Advances in Spatial Immune Profiling

The immune system is made up of cells that protect the body from harm. These cells guard against infection, prevent cancer growth, modulate autoimmune activity, and support countless biological processes. Understanding how these immune cells function and what happens when they fail is essential for advancing disease prevention and treatment. To do this, researchers must study immune cells in context, including their location, the cells they interact with, and the genes and proteins involved in both healthy and diseased tissues. Capturing this information once required combining data from separate techniques. Spatial profiling now allows it to be collected in a single experiment. As a result, scientists can explore immune behavior with greater clarity and biological relevance, supporting research in cancer, infectious disease, autoimmunity, neurodegeneration, transplantation, and more.

Advances in spatial profiling technologies

Early spatial profiling technologies, such as in situ hybridization, immunohistochemistry, and immunofluorescence, were limited to analyzing just one marker per slide. According to Traci Degeer, Director of Innovation at Leica Biosystems, recent technological advances have enabled tools like multiplex immunofluorescence and multiplex chromogen staining that allow scientists to detect numerous biomarkers on the same slide. “That’s especially important if the tissues are precious,” she stated, “because biopsies have gotten smaller, and the tissue researchers work with is increasingly limited.” These innovations have also made it possible to examine critical cell interactions, such as PD-1/PD-L1 dynamics in the tumor microenvironment.

Spatial transcriptomics builds on these advances by capturing the detailed gene expression within tissues of interest. “It gives us a unified view of the molecular identity of each cell anchored in its physical location, and that’s incredibly powerful,” stated Michael Schnall-Levin, Ph.D., Chief Technology Officer and Founding Scientist at 10x Genomics. This additional context, he said, enables researchers to identify immune cell subtypes, identify pro- and anti-inflammatory phenotypes, track tumor-driven immune modulation, and monitor therapeutic responses, such as CAR-T cell distribution.

Todd Dickinson, Ph.D., Chief Executive Officer and Board Member at Stellaromics, highlighted how spatial transcriptomics is now being extended into 3D, where researchers can localize gene expression to specific regions and cell types within intact tissue. “This has been critical in the study of tumor microenvironments, where the behavior of immune cells is often highly localized and dependent on spatial cues,” Dickinson stated. This spatial insight also makes it possible for researchers to observe distinct immune patterns that are not evident in dissociated analyses.

Alongside transcriptomic technologies, Imaging Mass Cytometry™ (IMC™) adds another important layer of protein-level data from immune environments. “This approach offers scalable and high-throughput acquisition while generating high-quality data without fluorescence-based limitations such as spectral overlap, autofluorescence, or signal amplification,” stated Jennifer Ellis, MS, Director, Technical and Scientific Content at Standard BioTools. She added that same-slide multiomics, whether integrating H&E and high-plex proteomics, layering multiple proteomics technologies on the same slide, or conducting proteomic imaging following spatial transcriptomics data collection, is also a developing area that offers the ability to generate RNA and protein expression data from the same cells in one tissue section, giving new insights into biological questions that previously required separate analytical approaches.

Current challenges in data analysis

Despite the significant insights these technologies provide, analyzing and interpreting their data remains a major hurdle. One key challenge in interpretation, Ellis pointed out, is the correlation and dissimilarity between protein and RNA levels. “The transcriptome and proteome each provide distinct layers of information that support the understanding of biological processes,” she noted. “While mRNA levels show the potential for protein production, protein levels indicate which proteins have been produced and are actionable targets.” Since RNA levels don’t predict protein abundance or vice versa, integrating data from both modalities is essential for an accurate picture of cellular function.

Another core challenge is segmentation, particularly in immune-rich environments. Schnall-Levin explained that “precise segmentation is crucial for immune cells as they are often found in close proximity to each other and have irregular boundaries,” a problem further complicated by interference from surrounding stromal tissue. He also shared that spatial analysis still lags behind single-cell analytics in terms of mature computational tools, and emphasized that combining spatial data with other biological information, such as pathological annotation, is key to unlocking its full value.

As spatial data becomes more multiomic, Dickinson stressed that integration and interpretation are the biggest challenges, and that each modality has its own resolution, noise characteristics, and analytical complexities. “When you combine them in a spatial framework, the data dimensionality increases dramatically,” he noted. To make sense of this, researchers need computational methods that can align these data layers within the same tissue section while also interpreting them in the context of tissue architecture and cell-to-cell interactions.

One of the other major issues in analyzing spatial immune data, Degeer explained, is the sheer volume and complexity of information generated, especially when combining high-plex protein markers, transcriptomics, and single-cell data. Integrating these diverse datasets requires advanced tools, but there are limited commercially deployable options. While some companies have internal solutions and there are some commercial solutions on the market, they are not comprehensive, aren’t always scalable or widely accessible, and much of the analysis still relies on manual effort or long, customized workflows. Degeer emphasized that data interpretation remains time-consuming and resource-intensive, creating a significant bottleneck in research and therapeutic development.

AI and image analysis in spatial profiling

With spatial profiling now generating high-dimensional datasets at a massive scale, researchers are turning to AI and advanced image analysis to uncover patterns that would otherwise remain hidden. “We’re at a moment when the scale and complexity of spatial immune data have far outpaced what manual analysis can keep up with,” stated Schnall-Levin. “There’s a growing recognition that we need to transition to more efficient, scalable image analysis and AI-generated annotations to accelerate research.” Applying AI effectively, however, requires training models on high-quality spatial and molecular datasets. Schnall-Levin pointed to a study from Nicholas Banovich's lab at TGen that used Visium and Xenium to identify molecular niches linked to epithelial detachment in idiopathic pulmonary fibrosis.¹ While not AI-driven, he believes that it shows how large datasets could enable AI to reveal key molecular patterns.

Dickinson also emphasized that AI is essential for interpreting the vast datasets produced by spatial transcriptomics. He noted that machine learning and advanced image analysis are revealing patterns such as immune cell neighborhoods and cytokine gradients that are difficult to detect manually. AI can also identify tissue architecture, classify cell types, and quantify immune interactions. Stellaromics is investing heavily in algorithm development on its Pyxa platform to enable these kinds of discoveries. As Dickinson stated, “It’s not just about data capture anymore; it’s about deriving insight at scale, and AI is central to that process.”

While some spatial techniques can be used without the need for AI, Degeer noted that many newer approaches increasingly incorporate AI to count mRNA signals, assess proximity between immune and tumor cells, and interpret complex spatial patterns. At Leica Biosystems, Degeer’s team focuses on developing AI-driven workflow enhancements, including advanced segmentation and tumor detection algorithms. She also noted an ongoing collaboration with Indica Labs focused on advanced cell segmentation, quantification, and spatial analysis.

Building on this, Ellis shared several specialized tools used to extract meaningful insights from the data generated by multiplex imaging technologies. To analyze this spatial and neighborhood data, she explained that researchers use tools like Phenoplex, Halo, Qupath, histoCAT, MCD Smartviewer, ImaCytE, MCMICRO, Spectre, and cytomapper, among other tools created by academic labs, all of which incorporate AI. Ellis also highlighted the use of AI-powered solutions like Biomics, designed to streamline data analysis and support biomarker discovery.

Applications of spatial profiling in immune research

In applied research, these technologies are advancing our understanding of how immune cells shape health and disease. For instance, a Broad Institute study using STARmap—the technology upon which Stellaromics’ Pyxa was developed—demonstrated that 3D spatial transcriptomics in thick tissue more accurately reveals tumor-immune interactions, such as interaction patterns between macrophages and T cells.² Focusing on tissue conservation in lymphoma, Leica Biosystems completed a spatial project that used chromogenic multiplex staining and AI-driven image analysis on small biopsies to improve identification of overlapping markers critical for accurate subtyping.

In another study, researchers used IMC to show that spatial organization influences response to immune checkpoint blockade in triple-negative breast cancer, by identifying predictive cell types and interactions and suggesting early biopsies could guide improved development of precision immuno-oncology strategies.³ A 10x Genomics study using Visium HD and Xenium further underscored the value of spatial profiling across cancers, revealing that macrophage behavior in colorectal cancer is shaped by the tumor microenvironment, with anti-tumor immune niches at the tumor edge offering potential targets for drug discovery.⁴

Future directions in spatial profiling

Spatial profiling technologies are advancing immune research by identifying transcripts, proteins, and cellular interactions within their native tissue context. Although data analysis and interpretation remain significant bottlenecks, the emergence of AI and other computational tools is beginning to address these challenges and support more integrated and scalable analyses of complex, multimodal datasets. The development and application of these technologies will be instrumental in accelerating discoveries across cancer, infectious disease, autoimmunity, and more, bringing us closer to a more complete understanding of immune function in health and disease.

References

1. Vannan A, Lyu R, Williams AL, et al. Spatial transcriptomics identifies molecular niche dysregulation associated with distal lung remodeling in pulmonary fibrosis. Nat Genet. 2025;57(3):647-658. 

2. Sui X, Lo JA, Luo S, et al. Scalable spatial single-cell transcriptomics and translatomics in 3D thick tissue blocks. Preprint. bioRxiv. 2024;2024.08.05.606553. Published 2024 Aug 8. 

3. Wang XQ, Danenberg E, Huang CS, et al. Spatial predictors of immunotherapy response in triple-negative breast cancer. Nature. 2023;621(7980):868-876.

4. Oliveira MF, Romero JP, Chung M, et al. High-definition spatial transcriptomic profiling of immune cell populations in colorectal cancer. Nat Genet. 2025;57(6):1512-1523.