河北医学2023，Vol.29Issue(12)：1967-1973,7.DOI:10.3969/j.issn.1006-6233.2023.12.07
基于MRS IDEAL序列构建非特异性腰痛患者椎旁多裂肌脂肪浸润程度的预测模型及效能评价
Construction of a Predictive Model and Efficacy Evaluation of Paraspinal Multifidus Fat Infiltration in Patients with Non-specific Low Back Pain Based on MRS and IDEAL Sequence
摘要
Abstract
Objective:To assess the relative content of paravertebral multifidus fat in patients with non-specific low back pain,and construct a predictive nomogram model utilizing magnetic resonance spectroscopy imaging(MRS)and iterative decomposition of water and fat with echo asymmetry and least squares estimation(IDEAL)quantitative analysis.Methods:The clinical data of 88 patients with nonspecific low back pain who were treated in our hospital from July 2021 to July 2023 were included.All patients were divided into small group and large group according to Goutallier grading standard.MRS and IDEAL sequences were used to de-tect fat and water content in both groups.The correlation between fat content and the degree of paravertebral multifidus muscle infiltration was analyzed.ROC curves were plotted to analyze the predictive value of fat and water content for the degree of paravertebral multifidus fat infiltration.Fat and water content detected by MRS and IDEAL sequences were included in the nomogram model.R software was used to evaluate the efficacy of the nomogram model using the C-index,ROC curves,and calibration curves.Results:There were no signifi-cant differences in gender,age,and duration between the two groups(P>0.05).Fat(%)detected by MRS and IDEAL sequences in the many group was significantly higher,while Water(%)in the few group was sig-nificantly higher,with statistical significance(P<0.05).Correlation analysis showed a positive correlation between Fat(%)and the degree of paravertebral multifidus fat infiltration,with significant values for MRS(r =0.420,P<0.05)and IDEAL sequences(r=0.477,P<0.05).Water(%)showed a negative correlation,with significant values for MRS(r=-0.469,P<0.05)and IDEAL sequences(r=-0.496,P<0.05).ROC curve results showed that the AUC values for evaluating the degree of paravertebral multifidus fat infiltration in patients with nonspecific low back pain were 0.800 and 0.841 for MRS and IDEAL sequences,respectively.The nomogram model based on Fat(%)and Water(%)detected by MRS and IDEAL sequences had good discrimination,with a C-index of 0.978.ROC curve analysis showed an AUC of 0.978(95%CI:0.948-1.000)for the modeling group and 0.935(95%CI:0.838-1.000)for the validation group,indicating good predictive efficacy.Conclusion:The nomogram model based on Fat(%)and Water(%)detected by MRS and IDEAL sequences provides an intuitive prediction of the degree of paravertebral multifidus fat infiltration in patients with nonspecific low back pain.关键词
磁共振波谱成像/非对称回波的最小二乘估算法迭代水脂分离技术/非特异性腰痛/椎旁多裂肌脂肪浸润Key words
Magnetic resonance spectroscopy imaging/Iterative water-fat separation technology by least square estimation method of asymmetric echo/Nonspecific low back pain/Fat infiltration of paraver-tebral multifidus muscle引用本文复制引用
刘云鹏,曹斌,左玉强,田玉良,陈秋敬,张耀杰..基于MRS IDEAL序列构建非特异性腰痛患者椎旁多裂肌脂肪浸润程度的预测模型及效能评价[J].河北医学,2023,29(12):1967-1973,7.基金项目
河北省卫健委医学科学研究课题计划,(编号:20210870) （编号:20210870）
