Introduction: The volume of enhancing tissue on MRI indicates only a fraction of proliferating tumor in glioblastoma (GBM). As the advancing edge of a tumor invades surrounding tissue, other histological changes (e.g. increased cellularity, mitosis, microvascular proliferation, apoptosis) may precede loss of BBB integrity, allowing tumor to spread, undetected, beyond the enhancing margin. In this study, the histological characteristics of tissues in and around malignant gliomas were analyzed, and the power of the enhancing margin to discriminate tumor from surrounding brain parenchyma was quantified.
Methods: 15 stereotactically registered positions were biopsied during 5 glioblastoma resections, yielding a total of 75 stereotactic biopsies. Each biopsy was formalin-fixed, stained with H&E and scored by a neuropathologist for cellularity, necrosis, microvascular proliferation, apoptosis, and percent tumor per sample. Independently, specimens were classified as labeled "core," "margin," or "outside tumor" based on location relative to the enhancing margin. A scoring function was written as a weighted sum of the individual pathology scores, with a threshold for "tumor" determined through calculation of an ROC curve. Weights were determined by optimization of area under the ROC curve, representing the ability of imaging to discriminate between tumor and non-tumor points.
Results: When points in the enhancing margin were categorized as "tumor," optimal weighting coefficients yielded a scoring function with 74.9% discriminatory power (as determined by area under ROC). The optimal scoring function was dominated by percent tumor per core (5.5), presence of apoptosis (4.25), necrosis (1.37), and enhanced cellularity (1.25), with little contribution from mitosis (0.05) or microvascular proliferation (0.1).
Conclusions: Biopsies outside of the enhancing margin tended to display histological characteristics consistent with glioblastoma infiltration: mitosis and microvascular proliferation. Biopsies within the enhancing margin tended to have the histological composition of a more mature tumor mass, displaying apoptosis, necrosis, and enhanced cellularity.
Patient Care: MR imaging is the clinical gold standard for defining the extent of glioblasotma infiltration. Difficulty identifying glioblastoma boundaries with MRI complicates treatment and may contribute to the high likelihood of recurrence after resection. We hope to better define the relationship between tissue histology and MR imaging, thereby improving our understanding of how surgical and non-surgical treatments should be delivered to glioblastoma patients.
Learning Objectives: By the conclusion of this session, participants should be able to 1)better understand regional variability in histology in GBM and to 2) understand how regional stereotactic sampling can be used as a tool to study glioma biology.