Fig. 3

Initial experiments: whole-slide image (WSI) similarity analysis, comparison of segmentation accuracy for different approaches, and visual evaluation of segmentation accuracy (A) Similarity analysis between annotated (n = 10) and non-annotated WSIs (n = 200). Slide cohort: TCGA colorectal cancer cohort. The heatmap on the left side depicts the cross-slide case-level similarity values between annotated cases, where darker colors indicate higher similarity. There is a broad range of similarities among unrelated cases (0.29–0.79). The values of similarities for the same cases are higher than for unrelated cases but not equal 1.0 as we take five most representative regions. These values, therefore, outline intra-case intratumoral heterogeneity. The box plot on the right side illustrates the distribution of similarities for a larger cohort of non-annotated cases (n = 200). The non-annotated cases are grouped to the most similar case among annotated cases and the points are similarity values measured against these most similar cases. (B) Comparison of segmentation accuracy for different training approaches using initial study setup (TCGA dataset, annotated slide n = 10, non-annotated slide n = 200; details see Fig. 2). Three different approaches (supervised learning, traditional SSL, and SSL using S–o-M) are compared concerning pixel-wise segmentation accuracy (measured by Dice Score). We concentrate on tumor and tumor stroma classes as two bottleneck classes for multi-class tissue segmentation problem (explanation in Fig. 1 and Methods). The accompanying table (right side) provides the average Dice score with standard deviation of 3 independent experiments for each approach, statistical significance of the combined tumor and tumor stroma scores was assessed using a paired t-test (*** p < 0.001, ** p < 0.01). Conclusion: For both tumor and tumor stroma classes, new SSL S–o-M approach provides better segmentation accuracies. (C) Visual comparison of segmentation accuracy/generated pseudo-labels in non-annotated slides for supervised model (as part of traditional SSL approach) and for most “similar” model (trained on the most similar annotated case). The review of the images by pathology experts shows higher quality of segmentation/pseudo-labels for S–o-M approach.