18F-Sodium fluoride PET-CT visualizes disease activity in chronic nonbacterial osteitis in adults

Abstract Chronic nonbacterial osteitis (CNO) is a rare disease spectrum, which lacks biomarkers for disease activity. Sodium fluoride-18 positron emission tomography/computed tomography ([18F]NaF-PET/CT) is a sensitive imaging tool for bone diseases and yields quantitative data on bone turnover. We evaluated the capacities of [18F]NaF-PET/CT to provide structural and functional assessment in adult CNO. A coss-sectional study was performed including 43 adult patients with CNO and 16 controls (patients referred for suspected, but not diagnosed with CNO) who underwent [18F]NaF-PET/CT at our expert clinic. Structural features were compared between patients and controls, and maximal standardized uptake values (SUVmax [g/mL]) were calculated for bone lesions, soft tissue/joint lesions, and reference bone. SUVmax was correlated with clinical disease activity in patients. Structural assessment revealed manubrial and costal sclerosis/hyperostosis and calcification of the costoclavicular ligament as typical features associated with CNO. SUVmax of CNO lesions was higher compared with in-patient reference bone (mean paired difference: 11.4; 95% CI: 9.4–13.5; p < .001) and controls (mean difference: 12.4; 95%CI: 9.1–15.8; p < .001). The highest SUVmax values were found in soft tissue and joint areas such as the costoclavicular ligament and manubriosternal joint, and these correlated with erythrocyte sedimentation rate in patients (correlation coefficient: 0.546; p < .002). Our data suggest that [18F]NaF-PET/CT is a promising imaging tool for adult CNO, allowing for detailed structural evaluation of its typical bone, soft-tissue, and joint features. At the same time, [18F]NaF-PET/CT yields quantitative bone remodeling data that represent the pathologically increased bone turnover and the process of new bone formation. Further studies should investigate the application of quantified [18F]NaF uptake as a novel biomarker for disease activity in CNO, and its utility to steer clinical decision making.


Introduction
Chronic nonbacterial osteitis (CNO) is an inflammatory bone disease affecting children and adults.The CNO spectrum is heterogeneous, ranging from cases of isolated sterile bone inflammation to those with varying additional (extraskeletal) features such as arthritis, psoriasis, acne, and pustulosis palmoplantaris.In absence of aligned terminology, adult patients with CNO may currently be referred to as synovitis, acne, pustulosis, hyperostosis, osteitis (SAPHO) syndrome as well. 1,2In adult CNO, bone inflammation mainly occurs in the anterior chest wall (involved in 89%), followed by the spine, mandible, and incidentally, the peripheral skeleton. 3Typical structural imaging features include sclerosis, hyperostosis, ossification of soft tissue, ankylosis, and degeneration of the adjacent joints, which all gradually accumulate over time. 4lthough the CNO disease course is generally chronic, inflammatory disease activity can vary over time and may be relapsing and remitting as well. 1 The evaluation of disease activity, however, is highly challenging currently, as biochemical markers of inflammation are only increased in a minority of patients.With regard to imaging biomarkers, bone scintigraphy using technetium-99m radio-labeled hydroxymethylene diphosphonate ([ 99m Tc]Tc-HDP) visualizes the increased bone turnover resulting from active inflammation, but continues to demonstrate increased uptake despite clinical remission, suggesting an imprinting pattern. 5With regard to magnetic resonance imaging (MRI), the key CNO features of sclerosis and hyperostosis hamper the proper detection of bone marrow edema on T2-weighted images with fat suppression and short tau inversion recovery (STIR)-weighted images, and similarly prevents the assessment of areas of restricted diffusion on diffusion-weighted imaging. 6,7odium fluoride-18 positron emission tomography/computed tomography ([ 18 F]NaF-PET/CT) is a widely used tool in metabolic bone disease, [8][9][10] and poses distinct advantages compared with other nuclear imaging techniques.0][11] Compared with a combination of [ 99m Tc]Tc-HDP bone scintigraphy with single positron emission computed tomography (SPECT/CT), it offers superior spatial resolution for anatomical orientation.In addition, [ 18 F]NaF-PET/CT readily yields quantitative bone turnover signals that have shown to strongly correlate with clinical disease activity in other metabolic bone diseases, qualifying the imaging tool as a disease-monitoring instrument. 9his study therefore aimed to evaluate the application of [ 18 F]NaF-PET/CT in adult CNO, both in providing structural assessments of CNO lesions, as well as functional evaluation via quantitative bone turnover parameters.

Materials and methods
This study was approved by the medical-ethical review board associated with the Leiden University Medical Center.Written consent was obtained from all individuals for the pseudoanonymized use of their data.The study conduct adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. 12

Study population and design
A cross-sectional study was conducted at the Center for Bone Quality of the Leiden University Medical Center, a tertiary referral and expert center for CNO.Adult patients with suspected CNO referred between 2019 and 2022 were included in the study as, from this time, [ 18 F]NaF-PET/CT was performed as part of standard care at presentation or during flare-ups.In absence of validated classification criteria for adult CNO, eventual diagnosis was expertbased on a combined clinical and imaging assessment of an experienced endocrinologist, radiologist, and nuclear imaging physician.Features regarded as indicative of CNO were chronic or relapsing-remitting inflammatory bone pain and sclerosis and hyperostosis on imaging, isolated or accompanied by extraosseous features, such as synovitis, skin manifestations, or other auto-inflammatory comorbidities.Importantly, quantified uptake on [ 18 F]NaF-PET/CT did not affect diagnosis as these data were retrospectively obtained for the present study.Controls were individuals presenting with CNO-like complaints (mostly pain in the anterior chest wall) but without sufficient features for diagnosis or in whom another diagnosis was made.Individuals were excluded in case of an indefinite diagnosis, if [ 18 F]NaF-PET/CT data were incomplete, or whenever [ 18 F]NaF-PET/CT was performed after the initiation of therapy.

[ 18 F]NaF-PET/CT acquisition
Whole-body [ 18 F]NaF-PET/CT was performed on a 5-ring Discovery MI PET/CT (GE Healthcare, Chicago, IL, USA).Data acquisition was performed approximately 30 minutes after intravenous Na 18 F administration of approximately 1 MBq/kg body weight.An emission scan was obtained using multiple bed positions with 50% overlap between bed positions and 40 seconds per bed position.Time-of-flight PET data were reconstructed using a 256 × 256 matrix, the point spread function and CT-based attenuation correction (120 kV, smart mA modulations with a noise index of 49.5 and a tube current ranging from 15 to 550 mA, 0.5-s rotation time).

Analysis of structural imaging features on [ 18 F]NaF-PET/CT
All [ 18 F]NaF-PET/CT scans were assessed by 1 experienced nuclear physician (F.S.) and radiologist (A.N.C.) in a Picture Archiving and Communication System (PACS) following a systematic reporting format.Upon disagreement, eventual assessment was consensus-based.For each possible lesion site (clavicles, costae, manubrium, sternoclavicular joints, costoclavicular ligaments, costosternal joints, costochondral transitions, mandible, and spine), the degrees of sclerosis, hyperostosis, erosions, ankylosis, and [ 18 F]NaF uptake were scored as appropriate (see table S1 for full systematic reporting format, including detailed definitions of abnormalities).

Quantitative analysis of [ 18 F]NaF-PET/CT
Volumes of interest (VOIs) to generate quantitative parameters of isotope uptake were determined on European Association of Nuclear Medicine Research GmbH (EARL)-compliant reconstructed PET images fused with low-dose CT.The VOIs were drawn using CT images, without blinding for patient or control status but with blinding for degree of uptake (fused PET images were not enabled during the drawing of VOIs).In patients, up to 2 areas of abnormal sclerosis in bone were selected to represent a "bone lesion."Also, up to 2 areas with soft tissue abnormalities in the anterior chest wall (ossification, ankylosis, degeneration; possibly in the sternoclavicular joints, costosternal joints, costochondral transitions, costoclavicular ligaments, or manubriosternal joint) were selected to represent "soft tissue/joint lesions."If more than 2 diseased bone/soft tissue/joint areas were available, the 2 most pronounced areas were selected.In controls, 3 VOIs were drawn in the proximal clavicles and the manubrium sterni to represent control "bone lesions," and 5 in the ipsilateral sternoclavicular joint, first costosternal joint, first costochondral transition, costoclavicular ligament, and the manubriosternal joint to represent control "soft tissue/joint lesions."For both patients and controls, 2 VOIs were drawn in reference bones (ref-bone): the fifth vertebral body (Th5) and the proximal humerus, except in the case of confounding pathology in 1 of these locations (n = 3 and n = 1, respectively).From all VOIs, the maximum standardized uptake value (SUV max ) in g/mL was calculated, which is the decay-corrected activity concentration of the radiopharmaceutical (in Bq/mL) standardized for administered activity (in Bq) and subject body weight (in g).This parameter was selected for its reproducibility and widespread use in clinical practice.Absolute mean SUV max of bone and soft tissue/joint lesions plus the ratio between bone lesions/ref-bone were compared between patients and controls (overall and stratified per anatomical location).

Clinical and biochemical data
The following ancillary data were obtained from the electronic health records for patients and controls: age, gender, height and weight, auto-inflammatory comorbidity, and current medication use.Clinical disease status was evaluated by a visual analog scale (VAS) for pain (VAS 0-10, with 0 representing no pain and 10 representing the worst imaginable pain) on the Brief Pain Inventory 13

Statistical analysis
SPSS Statistics version 25 (IBM Corporation, Armonk, NY, USA) was used for statistical analyses.Data are presented as frequency of counts and percentages for categorical variables and mean ± SD or median (IQR) for continuous variables.Clinical and radiologic characteristics and SUV max of bone and soft tissue/joint lesions were compared between patients and controls using chi-square tests (or Fisher's exact test in case of ≥1 cells with expected frequencies <5) for proportions and categorical data, t test or 1-way ANOVA for parametric numerical data, and Mann-Whitney tests for nonparametric numerical data.The discriminatory capacity of individual radiologic features was evaluated with binary logistic regression; Haldane-Anscombe correction was applied in zero-cell cases. 14The number of bone and soft tissue/joint lesions and their SUV max was correlated with clinical disease activity (biochemical parameters and VAS score for pain), age, and disease duration with Pearson's correlation coefficient or Spearman's rank test, as appropriate, in patients.Diagnostic capacities of mean SUV max of bone and soft tissue/joint lesions were evaluated by multivariate logistic regression, adjusted for symptom duration.Cutoffs were optimized for specificity so as to minimize false-positives.The level of significance was set at p < .05.

Results
Eighty-nine individuals underwent [ 18 F]NaF-PET/CT scanning between 2019 and 2022, of whom 43 patients with CNO and 16 controls could be included in the present study (Figure 1).

Clinical characteristics of patients with CNO and controls
The clinical characteristics of patients and controls included for qualitative and quantitative analysis of [ 18 F]NaF-PET/CT images are displayed in Table 1.Gender, body mass index, comorbidity, [ 18 F]NaF incubation time, clinical symptoms, and medication use at the moment of scanning were all similar between patients and controls.Patients were slightly younger than controls (age: 45 ± 10 vs 49 ± 18 years).Controls were diagnosed with osteoarthritis (n = 7, 44%), functional complaints (n = 7, 44%), heterotrophic calcifications of the costal cartilage (n = 1, 6%), and a subchondral cyst (n = 1, 6%).

Structural evaluation of CNO on [ 18 F]NaF-PET/CT
Structural imaging characteristics as scored via standardized reporting formats for bone are depicted in Table 2.In patients, sclerosis and hyperostosis were frequently found in costae 1 and 2 (79% for both), manubrium (77% and 70%), and to a slightly lesser extent in the clavicle (70% and 33%).Bilateral involvement was most common for the costae (67% displaying sclerosis/hyperostosis on both sides).In controls, radiologic abnormalities were also detected, but bilateral sclerosis of the clavicle and hyperostosis of the clavicle or manubrium were exclusively seen in patients with CNO.For both patients and controls, increased [ 18 F]NaF uptake was generally found in the presence of radiologic abnormalities, with the exception of costae 1 and 2 for controls; here, increased uptake was also seen without signs of osteitis (56% increased uptake, only 6% radiologic abnormalities).Bone features that best differentiated patients and controls were sclerosis, hyperostosis, and increased uptake of the manubrium, and also sclerosis and hyperostosis of costae 1 and 2 (but not increased uptake) (p value for odds ratio [OR] = .004for manubrial hyperostosis, p value for OR < .001for others).Also (tending towards) differential were bilateral clavicular sclerosis (OR, 20; 95% CI, 0.9-424; p = .058)and bilateral increased uptake (OR, 10; 95% CI, 1-81; p = .034).In general, spinal features were only seen in few patients (19% with sclerosis/hyperostosis compared to 6% in controls).
Soft tissue/joint tissues (eg, joints, ligaments) were also evaluated for radiologic features (see Table 3).Abnormalities found in the sternoclavicular joints and manubriosternal joint did not discriminate between patients and controls.In contrast, calcification/beginning ankylosis with increased uptake of the costoclavicular ligament was almost exclusively seen in patients (58% vs 6% and 54% vs 0%; OR [95% CI], 21 [3-172] and 48 [2-1022], respectively).This was similar for radiologic involvement of the costosternal joints (77% in patients vs 6% in controls; OR, 57; 95% CI, 7-488; p < .001),but not for uptake at this location, which only tended to be higher in patients than in controls.
SUV max stratified per lesion location is displayed in Figure 3. SUV max was higher in patients at all comparable locations (bone and soft tissue/joint).This also applied to locations for which a qualitative description of [ 18 F]NaF uptake (Table 2) could not differentiate between patients and controls (ie, the clavicle, sternoclavicular joint, manubriosternal joint, costosternal joints, and costochondral transitions).Considering bones in patients with CNO, the clavicle, manubrium, and costa 1/2 yielded similar SUV max (mean, 22 ± 7.1, 20.8 ± 6.6, and 22.5 ± 6.6 g/mL, respectively).Spinal locations of CNO showed somewhat higher mean SUV max of 26.1 ± 10.0 g/mL (p = .07as compared to clavicle/manubrium/costae).With regard to joint and soft tissue involvement in CNO, the costoclavicular ligament proved most active, with a mean of 29.2 ± 13.0 g/mL, also exceeding the mean SUV max of all bones.

Discussion
Chronic nonbacterial osteitis is an impactful rheumatic disorder characterized by chronic bone inflammation and long-term hyperostotic transformation.Imaging takes on a pivotal role in diagnosis and disease activity assessment, because clinical and biochemical markers may be nonspecific. 3This study evaluated the capacities of [ 18 F]NaF-PET/CT in the structural and functional assessment of adult CNO.We first demonstrated that [ 18 F]NaF-PET/CT produces high-quality images that well display the structural Table 5.Odds ratios adjusted for symptom duration, cutoff optimized for specificity, and sensitivity and specificity for SUV max, (mean and highest, bone and soft tissue/joint).
hallmarks of CNO, as identified in previous studies.We found that features associated with adult CNO included sclerosis and hyperostosis of the manubrium and costae 1-2 and calcification/ankylosis of the costoclavicular ligament.][17][18][19][20][21] In addition to providing a thorough structural assessment, [ 18 F]NaF-PET/CT offers technical and logistical benefits compared with alternative imaging methods.Like [ 99m Tc]Tc-HDP SPECT/CT, [ 18 F]NaF-PET/CT conveniently combines CT and nuclear imaging in 1 procedure, but [ 18 F]NaF-PET/CT has a shorter incubation and scanning time, higher spatial resolution, and significantly lower radiation exposure due to its pharmacokinetics (total exposure estimated at 209 MBq compared to 672 MBq for [ 99m Tc]Tc-HDP SPECT/CT). 10,11,22Compared with MRI, the CT component of [ 18 F]NaF-PET/CT offers a superior structural evaluation, especially for the anterior chest wall, which is the primary localization of adult CNO, involved in 89% of patients. 3In this region, CT yields higher spatial resolution, being less compromised by motion and breathing artefacts.4][25] ) Indeed, a recent study performed both MRI and CT of the anterior chest wall in adult patients with CNO, where CT detected a greater number of structural lesions. 26Altogether, we argue that [ 18 F]NaF-PET/CT is preferred over [ 99m Tc]Tc-HDP SPECT/CT and MRI to provide structural characterization of adult CNO, especially in the anterior chest wall, being a technically easy, fast, and widely available tool that produces high-quality images of the relevant regions.
In addition to structural assessment, [ 18 F]NaF-PET/CT offers the advantage of simultaneously providing functional assessment in a single scan, readily yielding quantitative parameters of bone turnover (see Figures 4 and 5 for illustrative imaging slides).As Na 18 F precipitates in young osteoid, the degree of uptake reflects the degree of bone turnover, which is commonly increased as a result of inflammation in diseases like CNO.Indeed, we found markedly higher SUV max values in all evaluated CNO bone and soft tissue/joint lesions as compared with in-patient reference bone and with controls, suggesting that high [ 18 F]NaF uptake  The highest absolute SUV max values were found in soft tissue/joint lesions, specifically in the costoclavicular ligament (mean, 29.2 ± 13.0 in patients).This finding is of interest, given that these ligaments are known to undergo a calcification process over time, potentially resulting in complete ankylosis in long-term CNO. 20Our findings therefore suggest that quantified [ 18 F]NaF uptake may not only represent bone turnover but also the process of new bone formation.This property of [ 18 F]NaF uptake has been discovered in other diseases characterized by new bone formation as well, such as fibrodysplasia ossificans progressiva 27 , axial spondylarthritis, [28][29][30] and psoriatic arthritis. 31ur data also add to the ongoing discussion on the primary disease process in CNO and its chronology.It is generally recognized that calcifications of ligaments and joint capsules are prominent features of long-term CNO 17,32 , which corresponds to the idea that CNO starts with osteitis and affects adjacent joints and ligaments only secondarily. 20,33ontrarily, a landmark radiographic study proposed that the ossification of the costoclavicular ligament isan early phase of the disease 34 , suggesting that the primary pathology in CNO is inflammatory enthesitis rather than osteitis.This idea was confirmed in a recent study in which most inflammatory signs were found at the insertions of the ligaments and joint capsules. 32While our cross-sectional data cannot reflect the progression of changes during CNO's disease course, it remains striking that soft tissue/joint lesions demonstrate the highest SUV max values, and that the SUV max in soft tissue/joint lesions positively correlates with ESR.In that sense, our results modestly support the hypothesis of inflammatory enthesitis being a primary, if not key, disease process in CNO.
The clinical impact of our findings is substantial.First, [ 18 F]NaF-PET/CT bone turnover parameters form an important advancement in disease activity assessment in adult CNO, which is currently highly difficult as clinical and biochemical measures are nonspecific. 3[ 18 F]NaF-PET/CT provides radiologic biomarkers allowing for accurate and objective disease monitoring, which can support clinicians and patients in understanding the disease course and making better-informed therapeutic decisions.Second, [ 18 F]NaF-PET/CT visualizes the early stages of new bone formation in CNO, an irreversible and detrimental disease process associated with poor outcomes due to mechanical complications.Early detection of this process enables proactive and potentially more aggressive treatment to prevent permanent skeletal damage, which is highly relevant to long-term disease management. 21,34ur study has several limitations.First, diagnosis of CNO was based on the assessment of a multidisciplinary expert team.While there are published criteria for CNO available, these were devised for pediatric patients 35,36 or pertain to the broad entity of SAPHO syndrome, which can include also patients without the core disease characteristic of osteitis. 3,37,38In addition, none of the criteria sets have been externally validated.We therefore use expert-based diagnosis until better criteria are available.Even so, all of our patients also retrospectively fulfilled the Jansson criteria for CNO, allowing for comparison with other cohorts.
Second, the diagnosis of CNO-whether expert-based or criteria-based-relies in large part on radiologic features, specifically hyperostosis.Therefore, it is not surprising that hyperostosis appears associated with CNO, as it likely prompts diagnosis in the first place.However, our data additionally inform about other, less conventional features that we identify to be associated with CNO, including SUV max values.These were shown to be associated with CNO but could not have guided diagnosis as they were generated retrospectively for the purpose of this study and are not necessarily correlated with steering features like hyperostosis. 21Third, the sample size of the control group was small, limiting the power of our analyses and inducing substantial imprecision as reflected in the wide CIs for, particularly, the ORs of radiologic features.Due to the limited sample size, these radiologic features could also not meaningfully be adjusted for symptom duration, which would have been of interest as well.Also, our qualitative assessments were made by 1 radiologist and 1 nuclear medicine physician.However, both have many years of experience, also with CNO specifically, and we used a systematic reporting format to ensure fair reproducibility.A strength of this study is the choice of SUV max as an imaging parameter, which is recognized for its reproducibility between observers and stability across scanners and time, giving our results marked clinical outreach. 39,40n sum, [ 18 F]NaF-PET/CT is a promising imaging tool for adult CNO, especially for the anterior chest wall.It allows for detailed structural evaluation of bone and soft tissue/joint lesions and functional evaluation by yielding quantitative data representing bone turnover and new bone formation.Further studies should investigate the application and clinical utility of quantified [ 18 F]NaF uptake as a novel biomarker for disease activity in CNO.

Figure 3 .
Figure 3. Maximum standardized uptake value (SUV max ) in g/mL per disease location in patients with CNO and controls.Costa 1/2 and spine were not evaluated for controls.Values are means ± SD. * * * p < .01.Abbreviation: CNO, chronic nonbacterial osteitis; SUV max , standardized maximum uptake value.

Table 1 .
Clinical characteristics of patients with CNO and controls.

Table 2 .
Imaging characteristics of bone in patients with CNO and controls on [ 18 F]NaF-PET/CT.

Table 3 .
Imaging characteristics of soft tissue and joints in patients with CNO and controls on [ 18 F]NaF-PET/CT.

Table 4 .
Correlations between number of active lesions, their SUV max , and biochemical and clinical disease activity parameters in patients with CNO.