Katie Campbell
Katie Campbell, Ph.D. completed her undergraduate degree in Biochemistry at the Pennsylvania State University, graduating with honors from the Schreyer Honors College, and received her PhD in Molecular Cell Biology at Washington University in St. Louis in the laboratory of Dr. Obi Griffith.
Upon the completion of her PhD, Dr. Campbell joined the laboratory of Dr. Antoni Ribas at UCLA as a postdoctoral researcher in 2018. Her early postdoctoral work focused on establishing cloud-based computational pipelines to automate and parallelize the processing of genomics and transcriptomics data derived from clinical melanoma tumor biopsies and patient-derived melanoma models. She has since expanded these approaches to integrate multiplexed spatial profiling data to understand the complex molecular drivers and cellular interactions responsible for immunotherapeutic response in melanoma clinical specimens. Campbell’s current research aims to understand how the somatic alterations in the antigen presentation machinery modulate tumor-T-cell interactions, particularly through copy number alterations that result in imbalance or loss of human leukocyte antigen (HLA) genes. Her interests and approaches collectively enable the comprehensive molecular profiling of tumors, defined by interface of tumor drivers and immunogenicity, to improve immunotherapeutic strategies.
Abstract
Katie Campbell, Cynthia Gonzalez, Egmidio Medina, Philip Scumpia, Antoni Ribas
University of California, Los Angeles, Los Angeles, CA, USA
Sequencing technologies have become standard approaches for high-dimensional analysis of clinical tumor biopsies, leading to the identification of relevant molecular correlates and mechanisms for therapeutic response. However, despite advances in sequencing technologies and bioinformatics approaches, conclusions are often limited by the quality, size, and histologic or pathologic features of the collected samples. Sequential slices of tissues can be used to corroborate findings in bulk sequencing data using multiplexed imaging or single stain studies, but their analysis may still require manual assessment by trained pathologists. Here, we demonstrate the utility of the open-source software, Qupath, for digital pathology to standardize the analysis of sequential slides cut from formalin-fixed, paraffin-embedded melanoma tumor biopsies before and during treatment with immunotherapy, individually stained for hematoxylin and eosin (H&E), melanoma tumor cells (S100), or cytotoxic T cells (CD8). By overlaying consecutive images, the densities of melanoma and CD8+ T cells were calculated in consistent regions across biopsy slices, based upon chromogen intensities, for accurate quantitation of CD8 T-cell infiltration. Furthermore, a dermatopathologist identified additional features that may digitally or manually limit the precision of these measurements, such as the presence of melanophages (pigment-laden macrophages) and pigmented melanocytes. These features are especially critical in assessing immunotherapy-treated biopsies, where antitumor immune reactions complicate pathologic review. This semi-automated approach for pathologic assessment demonstrates a standardized workflow for orthogonal validation of bulk sequencing-based metrics, including tumor cellularity and immune cell deconvolution.