48 Advances in multiplexed ion beam imaging (MIBI) for immune profiling of the tumor microenvironment

Background Multiplexed ion beam imaging (MIBI) combines time-of-flight secondary ion mass spectrometry (ToF-SIMS) with metal labeled antibodies to image 40+ proteins in a single scan at subcellular spatial resolution. Here, we show that the recently released MIBIscope provides improved sensitivity for detecting immune checkpoint markers and offers greater throughput at higher resolution than the alpha instrument. Methods Serial sections from three FFPE NSCLC samples, in addition to a control slide consisting of various unremarkable tissues, were stained with a panel of 25 metal labeled antibodies. The tissue was imaged at subcellular resolution using the MIBIscope and the alpha instrument. Masses of detected species were assigned to target biomolecules given the unique label of each antibody and multi-step processing was used to create images. Cell classification was performed using two complementary methods that differed in the need for cell segmentation to phenotypically characterize the tissue environments and quantify marker expression. Results Replicate regions of interest (ROIs) were collected on both instruments with similarly sized ROIs acquired in 17 minutes with the MIBIscope compared to 280 minutes with the alpha instrument. Fourier Ring Correlation (FRC) showed the resolution to be greater on the MIBIscope as compared to the alpha instrument with FRC also demonstrating uniform resolution across an ROI 2.5X greater in size. Even with the 16X greater speed of the MIBIscope, the signal of the 25 markers across replicate ROIs was increased (y=x^1.07) and showed similar expression patterns to those observed on the alpha instrument (figure 1). This resulted in greater sensitivity to markers with low expression, such as checkpoint markers. Eleven cell populations were classified across the ROIs utilizing two methods, with both methods showing a similar frequency of tumor cells and B, T, and myeloid cell subsets between instruments. Segmentation enabled the number of cells within a population to be calculated but defining boundaries is laborious and signal from neighboring cells can result in misclassification. Performing classification at the pixel level, without segmentation, enabled the fraction of the tissue that is tumor or any other cell type to be rapidly determined. Conclusions The MIBIscope enables the phenotypic characterization of tumor and non-tumor microenvironments. Co-expression of markers can be used to classify tumor and immune populations and to quantify the expression of markers associated with immune suppression. The increased sensitivity and throughput of the MIBIscope, in combination with the 40-parameter capability and subcellular resolution, provides a platform uniquely suited to understanding the complex tumor immune landscape.