A spatially resolved multi-omics atlas of tertiary lymphoid structures in ovarian cancer

Tertiary lymphoid structures (TLSs) coordinate localized immune responses in the ovarian tumor microenvironment. However, the maturation transitions of TLS remain poorly understood at both cellular and molecular levels. We constructed a spatially resolved multi-omics atlas that integrates histopathology, spatial transcriptomics, proteomics, and immunogenomics to characterize TLS organization, maturation, and clinical relevance in ovarian cancer. H&E-stained tumor sections from a prospective cohort of 334 ovarian cancer patients across histologies were analyzed using deep learning-assisted segmentation to identify lymphoid aggregates (LAggs), immature TLSs (iTLSs), and mature TLSs (mTLSs) with germinal centers (GCs). From these, 45 post-neoadjuvant chemotherapy omental samples were used for tissue cyclic immunofluorescence (t-CyCIF) imaging with a 30-marker immune panel, followed by Mesmer segmentation and spatial phenotyping. 30 adjacent sections underwent spatial transcriptomic profiling by GeoMx DSP. Data were batch-corrected and analyzed for differential expression, pathway enrichment, and cell-type deconvolution. B-cell receptor (BCR) repertoires were reconstructed from bulk RNA-seq data with MiXCR. Preliminary analyses indicate substantial heterogeneity in TLS and LAgg abundance and spatial localization across ovarian cancer histologies. LAggs were present in 46% of samples, whereas iTLSs or mTLSs appeared in 16% and were confined to omental metastases. Spatial metrics quantified TLS distribution and proximity to tumor nests. Early t-CyCIF results revealed distinct TLS microstructures at different maturity stages: mTLSs contain well-defined GCs while iTLSs exhibited only early B/T compartments. Processed t-CyCIF data are being analyzed using the SPACEStat spatial analysis algorithm to uncover functional spatial immune composition in TLSs. 268 selected regions of interest (ROIs) were profiled in GeoMx. GC, T-B zone, and stromal signature scores enabled molecular labeling of TLS regions. SCORPIUS pseudotime analysis delineated a maturation trajectory from LAggs to iTLSs and mTLSs. By integrating spatial data, gene expression dynamics are mapped across maturation stages to investigate key chemokine axes and pathway enrichment. Cell-type deconvolution is expected to reveal maturation-associated spatial remodeling and functional specialization of immune subsets. Ongoing BCR profiling aims to explore clonotype patterns that may reflect class switching and somatic hypermutation. Our spatially resolved multi-omics atlas captures coordinated cellular and molecular remodeling that underlies TLS maturation and functional specialization, providing mechanistic insights into anti-tumor immune activation linked to clinical outcomes.