HistoLoop improves gene activity maps from tissue images using iterative context

HistoLoop: Iterative Contextual Inference with a Looped Transformer for Histology-Based Spatial Gene Expression Prediction

Computer Vision and Pattern Recognition

Summary

Figuring out where genes are active in tissue samples helps scientists understand diseases like cancer. The authors created HistoLoop, a computer method that looks at microscope images of tissues and repeatedly improves its guesses about gene activity by considering neighboring spots. This step-by-step approach lets HistoLoop understand complex gene patterns better than previous methods. Their tests show it works well across different cancers and organs.

What this means in practice

  • For biotech product developers: Enhance diagnostic tools by integrating HistoLoop to predict gene activity patterns from standard tissue images without extra costly experiments.$Commercial implications: Enables development of advanced, image-based gene expression mapping products for healthcare providers improving diagnosis and treatment planning.
  • For computational pathology teams: Incorporate iterative refinement methods like HistoLoop to better capture spatial gene patterns in cancer tissue analysis workflows.

Authors

Ziang Liu, Jie Liu, Shuai Li, Yigui Feng, Qinglin Wang

Abstract

Predicting spatial gene expression from H&E images offers a scalable complement to experimental spatial transcriptomics. However, despite progress in spatial context modeling, how to construct subsequent context from spot states updated by earlier contextual interactions remains underexplored. We introduce HistoLoop, a looped Transformer that formulates prediction as iterative contextual inference. At each refinement step, HistoLoop aggregates the current spot states into context tokens at the corresponding resolution, processes them with a shared spatial Transformer, and retrieves spot-specific contextual information through cross-attention. A memory-mediated update integrates this information into persistent spot states, which are then used to construct context at the next, finer resolution. Under the HEST-Bench HVG50 evaluation protocol, HistoLoop achieves the highest macro-averaged gene-wise Pearson correlation among the compared methods across six cancer-specific tasks spanning five organs in HEST-1K. Ablation studies support the importance of recurrent state refinement and reconstructing context from updated spot states at progressively finer resolutions. Expression-derived clustering further suggests that HistoLoop better captures spatially organized multigene patterns.