Read the full review here: doi.org/10.1002/1878.... Exciting times for spatial cancer biology.
 Thanks to @mariasecrier.bsky.social and all co-authors!
#SpatialOmics #AI #Cancer

ST can uncover biomarkers of therapy response, resistance niches and new targets. We discuss how integration with pathology and AI can bridge the gap to patient care.

Linking ST with histology and AI could make spatial insights accessible in the clinic. We outline current approaches, from gene expression prediction to super-resolution mapping.

Graph neural networks, geospatial statistics and self-supervised learning are transforming niche detection and cell–cell communication analysis.

Niches are ecosystems of interacting cells driving immune evasion and therapy resistance. We review statistical, geospatial and AI-based tools to detect and model these niches.

Cell identity is more than type. States like EMT or T-cell exhaustion shape tumour behaviour. We discuss methods to capture these states and why spatial context changes the game.

Key challenges in ST for cancer

(a) Defining cell states beyond cell types

(b) Characterising cellular niches and interactions

(c) Integrating spatial data with other modalities for clinical use

Spatial transcriptomics (ST) is changing how we study cancer. It maps gene expression to tissue structure, revealing how cells interact within the tumour microenvironment. But big challenges remain. Here’s what we cover.

Thanks all members of @mariasecrier.bsky.social lab for their comments and critique during the development of EnrichMap.

Want a sneak peek or to try it on your own data? Try EnrichMap enrichmap.readthedocs.io/en/latest/
#SpatialTranscriptomics #ComputationalBiology #EnrichMap #spatialomics

Generalisable to any tissue.
EnrichMap isn’t just for tumours – it’s for anyone working with spatial transcriptomics.

We’ve designed it to be fast, flexible and easy to use:
- Works seamlessly with Scanpy, Squidpy and AnnData (@scverse_team compatible)
- Plug-and-play any ST dataset
- Outputs spatially informed scores
- Geostatistics-inspired metrics to evaluate gene sets

What can EnrichMap do?
- Correct for spatial confounders
- Support batch correction and gene weighting
- Map gradients of tumour and immune programmes
- Compare microenvironments
- Reveal niche-specific signalling

Why does this matter?
Tumours are messy. Immune niches are local. Cell states vary in space. If you’re collapsing expression without space, you’re missing the story. EnrichMap keeps that story intact.