Additional file 2: Figure S1. of An ecological measure of immune-cancer colocalization as a prognostic factor for breast cancer

Supplementary Figures and Tables. This file contains the following figures and tables: Validation of image analysis with pathological scores. Figure S2. Comparison of Voronoi and Square tessellation. Figure S3. Comparing the Morisita-Horn index with Pearson correlation. Figure S4. Heatmap to show correlation among Pearson correlation and the Morisita-Horn index computed over different scales with square (S) or Voronoi (V) tessellation. Figure S5. Association of immune-cancer correlation with survival according to different spatial scales. Figure S6. Optimizing the Morisita index cutoff in the discovery cohort. Figure S7. Immune-cancer colocalization sub-stratifies clinical parameters and immune abundance, estrogen receptor (ER) status and TP53 mutation status in human epidermal growth factor receptor 2 (HER2)â +â subtype. Figure S8. Robustness of prognostic value of Morisita index in HER2-amplified subtype. Figure S9. Intra-slide score variability. Figure S10. Kaplan-Meier curves of the Morisita index estimated using decreasing amount of tissue. Table S1. Prognostic value of immune-cancer colocalization measures with Voronoi tessellation using univariate and multivariate Cox regression results in two independent cohorts. Table S2. Univariate and multivariate Cox regression analysis result with the Morisita index, node, size and grade in four PAM50 subtypes and HER2-amplified subtype defined by SNP6 microarray with two cohorts. Table S3. Groups of multivariate Cox proportional hazard analysis of the Morisita index and treatment, known clinical or other variables in HER2-amplified, HER2-amplified and chemotherapy (CT)-treated, and HER2-amplified and radiotherapy (RT)-treated patients. (DOCX 6725 kb)