363

---

Click the image below to view the full size version of this cover.

Issue 363: Multiscale Spatial Analysis (MSA) involves extracting and examining spatial information from data across multiple scales—such as images, quadrats, and cells. This cover art provides a conceptual visualization of MSA using multi-stain images from Lung cancer. The background image is a risk map: a single multi-stain image depicting gradients of disease-progression probabilities, computed by inferring the spatial neighborhoods of cells and quadrats. The foreground features a collage of t-SNE renderings of cells and quadrats extracted from several such multi-stain images. In our paper, we show that integrating localized information with a broader spatial context can improve predictions of disease progression.