Background: Maximal safe resection remains a central determinant of outcomes in glioma surgery, yet intraoperative discrimination between tumor and normal brain tissue is limited by the speed and subjectivity of frozen-section analysis. Label-free techniques such as Raman spectroscopy, mass spectrometry (MS), and optical coherence tomography (OCT) offer real-time biochemical and structural characterization that may enhance surgical precision. Their comparative diagnostic accuracy across clinically relevant endpoints has not been comprehensively evaluated. Methods: Following PRISMA 2020 guidelines, a systematic review and quantitative meta-analysis were conducted using PubMed, Embase, Scopus, and Web of Science through December 2024. Original human studies evaluating Raman, MS, or OCT for intraoperative glioma margin assessment were included. Pooled sensitivity, specificity, and diagnostic odds ratios (DORs) were calculated using a random-effects model. Subgroup analyses addressed tumor versus normal brain tissue, infiltrated versus non-infiltrated margins, and IDH-mutant versus wild-type gliomas. Results: Twenty-four studies comprising 1768 patients met the inclusion criteria. Across all modalities, pooled sensitivity and specificity were 0.89 (95% CI 0.86–0.92) and 0.88 (95% CI 0.84–0.91), with a pooled DOR of 65.7 (95% CI 42.3–101.8; logDOR 4.18), indicating high overall discriminative performance. Tumor versus normal differentiation achieved DOR 72.4 (logDOR 4.28; I2= 26%), infiltrated margin detection DOR 41.8 (logDOR 3.73; I2= 41%), and IDH classification DOR 52.3 (logDOR 3.96; I2= 29%). No publication bias was observed. Raman and MS outperformed OCT. Conclusions: Raman spectroscopy, mass spectrometry, and OCT demonstrate strong diagnostic accuracy for real-time intraoperative glioma evaluation, enabling reliable tissue differentiation and molecular profiling that may enhance resection extent and support precision, molecularly informed neurosurgery.
背景：最大安全切除仍是决定胶质瘤手术预后的核心因素，但术中肿瘤与正常脑组织的鉴别受限于冰冻切片分析的速度和主观性。拉曼光谱、质谱（MS）和光学相干断层扫描（OCT）等无标记技术能提供实时生化和结构特征分析，有望提升手术精准度。但这些技术在临床相关终点上的诊断准确性尚未得到全面评估。方法：遵循PRISMA 2020指南，系统检索截至2024年12月的PubMed、Embase、Scopus和Web of Science数据库，对拉曼光谱、质谱或OCT用于术中胶质瘤切缘评估的原创性人类研究进行系统综述和定量荟萃分析。采用随机效应模型计算汇总敏感性、特异性及诊断比值比（DOR）。亚组分析涉及肿瘤与正常脑组织鉴别、浸润与非浸润切缘区分以及IDH突变型与野生型胶质瘤分类。结果：共纳入24项研究，涵盖1768例患者。所有技术汇总的敏感性和特异性分别为0.89（95% CI 0.86–0.92）和0.88（95% CI 0.84–0.91），汇总DOR为65.7（95% CI 42.3–101.8；logDOR 4.18），表明整体区分效能较高。肿瘤与正常组织的鉴别DOR为72.4（logDOR 4.28；I²=26%），浸润切缘检测DOR为41.8（logDOR 3.73；I²=41%），IDH分型DOR为52.3（logDOR 3.96；I²=29%）。未观察到发表偏倚。拉曼光谱和质谱表现优于OCT。结论：拉曼光谱、质谱和OCT在术中实时胶质瘤评估中表现出强大的诊断准确性，能够实现可靠的组织区分和分子特征分析，有望提升切除范围并支持精准化、分子信息引导的神经外科手术。
肿瘤(癌症)患者之家