目的 探讨多模态功能磁共振成像（functional magnetic resonance imaging，fMRI）联合磁共振波谱（magnetic resonance spectroscopy，MRS）在脑胶质瘤术后复发与假性进展（pseudoprogression，PsP）鉴别诊断中的价值。 方法 选取2021年9月至2024年3月就诊于石家庄平安医院的104例经手术病理证实为脑胶质瘤的患者，并根据修订版脑胶质瘤治疗反应评估标准将脑胶质瘤患者分为两组，70例术后复发的患者分为复发组，34例术后PsP的患者分为PsP组，并选择同期就诊的73例脑胶质瘤患者为模型验证对象。比较两组的一般资料、多模态fMRI参数表观扩散系数（apparent diffusion coefficient，ADC）、标准化脑血容量（cerebral blood volume，CBV）与MRS指标，采用logistic回归分析脑胶质瘤术后复发的影响因素，并采用受试者工作曲线（receiver operator characteristic，ROC）分析多模态fMRI和MRS在诊断脑胶质瘤术后复发与PsP的临床效能。 结果 复发组的ADC水平低于PsP组（P＜0.05）；CBV、胆碱（choline，Cho）/肌酸（creatine，Cr）和Cho/N-乙酰天门冬氨酸（N-acetylaspartic acid，NAA）水平高于PsP组（P＜0.05）；logistic回归分析结果显示ADC、CBV、Cho/Cr和Cho/NAA水平均是脑胶质瘤术后复发的危险因素（P＜0.05）；ROC结果显示多模态fMRI联合MRS在脑胶质瘤术后复发与PsP鉴别诊断中的AUC为0.916，联合后的AUC优于各指标单独诊断。对多模态fMRI联合MRS构建的联合模型进行预测验证，结果显示AUC为0.929，95%CI（0.844 ~ 0.976），敏感度88.46%，特异度91.49%，与前期模型建立时联合预测ROC曲线AUC比较，差异无统计学意义（P > 0.05）。 结论 多模态fMRI联合MRS在脑胶质瘤术后复发与PsP鉴别诊断中具有较高的临床应用价值，值得推广使用。

Objective To investigate the combined value of multimodal functional magnetic resonance imaging （fMRI） and magnetic resonance spectroscopy （MRS） in the differential diagnosis of postoperative recurrence and pseudoprogression （PsP） of glioma. Methods One hundred and four patients with glioma confirmed by surgical pathology were selected from our hospital from September 2021 to March 2024， and the patients were divided into recurrence group （n = 70） and PsP group （n = 34） according to the revised glioma treatment response assessment criteria. Another group of 73 patients with glioma were selected for model validation. The general data， apparent diffusion coefficient （ADC） of multimodal fMRI parameters， standardized cerebral blood volume （CBV） and MRS indexes of the two groups were compared. The influencing factors of postoperative glioma recurrence were analyzed by logistic regression， and clinical efficacy of the combined treatment in the diagnosis of postoperative glioma recurrence and PsP was analyzed by receiver operating curve （ROC）. Results The level of ADC in the recurrence group was lower than that in the PsP group （P < 0.05）， and the levels of CBV， choline （Cho）/creatine （Cr） and Cho/ N-acetylaspartic acid （NAA）were higher （P < 0.05）. Logistic regression analysis showed that the levels of ADC， CBV， Cho/Cr and Cho/NAA were risk factors for postoperative recurrence of glioma （P < 0.05）. ROC showed that the combined use of multimodal fMRI and MRS in the differential diagnosis of postoperative glioma recurrence and PsP had an AUC of 0.916， outperforming the diagnostic accuracy of each individual modality. Validation of the combined model constructed with multimodal fMRI and MRS yielded an AUC of 0.929， 95% CI （0.844 ~ 0.976）， with a sensitivity of 88.46% and specificity of 91.49%. There was no statistical difference when compared to the AUC of the combined predictive model established in the earlier phase. Conclusion Multi-mode fMRI combined with MRS demonstrate high clinical value in the differential diagnosis of postoperative glioma recurrence and PsP， and it is worth to be popularized.