Fig. 4

Analysis of a heterogeneous tissue with local indication of spatial association (LISA). a IF image of a tissue with the definition of two regions of interest c and d, in which we will analyze the heterogeneity in gene amplification. b Model of the tissue in a using the same cell-by-cell representation as in Fig. 2d, e. c Heterogeneity analysis of the region of interest c of the tissue in a, i IF image: blue = nuclei, red = HER2, green = CK. ii LISA analysis of the cells in i. Cells are classified based on their own HER2 IF-status (High or Low) and on the IF-status of their neighbors (High or Low), resulting in: High–High (HH), High–Low (HL), Low–High (LH), Low–Low (LL)-type cells. Cells and their neighbors are classified as High (respectively Low) if their HER2/CK ratio is higher (respectively lower) than a threshold of 0.25 as obtained from Fig. 3a. iii FISH image of the region. Blue = nuclei, green = CEP17, red = HER2. iv Automatic scoring of HER2 loci and CEP17 of the region in iii. v Spatial association status of the cells in iv. Cells are classified as High (respectively Low) if their HER2 loci number is ≥ 6 (respectively < 4).They are non-classified (NC) in the intermediate interval of HER2 loci number from 4 to 6, where the FISH HER2 status is equivocal. (di-v) Heterogeneity analysis of a HER2 equivocal region of the region of interest d of the tissue in a, following the same procedure as in (c). While the IF readout is similar, the FISH status is clearly distinct. e Spatial association analysis of the HER2 protein expression for the whole tissue in (a). f Spatial association analysis of HER2 amplification for the whole tissue in a, showing cluster heterogeneity, i.e., having clusters of HH cells that span more than 10% of the tissue area. Scale bars: 10 µm