PATH-51. Identification of molecular markers predicting malignant transformation in low grade gliomas

Abstract Low-grade gliomas generally have an indolent course and good prognosis after maximal safe resection; however, progression to high-grade gliomas via malignant transformation is a poorly understood process. In order to further elucidate the molecular drivers of malignant transformation of low-grade gliomas, we have assembled a cohort of patients for whom we have banked tissue from both the initial resection of the low-grade glioma and the high-grade recurrence. As known clinically relevant markers exist at the RNA, DNA and protein levels, we use a novel spatial multi-omics profiling method known as seqFISH+ which allows us to molecularly profile cells at all of these levels. We have used seqFISH+ to profile the spatial RNA expression of over 1000 genes at single cell resolution in over 1 million cells across 25 different tumors in 16 patients. Based on the RNA expression data, we have validated known and identified novel gene expression signatures associated with MT. Furthermore, as the spatial context of these cells are left intact, we can begin to look at the spatial organization of the tumor and tumor-microenvironment. In particular, we find that a micro-environmental motif that is enriched in endothelial cells is much more prevalent in LGGs that eventually malignantly transformed versus LGGs that did not transform. This suggests that angiogenesis may be an early independent predictor for MT. As tumors with neo-vascularization on H&E are generally characterized as high grade, our findings suggest that seqFISH+ is able to detect very early angiogenesis undetectable by conventional H&E staining and can possibly be used as an early predictor of MT. In conclusion, by using seqFISH+ to spatially profile the expression patterns of genes we find transcriptional and tumor microenvironment factors that predict MT.