Thirteen patients diagnosed with ECD in Peking Union Medical College Hospital (PUMCH，Beijing, China) were enrolled in the study between October 2018 and June 2019. The diagnosis of ECD was based on typical pathological findings in the context of appropriate clinical and radiological manifestations (4). Patients were excluded if they were also diagnosed with autoimmune diseases, osteoporosis, other osteosclerotic diseases, other malignancies, bone fractures of radius or tibia, or have been treated with glucocorticoids in the past one year.
Demographic characteristics and clinical manifestations were documented at the time of enrollment, and HR-pQCT was completed within two weeks. For newly diagnosed patients, HR-pQCT was done before treatment was initiated. Treatment-experienced patients were treated with high-dose interferon-α therapy, defined as 600 or 900MIU three times per week.
Ethics approval was obtained for the study from the institutional board. All participants were informed both in person by TH, and written informed consents were obtained. The study was performed in accordance with the ethical standards of the 1964 Declaration of Helsinki and its later amendments.
Evaluation and definition
Multisystem involvement was confirmed by either clinical symptoms or radiological findings. Cardiovascular system involvements, valve abnormalities or periaortic fibrosis, were evaluated with cardiac MRI or CT, or echocardiograph. CT scans of the chest and abdomen were also applied for assessing retroperitoneal tissue fibrosis (“hairy kidney”) and pulmonary involvement. For central nervous system involvement, we introduced MRI for patients with related symptoms. PET-CT has also been performed for evaluating whole-body multisystem involvement.
High-resolution peripheral quantitative computed tomography (HR-pQCT)
Bone geometry, volumetric bone mineral density (vBMD) and bone microarchitecture were assessed using HR-pQCT (XtremeCT II; Scanco Medical, Zurich, Switzerland) at the non-dominant distal radius and distal tibia. Each measurement was initiated at 9.5mm and 22.5mm from the mid-endplate at the radius and tibia, respectively. 168 parallel slices were obtained in the axial direction, providing images with an isotopic image voxel size of 82μm. The image quality was assessed by an experienced technician using the 5-step scale, and the images with quality worse than grade 3 were excluded from the analysis (16).
The contour between cortex and trabecula was defined automatically using the manufacturer’s standard software with manually assistance by one of the authors (TH). The following measurements were calculated: 1) Bone geometrical measures including total (Tt.Ar), cortical (Ct.Ar) and trabecular (Tb.Ar) areas (mm2), 2) vBMD (mg hydroxyapatite/cm3) of the entire cross section (Tt.vBMD), cortical section (Ct.vBMD) and trabecular section (Tb.vBMD), and 3) microarchitectural parameters including trabecular bone volume to tissue volume fraction (BV/TV), trabecular number (Tb.N, mm-1), trabecular thickness (Tb.Th, mm), trabecular separation (Tb.Sp, mm), standard deviation of 1/Tb.N to represent trabecular network inhomogeneity (Tb.1/N.SD, mm), cortical thickness (Cr.Th, mm) and cortical porosity (Ct.Po, %).
Other bone imaging
All patients had the 18F-Fluorodeoxyglucose (FDG) positron emission tomography
(PET) scan and 99m methylene diphosphonate (99mTc-MDP) bone scintigraphy for a better definition of the range of bone lesions. We also documented X-rays of affected regions as additional evaluations. One patient has also taken MRI of the affected ankle to determine the nature of bone lesions.
Data processing and statistics
All statistical analyses were performed using R (R version 3.6.0, 2019-04-26, 2019 The R Foundation for Statistical Computing Platform) (17). Age-, gender- and site specific distributions were derived from generalized additive models for location, scale, and shape (GAMLSS) with age as the only explanatory variable, based on an ongoing population-based study of HR-pQCT for Mainland Chinese (18, 19) (Supplementary Table 1). For each patient, Z-scores of each measurement were derived from the fitted distribution models, as well as the reference mean of the healthy population. All measurements of HR-pQCT were compared with and adjusted to the corresponding reference normal mean. One sample t-test was performed comparing each relative value to 1, and p value <0.05 was considered to be statistically significant.