骨质疏松性椎体压缩骨折患者术前MRI影像学参数与经皮椎体成形术后骨水泥渗漏的相关性

Correlation between preoperative MRI imaging indicators and cement leakage after percutaneous vertebroplasty in patients with osteoporotic vertebral compression fractures

中文关键词:  骨质疏松性腰椎压缩性骨折  经皮椎体成形术  骨水泥渗漏  椎基静脉孔  MRI

英文关键词:Osteoporotic lumbar vertebral compression fracture  Percutaneous vertebroplasty  Cement leakage  Basivertebral foramen  MRI

中文摘要:

  【摘要】　目的：探讨骨质疏松性腰椎压缩性骨折（osteoporotic lumbar vertebral compression fracture，OLVCF）患者术前MRI的影像学参数与经皮椎体成形术（percutaneous vertebroplasty，PVP）术后骨水泥渗漏（cement leakage，CL）的相关性。方法：回顾性分析2020年1月~2023年12月在广州中医药大学第一附属医院的行PVP治疗的84例OLVCF患者，根据术后3d腰椎X线和CT评估骨水泥是否渗漏，将患者分为渗漏组与非渗漏组，渗漏组又分为经椎基静脉孔渗漏（B型）和经节段静脉渗漏（S型）亚组。收集患者一般资料包括年龄、性别、BMI、骨密度T值、伤椎节段以及骨水泥注入量，测量并记录术前MRI中椎体骨质量评分（vertebral bone quality，VBQ）、伤椎椎基静脉孔高度（basivertebral foramen height，BFH）和椎基静脉孔深度（basivertebral foramen depth，BFD）相关影像学参数，并记录术中骨水泥注入量。采用t检验、方差分析、卡方检验进行组间比较；运用logistic回归分析CL的影响因素，通过中介模型检验骨水泥注入量的中介效应，并绘制ROC曲线评估预测性能。结果：纳入的OLVCF患者年龄75.8±8.1岁（60~90岁），其中渗漏组53例（B型33例，S型20例），非渗漏组31例。组间基线比较：渗漏组与非渗漏组、渗漏组亚组（B型和S型）与非渗漏组在年龄、BMI、性别、骨密度T值、伤椎位置、骨水泥注入量、VBQ、伤椎BFD及骨折程度均无统计学差异（P＞0.05），渗漏组与非渗漏组、渗漏组亚组（B型和S型）与非渗漏组伤椎BFH存在显著性差异（P＜0.05）。渗漏组与非渗漏组二元logistics回归显示，骨水泥注入量是CL的保护因素（OR=0.677，95%CI：0.514~0.892，P=0.006），伤椎BFH是CL的独立危险因素（OR=1.740，95%CI：1.303~2.324，P＜0.001）；在亚组分析中无序多分类logistics回归表明，以非渗漏组为参考，伤椎BFH是B型（OR=1.906，95%CI：1.387~2.619，P＜0.001）与S型渗漏（OR=1.550，95%CI：1.128~2.131，P=0.007）的共同高独立危险因素，骨水泥注入量则是两类渗漏的保护因素（B型：OR=0.646，95%CI：0.474~0.881，P=0.006；S型：OR=0.720，95%CI：0.523~0.991，P=0.044）。中介效应分析显示，骨水泥注入量在BFH与CL间的中介作用不显著（95%CI包含0）；BFH对B型渗漏的ROC曲线下面积（AUC）为0.749，临界值10.99，提示该指标对B型渗漏的发生具有中等强度的预测效能，对S型渗漏的AUC为0.684，临界值7.55，有着一定的提示作用；骨水泥注入量对B型（AUC=0.347）或S型（AUC=0.395）渗漏均未展现出良好的临床预测价值。结论：伤椎术前BFH是OLVCF患者PVP治疗术后B型与S型CL共同的独立风险因素，且对B型渗漏的预测效能优于S型。

英文摘要:

  【Abstract】　Objectives: To explore the correlation between preoperative MRI-measured imaging parameters and cement leakage(CL) after percutaneous vertebroplasty(PVP) for osteoporotic lumbar vertebral compression fracture(OLVCF). Methods: A total of 84 patients with OLVCF who underwent PVP in the First Affiliated Hospital of Guangzhou University of Chinese Medicine from January 2020 to December 2023 were retrospectively analyzed. Based on bone cement leakage or not according to lumbar X-ray and CT reexamination at postoperative 3d, the patients were divided into a leakage group and a non-leakage group. The leakage group was further subdivided into basivertebral foramen leakage(type B) and segmental venous leakage(type S). General data of the patients were collected, including age, sex, BMI, bone mineral density T-score, injured vertebral level, and bone cement volume. Preoperative MRI parameters including the vertebral bone quality(VBQ), basivertebral foramen height(BFH), and basivertebral foramen depth(BFD) of the injured vertebra were measured and recorded, along with the volume of bone cement injected during surgery. Intergroup comparisons were performed using t-tests, ANOVA, and chi-square tests. Logistic regression was employed to analyze factors influencing CL, a mediation model was used to examine the mediating effect of cement volume, and ROC curves were plotted to evaluate predictive performance. Results: The OLVCF patients enrolled aged 75.8±8.1 years(60-90 years). There were 53 cases in the leakage group(33 cases of type B and 20 cases of type S), and 31 cases in the non-leakage group. Baseline comparisons showed no significant differences in age, BMI, gender, bone mineral density T-score, injured vertebral level, cement volume, VBQ score, BFD, or fracture severity between the leakage and non-leakage groups or between type B and type S subgroups and the non-leakage group(P>0.05), except BFH(P<0.05). Binary logistic regression indicated in the leakage group and non-leakage group that the cement volume was a protective factor for CL(OR=0.677, 95%CI: 0.514-0.892, P=0.006), while BFH was an independent risk factor(OR=1.740, 95%CI: 1.303-2.324, P<0.001). Multinomial logistic regression for subgroup analysis, using the non-leakage group as reference, showed that BFH was a common independent risk factor for both type B(OR=1.906, 95%CI:1.387-2.619, P<0.001) and type S leakage(OR=1.550, 95%CI: 1.128-2.131, P=0.007), while cement volume was a protective factor for both types(type B: OR=0.646, 95%CI: 0.474-0.881, P=0.006; type S: OR=0.720, 95%CI: 0.523-0.991, P=0.044). To further explore the underlying mechanism of BFH and cement volume on leakage, further mediation analysis showed no significant mediating effect of cement volume between BFH and CL(95%CI included zero). ROC analysis indicated that BFH had a moderate predictive value for type B leakage(AUC=0.749, cutoff=10.99) and certain predictive utility for type S leakage(AUC=0.684, cutoff=7.55), whereas the amount of bone cement injected did not demonstrate good clinical predictive value for either type B(AUC=0.347) or type S(AUC=0.395) of leakage. Conclusions: The preoperative BFH of the injured vertebra is a common independent risk factor for both type B and type S CL after PVP for OLVCF, and its predictive efficacy for type B leakage is superior to that for type S.