Abstract
Multiple myeloma (MM) is a plasma cell malignancy shaped by dynamic interactions between MM cells and non-malignant cells in the immune microenvironment. To spatially profile the influence of cellular context on MM and immune cell expression, we developed a multimodal framework integrating 10x Genomics Visium HD, 10x Genomics Xenium, and clinically annotated single-cell RNA (scRNA-seq) sequencing datasets. Visium HD enabled unbiased, whole transcriptome, spatial discovery at 16 μm resolution, Xenium provided orthogonal validation at single-cell resolution, and scRNA-seq extended findings by mapping spatial labels and leveraging the greater sequencing depth. We developed a custom framework for cell type annotation within Visium HD spatial bins. Our approach enabled identification of plasma cell-dense niches enriched for non-canonical Wnt signaling, associated with gene expression supporting cell adhesion mediated drug resistance, inferior progression-free survival, and extramedullary lesions. Immune cells within these neighborhoods exhibited suppressed transcriptional states, including increased inhibitory receptor expression such as LAG3. Utilizing the niche-driven transcriptional states in MM and immune cells, we were able to develop a 15-gene signature independently predictive of progression free survival (HR = 2.00, p < 0.0001). Collectively, this study demonstrates the potential of integrated spatial and single-cell transcriptomics to define niche-specific programs supporting MM progression.