Comprehensive characterization of the tumor microenvironment benefits from simultaneous measurement of transcripts, proteins, and morphology at single-cell resolution. The G4X Spatial Sequencer is a high-throughput in-situ platform delivering co-registered RNA, protein, and morphological readouts from the same FFPE section, supporting up to 40 cm² of tissue per run.

Four targeted ~300-gene RNA panels for kidney, lung, colon, and breast tissues were designed — each including 150 shared immuno-oncology genes, 50 stromal genes, and ~100 tissue-specific targets. A paired 16-plex protein panel provides spatially resolved multiplexed proteomics. All analytes are acquired from single 5 µm sections and directly aligned with fH&E™.

Per-cell detection ranged from 50–200 transcripts and 20–50 unique genes, with dynamic ranges exceeding 500 transcripts in high-expressing cells. RNA FDR ≤ 0.5% (typically < 0.1%). Protein data quality matched single-plex immunofluorescence, and integration improved cross-modal specificity. G4X is deployed in ongoing atlas-scale lung and colorectal cancer studies and has been used for 3D reconstruction across serial sections.

A critical gap remains in sequencing highly variable transcript regions in situ at subcellular resolution — essential for mapping B- and T-cell clonotypes in their native tissue context. Direct-Seq™ employs probes targeting V and J regions flanking the diverse CDR3 domain of IgH and TCRβ transcripts.

Applied to FFPE tonsil and renal cell carcinoma sections: up to 9% of B cells profiled in FFPE tonsil and ~20% of B and T cells in fresh frozen tissue. Clonally expanded B cells with highly similar CDR3 sequences localized within germinal centers. In RCC, nine serial FFPE sections revealed persistent T-cell clones across serial sections.

Direct-Seq was combined with multiplexed protein detection on the same sections, confirming spatial concordance between transcript identity and protein phenotype.

Metastasis accounts for over 90% of CRC mortality. SPOT-Met integrates subcellular-resolution spatial multi-omics with AI-driven modeling to infer the rules governing organ-specific metastasis.

1,000 CRC primary tumors and ~100 matched metastatic tissues were profiled on G4X, generating > 300 million cell profiles at submicron resolution. Preliminary data suggest immune organization and cell-cell topology — rather than total immune content — may distinguish therapeutic outcomes. Liver-tropic tumors preferentially form perivascular stromal hubs enriched for metabolic and ECM signatures.

SPOT-Met is being developed as a biopsy-compatible diagnostic assay to predict metastatic potential and organotropism at the time of diagnosis.

Thirty CRC samples underwent G4X spatial profiling with a 360-gene panel across 3,088,252 cells. STAT1 was identified as the top upregulated gene in cancer-associated cells (log2 FC = 2.11; adj. p < 0.001).

Clustering revealed 10 distinct cell clusters including CAF subtypes, TAMs, mixed T cells, and plasma cells. Regression analyses indicated a negative association between STAT1-positive plasma cells and CRC recurrence (OR = 0.19; 95% CI = 0.03–1.13; p = 0.068), supporting spatial transcriptomic insights for patient stratification.

FFPE gastric tissues from 12 patients (38 ROIs) were profiled on G4X using a custom 358-gene panel, 16-plex protein panel, and fH&E on the same slide. 3D holotomography was used to embed 2D molecularly defined states into gland and mucosal context.

Multimodal analysis resolved distinct epithelial, stromal, and immune compartments. Same-slide RNA-protein measurements enabled mapping of CD4⁺ and CD8⁺ T-cell distributions relative to gastric epithelium. 3D HT refined estimates of lesion extent and immune proximity along the gland axis.

Serial 2D slices showed discordant cell-type compositions and neighborhood enrichments — strong variability even within a single tumor block. In contrast, 3D reconstruction revealed a coherent exhausted T-cell domain that failed to reach significance in any individual slice.

Holotomography confirmed dense, continuous interfaces among T cells, ECM, and tumor cells across depth, supporting true biological continuity. This provides a practical workflow for volumetric tumor-immune profiling that can accelerate biomarker discovery.

Profiled primary DGCs from 26 patients across the full histologic continuum using GeoMx DSP, Visium, Xenium Prime 5K, CosMx, and G4X with holotomography for 3D context.

Spatial profiling revealed invasion-associated transcriptional programs along the SRC-to-invasive continuum. Copy-number inference indicated linear and branched evolution trajectories. Single-cell spatial multi-omics revealed lymphocyte-dominant states at early lesions and stromal/myeloid remodeling with deeper invasion.

Two 3D models from 7 and 9 serial sections (16 total) representing 35–45 µm tissue depth and 1.6M and 3.4M cells. Models recapitulate 3D morphology of epithelial glands, vasculature, and tertiary lymphoid structures.

TFF2 (a SPEM marker) was observed in tumor-adjacent gastric mucosa. Small clusters of TFF2⁺ cells within the superficial tumor-invasive area were only present in a few tissue layers — rare-cell events enabled by 3D modeling.

ProteoBridge is a deep learning framework that predicts protein expression across serial sections using routine H&E on remaining sections to infer protein intensities — trained on a single section with H&E, multiplex protein, and RNA data.

Using single-slide supervision, ProteoBridge accurately reproduced marker intensities and preserved cross-plane consistency, reducing cost and turnaround time while expanding effective proteomic coverage for 3D tumor proteome reconstruction.

Spatial multiomic analysis of 31 ROIs in matched primary tumor and metastatic lymph nodes from 8 PDAC patients yielded 2.4 million cells. CAFs subclustered into inflammatory, mechanoresponsive, and steady-state phenotypes.

7/8 patients contained putative TLSs, with CD20⁺/CXCR5⁺ B cells surrounded by CD3⁺/CD4⁺/CD8⁺ T cells. CXCR5 colocalized with CXCL13, establishing this axis as a target for TLS investigation in PDAC.

Spatial profiling of 30 SPN sections from 10 patients yielded 3.6 million cells. Five distinct tumor subpopulations were identified. CPB1⁺ tumor cells showed increased acinar markers and formed spatial gradients at tumor-acinar interfaces, suggesting an acinar lineage origin.

STAT1⁺/TAP1⁺/CD74⁺ cells localized in immune-infiltrated areas, corroborated proteomically. This is the first multiomic analysis of pancreatic SPNs.

ScRNA-seq on 15 recurrent ULMS tumors (204,250 cells) plus spatial transcriptomics on 29 sections yielded 2.3 million spatially resolved cells. Dedifferentiated stem-like cells with high ESR1, PGR, and AR expression were scattered throughout tumor sections.

These hormonal cells were the most sensitive to all interrogated drugs including first-line therapies. AR/PGR-expressing cells correlated with improved survival; ischemic cells with worst outcomes.