1544. Diagnostic and Therapeutic Values of Intraoperative Electrophysiological Neuromonitoring for Resection of Spinal Tumors: Meta-Analysis and Systematic Review Encompassing 26 articles and 1563 patients
Authors: Wataru Ishida, MD; Alexander Perdomo-Pantoja, MD; Risheng Xu; Seba Ramhmdani, MD; Ethan Cottrill, BS; YuanXuan Xia, BA; Ali Bydon, MD; Timothy Witham, MD; Sheng-fu Lo, MD (Baltimore, MD)
Introduction: While several guidelines support the use of intraoperative neuromonitoring (IONM) for spinal surgeries with class A recommendations, clinical evidence on its diagnostic and therapeutic value particularly in spinal tumor (ST) resection is surprisingly sparse and not well-summarized. Here we aimed to conduct a meta-analysis and systematic review regarding this controversial topic. Methods: A web-based literature search (1970-2018) on Pubmed in compliance with the PRISMA guidelines was performed, utilizing the keywords, "spinal tumor intraoperative monitoring", which yielded 499 articles. Inclusion criteria were: English-language clinical articles 1) reporting clinical outcomes of ST resection with the use of IONM and 2) documenting information necessary to calculate the diagnostic values of IONM to predict neurological deficits. Exclusion criteria were 1) non-human research and 2) studies including < 10 patients. Consequently, 26 articles encompassing 1563 patients were included, and the pooled diagnostic values of IONM were calculated. Results: There were 10 articles reporting intramedullary-ST (IMST), six articles describing intradural-extramedullary ST (ID-EM-ST), and four articles including extradural-ST (EDST), whereas six articles included miscellaneous STs, all of them being retrospective. Thresholds for MEP ranged from 50% declines to 100% declines. Only eight articles (30.8%) assessed the therapeutic value of IONM. The pooled diagnostic value of IONM was a sensitivity of 80.8%, specificity of 88.9%, PPV of 65.0%, NPV of 95.0%. PPV was lowest in EDST (EDST:31.6%, ID-EM-ST:57.9%, IMST:71.5%, p=0.15), whereas NPV was lowest in IMST (EDST:96.6% versus ID-EM-ST:96.0% versus IMST:92.8%). The overall diagnostic accuracy (area under the curve) was 0.893. Conclusion: The substantial heterogeneity of the 26 articles with regards to types of STs, thresholds for IONM, and the quality of clinical evidence was identified. IONM for ST resection was deemed an excellent diagnostic modality, while its therapeutic value remained to be investigated. IONM should be carefully interpreted in EDST due to its lowest PPV.