第一部分
目的：探讨SAPHO综合征患者腰椎骨质硬化对双能X射线吸收仪（dual-energy X-ray absorptiometry，DXA）测量的骨小梁分数（trabecular bone score，TBS）和腰椎骨密度（bone mineral density，BMD）的影响。分析TBS与BMD在SAPHO综合征患者椎体骨折风险评估中的临床应用价值。
方法：回顾性纳入70例SAPHO综合征患者（平均年龄50.1±10.7岁，57例女性）和70例年龄和性别匹配的正常对照组。采用DXA评估TBS和腰椎、股骨颈和全髋的BMD。使用全脊柱计算机断层扫描（computed tomography，CT）评估骨质硬化和脊柱椎体骨折。根据国际临床骨密度学会（International Society for Clinical Densitometry，ISCD）建立的椎体去除标准对校正后的椎体再次进行TBS和腰椎BMD评估。采用年龄校正后的多因素logistic回归模型评估TBS和BMD每下降一个标准差（standard deviation，SD）的比值比（odds ratio，OR）以及95%置信区间（confidence interval，CI）。通过受试者工作特征（receiver operating characteristic，ROC） 曲线分析TBS和BMD评估SAPHO综合征患者椎体骨折的诊断效能，记录曲线下面积（area under the curve，AUC）。
结果：有骨质硬化的SAPHO综合征患者的TBS显著低于匹配的对照组，而腰椎BMD与对照组相比未见显著的统计学差异。无骨质硬化的患者，TBS和腰椎BMD均显著低于对照组。根据ISCD标准校正椎体后，SAPHO综合征患者腰椎BMD显著低于对照组。经年龄校正后，TBS（OR：12.002，95% CI：3.565-40.410，p=0.000）和腰椎BMD（OR：4.662，95% CI：1.993-10.904，p=0.000）是SAPHO综合征患者椎体骨折的独立相关因素。评估SAPHO综合征患者的椎体骨折时，TBS在ROC曲线上的AUC值最大，为0.920，而腰椎、股骨颈和全髋BMD的AUC值分别为0.777、0.690和0.652。
结论：与腰椎BMD相比，TBS不受SAPHO综合征患者骨质硬化的影响，并且在椎体骨折识别方面优于腰椎BMD，强调了TBS在SAPHO综合征患者椎体骨折评估中对BMD的补充价值，尤其是有骨质硬化的患者。
第二部分
目的： 基于磁共振成像（magnetic resonance imaging，MRI）椎体骨质量（vertebral bone quality，VBQ）评分和椎旁肌肉质量探讨SAPHO综合征患者骨量和肌肉参数及评估椎体骨折的风险。
方法：回顾性纳入80例SAPHO综合征患者（平均年龄42.6±12.3岁，58例女性）和80例年龄和性别匹配的正常对照组。基于腰椎非增强MRI T1加权成像（T1-weighted imaging，T1WI）测量VBQ评分。使用全脊柱计算机断层扫描（computed tomography，CT）评估椎体骨折。使用Image J软件评估L2-3、L3-4、L4-5椎间盘水平椎旁多裂肌（multifidus，MF）面积及脂肪浸润程度（degree of fat infiltration，DFF）。基于VBQ评分三分类法将SAPHO综合征患者VBQ评分分为低、中和高分数组。经年龄校正后，多因素logistic回归分析与SAPHO综合征患者椎体骨折独立相关的因素，记录比值比（odds ratio，OR）以及95%置信区间（confidence interval，CI）。通过受试者工作特征（receiver operating characteristic，ROC）曲线分析相关参数评估SAPHO综合征患者椎体的诊断效能，记录曲线下面积（area under the curve，AUC）。Spearman秩相关分析SAPHO综合征患者VBQ评分与MF面积、MF-DFF之间的相关性。
结果：SAPHO综合征男性和女性患者VBQ评分、L2-3、L3-4、L4-5的 MF-DFF均显著高于对照组。男性患者VBQ评分中-高分数组比例显著高于对照组。SAPHO综合征患者椎体骨折组VBQ评分、VBQ评分中-高分数组比例均显著高于无骨折组，L2-3 MF面积显著低于无骨折组。经年龄校正后，VBQ评分是SAPHO综合征患者椎体骨折的独立危险因素（OR：3.980，95% CI：1.049-15.105，p=0.042）。ROC曲线分析显示VBQ评分预测椎体骨折具有最佳诊断效能（AUC=0.674，p=0.025）。Spearman秩相关分析显示SAPHO综合征患者VBQ评分与L2-3、L3-4、L4-5的MF-DFF呈弱-中度正相关（相关系数分别为0.349、0.309、0.265，均p＜0.05）。
结论：SAPHO综合征男性和女性患者均出现骨量显著减少和腰椎椎旁多裂肌脂肪浸润程度显著增加。SAPHO综合征患者椎体骨折组的骨量和L2-3 MF面积显著减少。VBQ评分为SAPHO综合征患者骨密度评估提供了额外的骨质量特征，并且可作为评估椎体骨折的有效、可靠指标。
第三部分
目的：基于计算机断层扫描（computed tomography，CT）评估SAPHO综合征患者胸椎椎体骨量及相应层面的肌肉面积和密度，并探讨相关参数在评估椎体骨折中的作用。
方法：回顾性纳入170例SAPHO综合征患者（平均年龄45.5±12.8岁，107例女性）和170例年龄和性别匹配的正常对照组。基于胸部CT平扫图像评估所有受试者T8、T10、T12水平的椎体CT值（Hounsfield Unit，HU）和肌肉横截面积（cross-sectional muscle area，CSMA）、骨骼肌衰减（skeletal muscle radio-attenuation，SMRA）。根据椎体CT值将SAPHO综合征患者分为正常骨量、低骨量和骨质疏松组。根据SMRA值将患者分为正常肌肉、肌肉脂肪浸润和重度脂肪浸润组。重建胸部CT矢状位图像用以评估SAPHO综合征患者椎体骨折。将单因素分析中p＜0.10的变量纳入多因素logistic回归分析与椎体骨折独立相关的因素，得出比值比（odds ratio，OR）和95%置信区间（confidence interval，CI）。基于受试者工作特征（receiving operator characteristic，ROC）曲线评估骨量和肌肉参数对预测椎体骨折的诊断效能，获取曲线下面积（area under the curve，AUC）。Pearson相关分析骨量和肌肉参数之间的相关性。
结果： SAPHO综合征男性和女性患者在三个椎体层面（T8、T10、T12）的CSMA、SMRA和椎体CT值均显著低于对照组。SAPHO综合征患者椎体骨折组年龄和绝经后女性比例显著增高，CSMA-T12、SMRA-T12、T8-CT值、T10-CT值和T12-CT值显著减低。多因素logistic回归分析显示CSMA-T12（OR：0.925，95% CI：0.867-0.987，p=0.018）和T12-CT值（OR：0.949，95% CI：0.907-0.992，p=0.021）是SAPHO综合征患者椎体骨折的独立保护因素。ROC曲线分析显示CSMA-T12和T12-CT值参与构建的综合诊断模型对于评估椎体骨折具有最佳诊断效能（AUC=0.834，p=0.000）。Pearson相关分析显示同一水平的椎体CT值与SMRA均呈中度正相关（T8、T10、T12层面的相关系数分别为0.339、0.350和0.418，均p＜0.05）。
结论：SAPHO综合征男性和女性患者均出现骨量减少、肌肉面积减少和肌肉脂肪浸润，相同层面的椎体CT值与肌肉衰减值之间有显著相关性，并且综合骨量和肌肉参数能够更好地评估SAPHO综合征患者椎体骨折。

Part Ⅰ 

Analysis of Trabecular Bone Score to Assess Vertebral Fracture Risk in SAPHO Syndrome Patients

Objective: To investigate the impact of lumbar spine (LS) osteosclerosis on the trabecular bone score (TBS) and bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA) and analyze the clinical utility of TBS evaluation for vertebral fracture (VF) risk assessment in SAPHO syndrome patients compared with BMD assessment.

Methods: Seventy SAPHO syndrome patients (mean age 50.1±10.7 years, 57 women) and 70 age- and sex-matched normal controls were enrolled. TBS and the BMD of LS, femoral neck (FN), and total hip (TH) were assessed using DXA. Osteosclerosis and VF of the spine were evaluated using whole-spine computed tomography (CT). According to the vertebral exclusion criteria established by the International Society for Clinical Densitometry (ISCD), the adjusted vertebral bodies were re-evaluated for TBS and LS BMD. Odds ratio (OR) per standard deviation (SD) decrease in TBS and BMD and 95% confidence interval (CI) were estimated from multivariate logistic regression models with adjustment for age. The diagnostic efficacy of VF in SAPHO syndrome patients by analyzing TBS and BMD using receiver operating characteristic (ROC) curve, and recording the area under the curve (AUC).

Results: In patients with osteosclerosis, TBS, but not LS BMD, was significantly lower than matched controls, and in those without osteosclerosis, both TBS and LS BMD  were significantly lower than controls. After adjustment according to the ISCD criteria, LS BMD was significantly lower in the SAPHO syndrome patients than in the controls. TBS (OR: 12.002, 95% CI: 3.565-40.410, p=0.000) and LS BMD (OR: 4.662, 95% CI: 1.993-10.904, p=0.000) were independent associated factors for VF in SAPHO syndrome patients. For evaluating VF in SAPHO syndrome patients, TBS showed the greatest AUC in the ROC curve, with a value of 0.920 compared with 0.777, 0.690, and 0.652 for LS, FN, and TH BMD, respectively.

Conclusions: TBS seems not to be influenced by osteosclerosis in opposition to LS BMD and has a better discriminatory value than BMD for VF in SAPHO syndrome patients, which highlighted the value of TBS to complement BMD in the assessment of VF in SAPHO syndrome patients, particularly those with osteosclerosis.

Part Ⅱ

Analysis of Vertebral Fracture Risk in SAPHO Syndrome Patients Assessed Based on MRI Vertebral Bone Quality Score and Paraspinal Muscle Quality

Objective: To explore bone and muscle parameters in SAPHO syndrome patients based on magnetic resonance imaging (MRI) vertebral bone quality (VBQ) score and paraspinal muscle quality and evaluate the risk of vertebral fracture (VF).

Methods: Eighty SAPHO syndrome patients (mean age 42.6±12.3 years, 58 women) and 80 age- and sex-matched normal controls were retrospectively enrolled. VBQ score was measured based on non-enhanced MRI T1-weighted imaging (T1WI) of the lumbar spine. VF was assessed using whole-spine computed tomography (CT). The Image J software was used to evaluate the area of multifidus (MF) and degree of fat infiltration (DFF) at the level of the L2-3, L3-4, and L4-5 intervertebral discs. Based on the VBQ score three classification method, SAPHO syndrome patients were divided into low, medium, and high scores. After age adjustment, the multivariate logistic regression analysis was conducted to determine the factors independently related to VF in SAPHO syndrome patients. The odds ratio (OR) and 95% confidence interval (CI) were recorded. The diagnostic efficacy of VF in SAPHO syndrome patients was analyzed using the receiver operating characteristic (ROC) curve and recorded the area under the curve (AUC). Spearman rank correlation analysis assessed the correlation between VBQ score, MF area, and MF-DFF in SAPHO syndrome patients.

Results: The VBQ score, and MF-DFF of L2-3, L3-4, and L4-5 in men and women patients with SAPHO syndrome were significantly higher than those in the controls. The proportion of high-score groups in the VBQ scores of SAPHO men patients was significantly higher than that of the controls. The VBQ score and the proportion of medium to high VBQ scores in the VF group of SAPHO syndrome patients were significantly higher than those in the non-fracture group, and the L2-3 MF area was significantly lower. After age adjustment, the VBQ score was an independent risk factor for VF in SAPHO syndrome patients (OR: 3.980, 95% CI: 1.049-15.105, p=0.042). ROC curve analysis showed that the VBQ score had optimal diagnostic efficacy in predicting VF (AUC=0.674, p=0.025). Spearman rank correlation analysis showed that the VBQ score of SAPHO syndrome patients was weakly to moderately positively correlated with L2-3, L3-4, and L4-5 MF-DFF (correlation coefficients were 0.349, 0.309, and 0.265, respectively, all p<0.05).

Conclusions: Both men and women patients with SAPHO syndrome exhibited significant bone loss and increased degree of lumbar paraspinal MF fat infiltration. The bone mass and L2-3 MF area were significantly reduced in the fracture group of SAPHO syndrome patients. The VBQ scores provide additional bone quality features for assessing bone mineral density in SAPHO syndrome patients and can serve as an effective and reliable indicator for evaluating VF.

Part Ⅲ

Assessment of Bone and Muscle Loss and Related Factors with Vertebral Fracture in SAPHO Syndrome Patients Based on CT

Objective: To evaluate the bone mass and corresponding muscle area and density of thoracic vertebrae in SAPHO syndrome patients based on computed tomography (CT), and to explore the role of relevant parameters in evaluating vertebral fracture (VF).

Methods：One hundred and seventy SAPHO syndrome patients (mean age 45.5±12.8 years, 107 women) and 170 of age- and sex-matched controls were retrospectively enrolled. The Hounsfield Unit (HU), cross-sectional muscle area (CSMA), and skeletal muscle radio-attention (SMRA) of all participants at T8, T10, and T12 vertebral levels based on chest CT plain scan images were evaluated. SAPHO syndrome patients were divided into normal bone mass, osteopenia, and osteoporosis groups based on vertebral CT values. Patients were divided into normal muscle mass, muscle fat infiltration, and severe fat infiltration groups based on SMRA values. The chest CT sagittal images were reconstructed to evaluate VF in SAPHO syndrome patients. Variables with p<0.10 in univariate analysis were included in multivariate logistic regression analysis was conducted to identify factors independently associated with VF, resulting in odds ratio (OR) and 95% confidence interval (CI). Based on the receiver operating characteristic (ROC) curve, the diagnostic efficacy of bone and muscle mass parameters in assessing VF was evaluated, and the area under the curve (AUC). Pearson correlation analysis was used to evaluate the correlation between bone and muscle mass parameters.

Results: Both men and women patients with SAPHO syndrome had significantly lower CSMA, SMRA, and vertebral CT values at three vertebral levels (T8, T10, T12) compared to the controls. The age and the proportion of postmenopausal women in the VF group were significantly higher in SAPHO syndrome patients, and CSMA-T12, SMRA-T12, T8-CT value, T10-CT value, and T12-CT value were significantly decreased. Multivariate logistic regression analysis showed that CSMA-T12 (OR：0.925，95% CI：0.867-0.987，p=0.018) and T12-CT value (OR：0.949，95% CI：0.907-0.992，p=0.021) were independent protective factors for VF in SAPHO syndrome patients. ROC curve analysis showed that the comprehensive diagnostic model constructed with CSMA-T12 and T12-CT value had the best diagnostic performance for evaluating VF (AUC=0.834, p=0.000). Pearson correlation analysis showed moderate positive correlations between CT values of vertebral bodies at the same level and SMRA (correlation coefficients of 0.339, 0.350, and 0.418 at T8, T10, and T12 levels, respectively, all p<0.05).

Conclusions: Both men and women patients with SAPHO syndrome exhibited decreased bone mass and muscle area, as well as muscle fatty infiltration. There was a significant correlation between vertebral CT values and muscle attenuation values at the same level, and comprehensive bone mass and muscle mass parameters can better evaluate VF in SAPHO syndrome patients.