摘要目的:探讨手持微型落射荧光宽场显微镜(MEW-M)在中枢神经系统(CNS)病变术中的诊断价值。方法:前瞻性纳入16例2022年9月至2023年4月于中国医学科学院北京协和医学院北京协和医院神经外科行手术切除的CNS病变患者进行研究。均在神经导航引导下对病变边界组织进行活组织检查(简称活检)，术中分别采用MEW-M及冰冻组织病理学(简称冰冻病理)诊断，以术后石蜡包埋组织HE染色病理学(简称石蜡组织病理)诊断为标准，评估MEW-M与冰冻病理在术中鉴别病变组织与正常组织的准确性及所需时长。结果:16例患者共活检52块组织样本，石蜡组织病理显示42块为病变组织，10块为正常组织。使用MEW-M可识别血管等组织结构和细胞核等亚细胞结构的形态。52块样本中，MEW-M观察下有8块病变组织被错误地判断为正常组织(3块为原发性CNS弥漫大B细胞淋巴瘤，样本5块为神经节细胞胶质瘤样本)；而冰冻病理错误判断为3块(均为原发性CNS弥漫大B细胞淋巴瘤样本)。以石蜡组织病理诊断为标准，MEW-M鉴别病变组织的灵敏度为0.81(34/42)，特异度为1.00(10/10)，准确度为0.85(44/52)，受试者工作特征曲线的曲线下面积(AUC)为0.90；术中冰冻病理诊断的灵敏度为0.93(39/42)，特异度为1.00(10/10)，准确度为0.94(49/52)，AUC为0.96。以冰冻病理为标准，MEW-M鉴别CNS病变组织的灵敏度为0.87(34/39)，特异度1.00(13/13)，准确度为0.90(47/52)。41块恶性病变样本，以石蜡组织病理诊断为标准，MEW-M鉴别病变组织的灵敏度为0.91(29/32)，特异度为1.00(9/9)，准确度为0.94(38/42)，AUC为0.95；MEW-M判断结果与冰冻病理结果完全符合。11块良性病变样本，MEW-M判断病变组织的灵敏度为0.50(5/10)，准确度为0.55(6/11)；而冰冻病理结果与石蜡组织病理诊断完全符合。MEW-M诊断每块样本所需的中位时长为5(3～7)min，冰冻病理的中位诊断时长为27(22～34)min，差异有统计学意义( P<0.001)。 结论:初步研究显示，MEW-M鉴别CNS恶性肿瘤组织与正常组织的能力具有非劣于冰冻病理的潜力；但在良性病变判读上仍需要进一步积累经验。应用MEW-M鉴别所需的时长远低于冰冻病理诊断。

abstractsObjective:To evaluate the hand-held effectiveness of the miniaturized epi-fluorescent widefield microscope (MEW-M) system in intra-operative differentiation of lesions from normal brain tissues in the central nervous system (CNS).Methods:Sixteen patients who underwent craniotomy and spinal surgery at the Department of Neurosurgery, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College from September 2022 to April 2023 were prospectively enrolled. All patients underwent tissue biopsy at the lesion boundary under neuronavigation guidance, and MEW-M and frozen tissue pathology (referred to as frozen section pathology) diagnosis were performed during the operation. The diagnosis based on HE staining pathology of postoperative paraffin-embedded tissue sections (referred to as paraffin histopathology) was used as the standard. The accuracy of diagnosis and procedure duration were compared between MEW-M and frozen section pathology.Results:A total of 52 tissues were examined, including 42 abnormal tissues and 10 tissues as diagnoses by HE staining of paraffin sections. The morphology of vessels, cells, nuclei and other subcellular structures were clearly identified using MEW-M. Among 52 tissues, 8 abnormal samples were incorrectly interpreted as normal tissue through MEW-M, including 3 tumor samples of primary central nervous system lymphoma (PCNSL) and 5 samples of ganglioglioma. Three tumor samples of PCNSL were incorrectly interpreted through the frozen section examination. Setting paraffin-embedded section with HE staining examination as standard, the MEW-M had a sensitivity of 0.81 (34/42), specificity of 1.00 (10/10), accuracy of 0.85 (44/52) and an area under the curve (AUC) value of 0.90. The frozen section pathology had a sensitivity of 0.93 (39/42), specificity of 1.00 (10/10), accuracy of 0.94 (49/52) and AUC value of 0.96. The MEW-M exhibited a sensitivity of 0.87 (34/39), specificity of 1.00 (13/13) and accuracy of 0.90 (47/52) when compared with the frozen section pathology. Among malignant tumors, MEW-M exhibited a sensitivity of 0.91 (29/32), specificity of 1.00 (9/9), accuracy of 0.94 (38/42) and AUC value of 0.95 when compared to paraffin-embedded section with HE staining examination. MEW-M showed completely consistent results as the frozen section pathology for samples from malignant tumors. For non-malignant lesions, MEW-M exhibited a sensitivity of 0.50 (5/10) and accuracy of 0.55 (6/11), and the frozen section pathology showed completely consistent results with HE staining examination. The median time required for MEW-M to diagnose each sample was 5 min (range: 3-7 min), the median time required for frozen pathology diagnosis was 27 min (range: 22-34 min), and the difference was statistically significant ( P<0.001). Conclusions:Preliminary research shows that the ability of MEW-M to distinguish CNS malignant tumor tissue from normal tissue has the potential to be non-inferior to frozen section pathology. However, further experience is needed in the interpretation of benign lesions. The time required for identification using MEW-M is much shorter than that of frozen section pathology diagnosis.