ANGI-12. THE IMPACT OF STANDARD-OF-CARE TREATMENT ON THE RELATIONSHIP BETWEEN CELL DENSITY AND SOX2 POSITIVITY IN GLIOBLASTOMA PATIENTS AT AUTOPSY

Abstract

Abstract INTRODUCTION Sex-determining region Y-box 2 (SOX2) staining highlights pluripotent cells and provides an early sign of tumor cell invasion in glioblastoma patients. This infiltration can often occur at the single cell level and can occur in the absence of lower resolution histological tumor features such as hypercellularity. This study assessed the impact of standard treatment regimens on the relationship between cellularity and SOX2+% at death using autopsy samples as ground truth. METHODS A total of 36 IDH-wildtype glioblastoma patients were included for this study (23M/13F, mean age=61.9). Tissue samples from both suspected tumor invasion and suspected normal tissue were collected for each patient, which were then stained for SOX2, digitized at 40X resolution, and segmented into positive staining nuclei percentage using a color deconvolution algorithm. These samples were then aligned to the MRI using in-house MATLAB-based software, where manually defined control points placed on areas of architectural similarity are used to compute a transform that corrects for nonlinear shrinkage of the tissue sample due to post-mortem processing. The association between SOX2+% and cell density was then computed per subject, and compared between patients who received SOC (radiation and temozolomide) treatment following surgery (as well as additional treatment per case, n=31) and patients who opted for no additional treatment following surgery (n=5). RESULTS Patients who received no additional treatment showed substantially higher correlation between cell density and SOX2+% than patients who had received SOC (t=2.90, p=0.006). No association with overall survival was observed (r=0.17, p=0.329). CONCLUSIONS These results indicate a decoupling between hypercellularity and SOX2+% associated with standard GBM treatments, indicating that radiation and other treatments induce a more heterogenous tumor microenvironment that could lead to single-cell tumor invasion beyond gross histological markers of tumor presence.