How Whole-Transcriptome Spatial Profiling Reveals Hidden Tumor Heterogeneity in Breast Cancer

Why standard spatial analysis misses key tumor behavior

Spatial transcriptomics has enabled researchers to map gene expression within intact tissue architecture, providing a major step forward in understanding tumor organization¹. However, most current analytical workflows still struggle to fully resolve functional tumor heterogeneity.

The limitation is not spatial resolution, but biological representation.

Most spatial analyses rely on:

• Predefined marker genes
• Clustering-based cell type annotation
• Predefined or targeted gene panels with restricted gene coverage

These approaches work well for identifying major cell types, but they are less effective in tumors where biologically important states are:

• Rare
• Spatially restricted
• Defined by distributed transcriptional programs rather than single markers

As a result, functionally distinct cell states may be merged into broader epithelial or stromal categories, even when they play specific roles in invasion, immune regulation, or metabolic adaptation.

Whole-transcriptome spatial profiling addresses this limitation by removing dependence on predefined gene panels and enabling unbiased analysis of transcriptional programs directly in spatial context¹.

Why whole-transcriptome spatial profiling matters

Unlike targeted spatial platforms that measure a few thousand genes, whole-transcriptome approaches capture approximately 18,000–20,000 genes per spatial location.

This shift changes the analytical question: from “Are known markers present?” to “What transcriptional programs exist in this tissue?”

This is particularly important in tumors, where key biological processes are rarely defined by single genes. Instead, they emerge from coordinated programs involving immune regulation, hypoxia response, metabolic reprogramming, epithelial plasticity, and stromal interaction.

However, increased transcriptome coverage introduces a new challenge. Whole-transcriptome spatial experiments can generate hundreds of thousands to millions of spatially resolved cells, creating datasets that are difficult to interpret using standard clustering or marker-based approaches alone.

The opportunity is no longer simply generating more data—it is transforming that data into biological insight.

From whole-transcriptome spatial data to biological discovery

High-throughput whole-transcriptome platforms such as Atera generate rich spatial datasets that combine tissue morphology, spatial coordinates, and transcriptome-wide expression profiles. The challenge is transforming these measurements into biological insight.

Using SpatialX, researchers can move from raw spatial data to:

• Cell-type annotation
• Rare-state discovery
• Spatial neighborhood analysis
• Cell-cell interaction mapping
• Region-specific biological interpretation

Workflow overview:

The following case study illustrates how this workflow reveals hidden tumor heterogeneity in breast cancer tissue.

Figures illustrate from SpatialX:

Figure 1. Overlay of morphology and spatial transcriptomic data.

Figure 2. Cellular neighborhoods analysis.

Figure 3: Spatial domains identification

Figure 4: Cell-cell interaction analysis.

Case study: resolving hidden structure in breast cancer tissue

To illustrate how whole-transcriptome spatial datasets can be translated into biological insight, we highlight findings from a recent breast cancer study ². The study measured more than 18,900 genes while preserving single-cell spatial resolution across breast cancer tissue.

By integrating gene expression with tissue architecture, the study enabled simultaneous analysis of malignant, immune, and stromal compartments within their native spatial context.

Three key findings emerged.

1. Rare invasive tumor states are revealed by transcriptome-wide analysis

A rare ITGB6⁺ tumor population was identified, representing approximately ~0.7% of tumor cells.

This population:

• Exhibited elevated expression of ITGB6, MME, and PMEPA1
• Showed transcriptional programs associated with epithelial-to-mesenchymal transition (EMT)
• Was not resolved as a distinct cluster under marker-constrained analysis

Although rare, these cells represent a transcriptional state associated with invasive potential, highlighting the importance of whole-transcriptome resolution for detecting low-abundance but biologically meaningful tumor populations.

2. Tumors organize into spatial functional ecosystems

Beyond individual cell states, the tissue resolved into spatially distinct domains characterized by coherent transcriptional programs.

These domains broadly corresponded to:

• Immune-rich regions with T cell infiltration
• Hypoxic tumor cores with metabolic stress programs
• Stromal remodeling zones
• Invasive boundary regions associated with tumor expansion

These spatial domains were not defined by morphology alone but emerged from transcriptome-wide expression patterns.

For example:

• NF-κB-associated signaling was enriched near tumor margins, suggesting localized immune-modulatory activity
• Stromal regions exhibited CCR2 and EVL expression patterns consistent with migration and remodeling programs ^3,4.

Together, these findings show that tumor behavior is strongly shaped by spatial context rather than being uniformly distributed across malignant cells.

3. Spatial niches reveal immune and metabolic organization

Whole-transcriptome spatial profiling further revealed that immune and metabolic programs vary systematically across tissue space.

Different T-cell populations were distributed unevenly:

• CD8⁺ T cells were more frequently observed within tumor-associated regions
• CD4⁺ and naïve T cells were largely excluded from interior tumor compartments
• Cytotoxic CD8⁺ T cells were enriched in stromal and tumor-adjacent regions

Notably, cytotoxic CD8⁺ T cells represented only ~0.6% of tumor-associated cells yet retain substantial effector potential. Their spatial restriction suggests that immune activity may be limited more by localization than by abundance.

In parallel, metabolic programs were strongly associated with vascular proximity:

• Cells near blood vessels showed signatures of active growth and protein synthesis
• Cells further from vasculature activated glycolysis and nutrient stress pathways

These gradients highlight how immune activity, metabolism, and tissue architecture are tightly coupled across spatial organization.

Turning Atera-scale spatial datasets into biological insight with SpatialX

The biological patterns described above are not directly visible in raw spatial transcriptomic data. They emerge through structured analysis that connects gene expression, spatial organization, and cellular interactions.

SpatialX enables this transformation by providing an environment to:

• Identify rare transcriptional states
• Map their spatial localization
• Define cellular neighborhoods and tissue domains
• Infer cell-cell communication networks
• Compare biological programs across regions and samples

Applied to high-throughput Atera datasets, this workflow enables researchers to move beyond descriptive tissue maps toward a systems-level understanding of the tumor microenvironment.

Rather than asking only which cells are present, researchers can investigate how tumor, immune, and stromal populations organize into functional ecosystems and how those ecosystems shape disease progression.

As spatial transcriptomics scales from individual tissue sections to cohort-level studies, analytical workflows become just as important as the underlying assay.

Looking beyond tumor subtypes

The value of whole-transcriptome spatial profiling is not simply increased gene coverage. It is the ability to reveal biological structure that remains hidden in aggregated or targeted measurements.

Across breast cancer tissue, three consistent themes emerge:

• Rare cell states can carry disproportionate biological importance
• Tumors are organized into spatially distinct functional ecosystems
• Immune, metabolic, and invasive programs are strongly shaped by spatial context

Together, these observations suggest that tumor heterogeneity is not only present at the cellular level but also organized across tissue space in ways that conventional analyses often fail to capture.

As technologies such as Atera continue to expand the scale of spatial biology, researchers will increasingly be able to study tumors as spatially organized biological systems rather than collections of independent cell types.

The question is no longer whether tumors are heterogeneous. The question is how much of that heterogeneity remains unseen.

Curious how SpatialX streamlines analysis of Atera-scale spatial transcriptomics data? Request a demo with our team

References

1. Zormpas, E., Queen, R., Comber, A., et al. “Mapping the transcriptome: Realizing the full potential of spatial data analysis,” Cell, 186(26), 2023
2. Williams, C., Cui, Y., Patrick, M., et al. “Breast cancer through the lens of whole transcriptome spatial imaging,” bioRxiv, 2025.
3. Mouneimne, G., et al. Cancer Cell, 2012.
4. Tsuyada, A., et al. Cancer Research, 2012.