Mass spectrometry provides multiple options for the direct biomolecular characterization of tissue to support clinical decision-making and provides significant insight in the development of drugs targeting tumors of the central nervous system. Using an array of mass spectrometry (MS) applications, we rapidly analyze tumor markers ranging from small metabolites to proteins from surgical tissue for rapid diagnosis and surgical guidance. Using similar clinical protocols, we visualize drug and metabolites penetration in brain tumor tissue and correlate with tumor heterogeneity and response to support drug development.

Click here to watch Dr. Agar’s First Look presentation.

The methodologies developed and validated include the direct and quantitative analysis of specific biomarkers, and the application of machine learning algorithms to the complex data to identify diagnostic patterns. More specifically, we quantitatively monitor the distribution of the oncometabolite 2-hydroxuglutarate as a marker of the presence of cancer cells in glioma tumors harboring a mutation in the isocitrate dehydrogenase (IDH) gene and provide the information in real-time to the neurosurgeon to further inform surgical decision making. Using the broader mass spectral information containing hundreds to thousands of peaks, we apply classification workflows to distinguish brain tumor types, subtypes, grades, and assess tumor cell concentration, and continue to investigate different AI tools to facilitate analysis and extract more information. In an effort to increase our specificity in delineating pituitary microadenomas, we combine the analysis of specific pituitary hormones with machine learning. Furthermore, we apply similar approaches to assess tumor margins for breast conserving surgery in the treatment of breast cancer.

While we have made this technology available and continue to further validate it at BWH, true dissemination of the work will occur via commercialization of both clinically compatible mass spectrometry systems, and well validated human reference databases and algorithms allowing the rapid acquisition, analysis, and visualization of diagnostic classification.

For more information about Dr. Agar’s research, please contact Partners HealthCare Innovation by clicking here.

Top. 3D mapping of 2-HG detected by MS over MRI volume

reconstruction.

Bottom. Histopathology scoring of tumor cell concentrations

determined from H&E-stained tissue sections.

Santagata, S. et al (2014) PNAS. 111(30):11121-6. http://www.slicer.org/