Description:
The current state of the art:
Gliomas account for about 27% of all brain tumors and about 80% of the malignant tumors. Currently, Glioma is classified by histology and the gene mutation test (isocitrate dehydrogenase, IDH) defined by the world health organization (WHO).
The problems with the current art:
However, there is a big discrepancy between histology and gene mutation testing; their concordance rate is 82%. Moreover, histology diagnoses depend on human pattern recognition and are prone to errors and bias.
The advantages of our invention:
Scientists at AU developed novel transcriptomic profiling method to classify gliomas. The transcriptomic profiling method includes unsupervised and supervised clustering methods. In a retrospective study with about 1500 patient samples, both methods unbiasedly classify gliomas into four groups which is more accurately reflective of patient survival prognosis. Moreover, the transcriptomic profiling method has better concordance with the WHO-defined method; it is 90.2% and 95.5% concordant with histology and gene mutation testing respectively. Therefore, transcriptomic profiling can be a robust and objective method to classify gliomas.
AURI # 2020-019