plasma levels of chemokines, cytokines, growth factors, lectin adhesion molecules, osteogenic factors or fibrogenic factors between PXE patients and HVs (Table 4).

Assessment of Blood Circulating Factors
A significant difference was observed in MMP-2 and MMP-3 plasma levels between PXE patients and HVs (Figure 6A,B).
No difference was observed in plasma levels of chemokines, cytokines, growth factors, lectin adhesion molecules, osteogenic factors or fibrogenic factors between PXE patients and HVs (Table 4).

Discussion
The present study has demonstrated that 18F-FDG/18F-NaF activity was significantly greater in PXE-damaged skin regions and in the proximal aorta wall, whereas 18F-NaF activity alone was greater in the femoral arteries.

PXE: A Seemingly Non-Inflammatory Condition
We have previously demonstrated that specific skin regions (neck/flexural regions) and arteries (aorta/femoral/leg) are affected by PXE lesions while the lumbar skin region and popliteal arteries are spared [1,23]. Owing to the high levels of 18F-FDG activity observed in specific regions, the question arises as to whether LGCI is involved. No histological study has so far provided evidence of inflammatory cells in skin biopsies obtained from PXE patients [25]. Higher 18F-FDG uptake in specific skin regions may reflect pathological fibroblast proliferation in PXE [26]. Similarly, histological and ultrastructural analysis [27] of PXE artery walls has failed to detect inflammatory or immune cells such as those found in vasculitis [28] or atherosclerosis [7]. Moreover, no immune-inflammatory pathway in PXE patients was identified by the Luminex study, unlike in Takayasu arteritis where Th1 and Th17 cytokines drive inflammation [28].
An increase in ascending-aorta 18F-FDG activity is the only factor giving credence to early LGCI in PXE patients. Using multivariate analysis, we have shown that this 18F-FDG uptake appears to be significantly linked to BMI and HbA1c as commonly observed in DM [21]. These factors may therefore contribute to a LGCI state in PXE patients.
Furthermore, recent studies on the ascending aorta in atherosclerosis have demonstrated that VSMCs and fibroblasts, acting as macrophages, are capable of accumulating 18F-FDG [29,30]. They conclude that ascending-aorta 18F-FDG activity does not necessarily signify inflammation and advise against regarding this region as an imaging endpoint [30,31].
Taken together, our findings imply no LGCI in PXE.

PXE as a Prime Example of Chronic Skin and Arterial Calcification
In the 23 PXE patients 18F-NaF activity was higher in the neck and axillary folds than in the lumbar skin region, suggesting active calcification on damaged skin. These results are consistent with those recently published where patients with higher skin Phenodex scores exhibited higher 18F-NaF uptake in the neck [32]. In the vascular system 18F-NaF uptake was observed exclusively in the aorta and the femoral arteries, substantiating our previous characterization of vulnerability to PXE damage in these regions [1, 23,33]. Similarly, MVC and non-calcified atherosclerotic lesions constitute a predominant leg artery calcification type in the general population. In a series of 121 leg amputees, MVC was found in 71% of femoral and crural arteries versus 25% of calcified atherosclerotic lesions [34]. MVC is, moreover, a strong predictor for major cardiovascular events [35].
We found a trend for inverse correlation between 18F-NaF and PPi plasma levels. It is likely that the correlation was not significant due to lack of power. This observation did however lead us to hypothesize about the kinetics of ectopic calcification. As 18F-NaF activity increases [33], PPi rates decrease [4,5]. Nevertheless, it is possible that the decrease in PPi plasma levels alone does not account for ectopic calcification [36].
We also detected a decrease in MMP-2 and MMP-3 plasma levels in PXE patients. By contrast, elevated levels of circulating MMP-2 and MMP-9 reflecting extracellular matrix remodeling have been found in the sera of German PXE patients [37]. The assays were conducted on serum [37] and not on plasma. Serum is obtained after coagulation that results in thrombus formation, causing the release of high amounts of MMP-9 through neutrophil degranulation [38]. In addition, MMP-2 was found to be elevated exclusively in the sera of women in the German PXE cohort [37]. MMP3 degrades fibronectin, proteoglycans, laminin, basal lamina collagen IV, and collagen telopeptides. It enhances MMP-1 collagenolytic activity by enhancing fibrillar collagen hydrolysis [39]. A decrease in MMP3 may favor collagen accumulation, causing fibrosis as a result. Ectopic calcification conceivably leads to fewer MMP-2-and MMP-3-producing cells as in atherosclerosis, whereby calcification often corresponds to areas containing either no cells or dead cells [40].
Taken together, PXE is conceivably a prime example of chronic skin and arterial calcification.

Aortic Stiffness Correlated with 18F-NaF Not 18F-FDG in PXE
In 44 early-onset DM patients, aortic stiffness correlated with 18F-FDG [21]. In the 23 PXE cases studied herein, aortic stiffness correlated with 18F-NaF irrespective of SBP, DBP or both together but not with 18F-FDG. Two previous studies have shown that calcification in the tunica media of PXE patients increases arterial stiffness [41,42]. Arterial stiffness and MVC have been shown to correlate [43,44] and to be independent predictors of cardiovascular morbidity and mortality [45,46]. In the present study, multivariate analysis revealed that the adjusted risk factors for aortic MVC were DBP in all patients, and tobacco use in the subgroup with CS = 0HU. Nicotine can induce osteogenic transdifferentiation in VSMCs [47] resulting in tunica media calcification of the vessel wall [47].
Bartstra et al., evoked PXE as a prime example of accelerated peripheral vascular aging whereby MVC induces CV disease independently of atherosclerosis, inflammation and thrombosis [35]. In the present work, 18F-NaF activity in the vascular system has tended to correlate inversely with PPi levels. PPi is the major calcification inhibitor lacking in PXE patient plasma [5], and loss of a single calcification inhibitor can initiate MVC [4,48,49].

18F-NaF as a Diagnostic and Follow-Up Biomarker in PXE
18F-NaF-PET-CT is able to identify calcification that cannot be detected by CT-resolution alone [50]. It is recognized as a reliable detector for quantifying ectopic calcification [33] and tracking its progression [50].
We have demonstrated herein that 18F-NaF-PET-CT is able to detect early-onset calcification in patients with CS = 0HU. 18F-NaF is therefore a biomarker candidate for the diagnosis and follow-up of cardiovascular disease in PXE.
In the "Treatment of Ectopic Mineralization in Pseudoxanthoma Elasticum (TEMP)" trial, etidronate, a non-nitrogen-containing bisphosphonate and a PPi analog, reduced arterial calcification on CT-Scan but did not lower femoral 18F-NaF activity [52]. In this study, etidronate was administered similarly to treatment of osteoporosis (cyclical 20 mg/kg for two weeks every 12 weeks) [52]. In PXE, calcification is a slow and continuous process [27]. Discontinuous administration of etidronate in PXE is effective on clinically visualized CT calcification [52] but may not be effective at reducing molecular calcification as assessed by 18F-NaF. Since PPi is the main anti-calcifying agent that is lacking in PXE patients [4,5], we can hypothesize that discontinuous administration of PPI or its analogs could promote the restarting of molecular calcifications. This may explain the lack of 18F-NaF decay in the femoral arteries in PXE patients treated discontinuously with etidronate [52]. A clinical trial evaluating continuous versus discontinuous administration of etidronate might answer this question by retaining as primary endpoint 18F-NaF quantification of molecular calcification in the femoral arteries.

Study Limitations
The present study lacked a control group for dual PET-CT imaging since each patient was his/her own control for ethical reasons. Additionally, dual PET-CT imaging was conducted on non-digital PET scanners.

Conclusions
In the present cross-sectional study (using FDG/plasma biomarkers), no link could be established between inflammation and calcification in PXE patients.
PXE would appear to more closely resemble a chronic disease model of ectopic calcification than an inflammatory condition. To assess early ectopic calcification in PXE patients, 18F-NaF-PET-CT may be more relevant than CT imaging. It potentially constitutes a biomarker for disease-modifying anti-calcifying drug assessment in PXE.

Clinical Perspectives for PXE Patients
Should 18F-NaF-PET-CT prove to be an early biomarker of vascular and skin calcification in PXE patients, it may constitute an endpoint when assessing disease-modifying anti-calcifying drugs in PXE. Physicians caring for PXE patients are advised to keep CVR factors under control to minimize arterial stiffness and MVC.
Author Contributions: L.O., principal investigator, contributed to study conception and design, patient care, literature search, data collection and analysis, writing the report, revising the intellectual content and final approval of the version to be published. P.-J.M.: F.L. and O.C. contributed to PET-CT imaging, data collection, data analysis, literature search, revising the intellectual content and final approval of the version to be published. A.J. contributed to data analysis, writing the report, revising the intellectual content and final approval of the version to be published. E.L.P. contributed to data monitoring, statistical analyses, revising the intellectual content and final approval of the version to be published. S.B. and P.J. contributed to the Luminex study, data analysis, revising the intellectual content and final approval of the version to be published. G.K. contributed to the PPi determination, data analysis, revising the intellectual content and final approval of the version to be published. O.M. contributed to data analysis, writing the report, revising the intellectual content and final approval of the version to be published. N.N. contributed to data collection and analysis, revising the intellectual content and final approval of the version to be published. G.L. and L.M. contributed to patient care, data collection and analysis, revising the intellectual content and final approval of the version to be published. Each author has agreed both to be personally accountable for the author's own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature. All authors have read and agreed to the published version of the manuscript.