This invention is a novel prognostic and predictive diagnostic test that can be used to better risk-stratify patients and to guide therapy for clinically aggressive meningioma. Consisting of a prognostic gene expression signature and a gene signature-based risk score between 0 and 1, it significantly outperforms the current gold standard (the pathology-based WHO grading system) in predicting tumor control, time to treatment failure, and overall survival. The improved performance of this test is due to the application of bioinformatics and machine learning algorithms to a unique UCSF dataset, with subsequent validation in a publicly available dataset. This discovery dataset was significantly enriched for clinically relevant endpoints, allowing for targeted identification of novel gene expression patterns with direct correlation to patient outcomes. In contrast, the validation dataset was more representative of the general population of meningioma patients, thus lending credence to the generalizability of this diagnostic test. While further validation and optimization are still required to bring this invention to market, the core technology has been developed and tested, and several hundred more samples from the dataset are available for continued assessment. As there are currently no other commercial technologies available to supplement pathologic review of meningioma samples, this invention will help address the dire need for more informed therapy decisions through improved risk stratification and recommendations for adjuvant radiotherapy to patients with meningioma.
To commercialize the technology for public benefit