I am a PhD candidate advised by Dr. Young Hwan Chang in the Computational Biology Program and Biomedical Engineering Department at Oregon Health & Science University. My research focuses on the development of reproducible, scalable, and robust computational applications for digital pathology.
Spatial biology and digital pathology today are what clinical genomics was twenty years ago: rapidly evolving technologies with profound implications for precision medicine, but with much still to prove. To maximize their positive impact on both basic and clinical research, we must not only develop best-in-class algorithms for image analysis, but also invent data integration approaches which synergize biomarkers across imaging and molecular assays. Further, we must do it all in a way that scales up to billions of pixels and millions of cells, or more. These concerns have been central to my graduate research.
As a graduate trainee under the auspices of the Knight Cancer Institute and the National Cancer Institute’s Human Tumor Atlas Network and Cancer Systems Biology Consortium, I have helped lead digital pathology and image analysis projects from conception to publication and provisional patent filing. These projects include the development of a mathematical morphology-based epidermis segmentation algorithm; a GPU-boosted, machine learning-based framework which enables normalization, compilation, and phenotyping of millions of single-cell measurements from multiplex imaging data in minutes (see BCTMA paper); and deep learning-based algorithms for optimal sample selection and histological-to-immunofluorescent transformation of whole slide images (see SHIFT paper).
Some of these projects have been undertaken in the context of SMMART, an ongoing clinical trial focused on breast cancer precision oncology, through which I have learned to become an active collaborator with an interdisciplinary team of clinicians, pathologists, and other scientists.
With this skill set, experience, and the ability and desire to learn new things, I am poised to make immediate impact in a scientific role in industry.
Please find my CV here.
PhD in Biomedical Engineering, 2021
Oregon Health & Science University
MSc in Biology, 2017
University of Oregon
BSc in Biochemistry, 2016
University of Oregon
The emergence of megascale single-cell multiplex tissue imaging (MTI) datasets necessitates reproducible, scalable, and robust tools for cell phenotyping and spatial analysis. We developed open-source, graphics processing unit (GPU)-accelerated tools for intensity normalization, phenotyping, and microenvironment characterization. We deploy the toolkit on a human breast cancer (BC) tissue microarray stained by cyclic immunofluorescence and benchmark our cell phenotypes against a published MTI dataset. Finally, we demonstrate an integrative analysis revealing BC subtype-specific features.
Spatially-resolved molecular profiling by immunostaining tissue sections is a key feature in cancer diagnosis, subtyping, and treatment, where it complements routine histopathological evaluation by clarifying tumor phenotypes. In this work, we present a deep learning-based method called speedy histological-to-immunofluorescent translation (SHIFT) which takes histologic images of hematoxylin and eosin (H&E)-stained tissue as input, then in near-real time returns inferred virtual immunofluorescence (IF) images that estimate the underlying distribution of the tumor cell marker pan-cytokeratin (panCK). To build a dataset suitable for learning this task, we developed a serial staining protocol which allows IF and H&E images from the same tissue to be spatially registered. We show that deep learning-extracted morphological feature representations of histological images can guide representative sample selection, which improved SHIFT generalizability in a small but heterogenous set of human pancreatic cancer samples. With validation in larger cohorts, SHIFT could serve as an efficient preliminary, auxiliary, or substitute for panCK IF by delivering virtual panCK IF images for a fraction of the cost and in a fraction of the time required by traditional IF.