NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics

Spatial transcriptomics captures molecular states within cells and their organisation in tissue. However, integrating fine-grained gene information with spatial context at scale remains challenging for existing foundation models. Here we present NexuST, a hierarchical foundation model that repeatedly interleaves gene-level molecular modelling with cell-level spatial modelling, allowing the two levels to refine one another during end-to-end pretraining. For pretraining, we curated HumanST-46M, comprising 45.7 million human cells from 72 datasets across 11 organs and three imaging-based platforms. Across four held-out datasets totalling approximately 2.6 million cells, NexuST achieved state-of-the-art or competitive performance in cell-type annotation, region prediction, gene recovery and neighbourhood-composition prediction. We find that cell-intrinsic expression remains informative even for spatial tasks, as shown by an expression-only PCA baseline, while NexuST shows particularly strong gains where spatial context is essential. Overall, NexuST establishes a hierarchical framework that can serve as a general backbone for future spatial transcriptomics foundation models.