Current classification schemes of human tumors depend primarily on histologic appearance which, while of value, is subject to interpretive error and is not biologically based. As a consequence, many tumors elude precise classification, and a portion of cases are subject to interpretive error.

To address this problem, my primary research goal is to develop precise and reliable methods for the biologic classification of cancer using genomic/epigenomic data and bioinformatic techniques.

The adoption of molecular data into diagnostic neuropathology has resulted in an increase in clinically-relevant tumor types and their subsequent incorporation into the WHO Classification of Central Nervous System (CNS) Tumors. This paradigm shift reflects the significant biologic and, consequently, clinical heterogeneity of CNS tumors. However, current WHO nosology still retains traditional histologic classes and integrates them with available genetic information (e.g., mutations, copy number variants). There remains a need to refine and optimize current classification methods.

The goal of the lab is to improve cancer diagnostic precision using a critical diagnostic tool and emerging gold standard for pediatric and adult brain tumor diagnosis – DNA methylation array profiling. Successful implementation will lead to data-driven tumor classification schemes that are less prone to interpretive error.

To achieve this goal, the lab aims to: 1) adopt machine learning methodologies for reliable and accurate tumor class identification, 2) integrate multi-omic data (DNA methylation, RNA-seq, DNA-seq) for increased diagnostic resolution and precision, and 3) extend methylation- and multi-omics-based classification to systemic cancers.