Brain invasion in meningiomas: does surgical sampling impact specimen characteristics and histology?
Maximilian TimmeChristian ThomasDorothee Cäcilia SpilleWalter StummerHeinrich EbelChristian EweltFranz-Josef HansUta SchickMaximilian PuchnerUwe WildförsterBernhard BrunsHans Axel TrostMarkus HollingOliver GrauerKatharina HessBenjamin BrokinkelPublished in: Neurosurgical review (2019)
Brain invasion (BI) is a new criterion for atypia in meningiomas and therefore potentially impacts adjuvant treatment. However, it remains unclear whether surgical practice and specimen characteristics influence histopathological analyses and the accuracy of detecting BI. Tumor location, specimen characteristics, and rates of BI were compared in meningioma samples obtained from 2938 surgeries in different neurosurgical departments but diagnosed in a single neuropathological institute. Non-skull base tumor location was associated with CNS tissue on the microscopic slides (OR 1.45; p < .001), increasing specimen weight (OR 1.01; p < .001), and remaining tissue not subjected to neuropathological analyses (OR 2.18; p < .001) but not with BI (OR 1.29; p = .199). Specimen weight, rates of residual tissue not subjected to histopathological analyses, of BI and of brain tissue, on the microscopic slides differed among the neurosurgical centers (p < .001, each). Frequency of BI was increased in one department (OR 2.07; p = .002) and tended to be lower in another (OR .61; p = .088). The same centers displayed the highest and lowest rates of brain tissue in the specimen, respectively (p < .001). Moreover, the correlation of BI with the neurosurgical center was not confirmed when only analyzing specimen with evidence of brain tissue in microscopic analyses (p = .223). Detection of BI was not correlated with the intraoperative use of CUSA in subgroup analyses. Rates of brain invasion in neuropathological analyses are not associated with tumor location but differ among some neurosurgical centers. Evidence raises that surgical nuances impact specimen characteristics and therefore the accuracy of the detection of BI.