Back

Biography

James R Mansfi eld is a scientist with over 25 years of experience in spectral imaging, in-vivo spectroscopy and applied data analysis, directed towards fi nding of novel optical methods for the diagnosis and monitoring of medical conditions. He is currently the Director of Quantitative Pathology Applications at PerkinElmer where he is the senior application scientist for their multispectral and digital pathology product lines, which are being used in a wide range of fl uorescence and brightfi eld microscopy applications. Before PerkinElmer he worked at Cambridge Research & Instrumentation, where he helped develop their multispectral imaging systems. Prior to that he worked at the National Research Council of Canada as a research scientist and at several small companies developing non-invasive spectroscopic methodologies. His research has included projects ranging from the objective classifi cation of skin cancer spectra using mid-infrared spectral imaging, to developing methods for the non-invasive determination of the severity of rheumatoid arthritis, to the development of the fi rst of several spectral imaging systems able to map out skin oxygenation levels. He is an Associate Editor of the American Journal of Nuclear Medicine and Molecular Imaging, holds 6 patents, has over 50 publications and has served as an invited speaker, session chair and organizer at a variety of international conferences.

Abstract

There has been a rapid grown in the fi eld of tumor immunobiology in recent years as a result of recent successes in cancer immunotherapies, and it is becoming clear that immune cells play many sometimes confl icting roles in the tumor microenvironment. However, obtaining phenotypic information about the various immune cells that play these roles in and around the tumor has been a challenge. Existing methods can either deliver phenotypic information on homogenous samples (e.g., fl ow cytometry or PCR) or morphologic information on single immunomarkers (standard IHC). We present here a methodology for delivering quantitative per-cell marker expression and phenotyping, analogous to that obtained from fl ow cyt ometry, but from cells imaged in situ in FFPE tissue sections. Th is methodology combines: Th e sequential multi-marker labeling of up to 8 antigens using antibodies all of the same species in a single section; automated multispectral imaging (MSI) to remove the typically problematic FFPE tissue autofl uorescence and correct cross-talk between fl uorescent channels; and an automated analysis that can quantitate the per-cell marker expression, determine the cellular phenotype, count these cells separately in the tumor compartment and in the stroma and provide high-resolution images of their distributions. We present here three examples of this new methodology: Th e simultaneous labeling, analysis and validation of CD3, CD8 and FOXP3 multiplexed staining in follicular lymphoma; a CD4, CD8, CD20, pan-CK TH, TC and B cell panel in breast cancer; and PD-L1, CD8, CD34 and FOXP3 phenotyping and quantitation in melanoma. Each example will show the application of the multiplexed staining, per-cell quantitation and cellular phenotyping from multispectral images of FFPE tissue sections.