Image Mass Cytometry - IMC

Imaging Mass Cytometry (IMC) is a multiplexed image analysis of tissue achieved by the high content of mass cytometry enabled through the use of isotopically labeled probes and ICP-MS detection. Empowering researchers to interrogate up to 35 markers simultaneously on single section of tissue, IMC is an advanced single-cell proteomics technology and is a powerful tool to unveil new cell types, functions and biomarkers in cancer, immunology and more. Importantly, IMC is performed on FFPE tissue sections, so it can be applied to the retrospective research of disease and drug development using a limited number of pathological sections. LIDE has developed a validated panel if IMC antibodies, including marker of cancer, matrix, immunology and more and provides basic bioinformatical analysis. 

Representative mass cytometry image of a lung cancer patient’s tissue section
Representative mass cytometry image of a lung cancer patient’s tissue section showing the overlay multiple markers. Up to 35 markers were detected and selected markers could be represented by pseudo-colored raw ion images. 

Single-cell features could be computationally segmented using an algorithm and the single-cell marker expression data could be extracted. These single-cell data could be used for downstream data analyses to investigate cell subpopulations and micro environment.

Single-cell features
A. Single-Cell Identification and cell object image. The original data were segregated into single cells and the multiple markers were used to identify nuclei and cell membranes. The data were combined and analyzed using the algorithm to identify single-cell object masks, containing the cell location and boundary. 
B. t-SNE map displaying single cells from each cluster identified in heatmap images colored according to cluster. Single cells from the large cohort were clustered into groups according to their phenotypic similarity.
C. Heatmap showing the z-scored mean marker expression of the panel markers for each cluster. Clusters and markers are grouped according to expression profiles. The resulting 19 clusters were aggregated into larger groups following hierarchical clustering of their mean marker correlations. Multiscale bootstrap resampling was used to assess the uncertainty of each subtree.