Spatial biology is moving from research curiosity to clinical reality – and the results are already changing how we understand and treat disease. The key to unlocking its full potential lies in automated, reproducible, end-to-end workflows that turn tissue complexity into clear, actionable decisions

Every year, thousands of patients with Barrett’s oesophagus face an uncertain prognosis. A condition in which chronic acid reflux causes permanent changes to the oesophageal lining, Barrett’s affects approximately one in twenty adults in the US – yet conventional tissue examination cannot reliably tell a clinician which patients will remain stable and which will progress to the life-threatening oesophageal cancer. The consequences of that uncertainty can be serious: over-treatment of patients who would never have progressed and under-surveillance of those who will.

This is precisely the kind of clinical challenge that spatial biology was built to solve – and increasingly, it is doing so. Named Method of the Year by Nature Methods in both 2020 and 2024 and listed among the World Economic Forum’s Top 10 Emerging Technologies in 2023, spatial biology has moved rapidly from proof of concept to a tool with genuine clinical and translational impact.1,2,3

By mapping the precise location and interaction of multiple molecular markers within tissue, spatial biology reveals the biological architecture of disease in ways that single-marker tests simply cannot. And when paired with the right analytical workflow, it can turn that biological complexity into insights that can help guide actionable clinical decisions.4

The story of how that is happening – not as a future promise, but as a clinical reality today – begins with a test called TissueCypher.

While most patients with Barrett’s oesophagus will never develop oesophageal cancer, a small but significant proportion will progress to high-grade dysplasia (HGD) or oesophageal adenocarcinoma (EAC). Identifying that higher-risk group has remained a persistent challenge: traditional endoscopic surveillance and manual pathology review rely on visual and histologic features that may not fully capture the underlying biology of progression.

TissueCypher was developed by Castle Biosciences to specifically address this gap. The test applies multiplexed immunofluorescence (mIF) labelling to standard tissue biopsies to simultaneously evaluate nine protein biomarkers. These biomarkers are associated with key pathways in cancer progression: tumour suppression; cell cycle regulation; angiogenesis; inflammation and immune cell infiltration.5,6 The tissue is labelled and then digitally imaged using ZEISS Axioscan 7 system and SlideStream workflow automation, providing high-resolution, multichannel, whole-slide images that capture not just the presence of each biomarker but its precise spatial distribution within the tissue architecture.

From these images, Castle’s AI-driven analysis platform characterises 15 spatial and quantitative features from the multiplexed data, combining them into a fixed algorithm that generates a risk score from 0 to 10.5 Patients are stratified into low, intermediate or high-risk categories, with an associated five-year probability of progression to HGD or EAC. Castle Biosciences delivered more than 39,000 TissueCypher test reports in 2025 – a figure that underscores both the clinical demand for spatially-resolved prognostic tools and the operational maturity of the underlying workflow.6

The TissueCypher test measures nine protein biomarkers in seven cellular and tissue structures using multiplexed immunofluorescence, high definition digital microscopy and AI driven image analysis. The test is used to predict a patient’s five-year risk of progression to high-grade dysplasia or oesophageal adenocarcinoma in Barrett’s oesophagus

The TissueCypher case demonstrates something important for the broader industry: that spatial biology is not merely a research enabler, but a commercially viable clinical platform. The key enabling factors of multiplexed immunofluorescence labelling standardisation, high-throughput imaging with consistent output, and artificial intelligence (AI)-driven image analysis and risk classification are precisely the capabilities that a robust end-to-end workflow must deliver.

The workflow capabilities that made TissueCypher possible did not emerge overnight. Partnerships with pioneering diagnostic companies like Castle Biosciences were instrumental in shaping and stress-testing the end-to-end solution that exists today. What began as a trailblazing collaboration is now a production-ready platform, built to serve not just diagnostic laboratories but the full range of pharma, biotech and contract research organisation (CRO) applications.

To understand what that platform delivers and why it matters, it helps to look at what spatial biology actually does and why getting it right at scale has historically been so challenging. For decades, pathologists and researchers have known that understanding disease requires more than knowing which molecules are present in a tissue, but also knowing precisely where they are and how they interact. Two patients with identical diagnoses may respond very differently to the same therapy, not because their tumour cells differ in isolation, but because the broader tissue environment around those cells, such as the immune infiltrates, the stromal architecture and the spatial relationship between cell populations, tells a fundamentally different biological story. Spatial biology is the discipline built to read that story. By preserving and analysing the native tissue context of proteins, RNA transcripts and other molecular markers simultaneously, multiplexed imaging approaches offer a far more complete picture of disease biology than any single-marker or bulk assay can provide. Yet the transition from scientific promise to routine practice poses significant challenges. Every mIF workflow involves several distinct stages: panel design; assay development and optimisation; and transfer to operations – itself a complex multi-step process covering tissue staining, whole-slide imaging, data management, image analysis and reporting. Each stage has historically represented a potential failure point. Inconsistent staining protocols, operator-dependent scan settings, variable image quality between instruments or sites, and the computational complexity of interpreting multiplexed data have combined to make spatial biology workflows difficult to standardise and even harder to scale.

For pharmaceutical sponsors running multi-site clinical trials, these inconsistencies are not just inconvenient, they can also undermine the foundation of an entire study. If immune cell density measurements vary across sites because of differences in hardware calibration or operator technique rather than true biology, the data cannot be reliably pooled or used in regulatory submissions. These are the problems that tissue multiplexing workflows need to be designed to solve: not just to enable spatial biology, but to enable it consistently, at scale, and with the rigour that translational and clinical applications demand.

ZEISS's tissue multiplexing workflow integrates four core components: a co-optimised reagent ecosystem, high-throughput automated imaging, intelligent workflow management and AI-driven image analysis, into a single, streamlined process designed to minimise hands-on time, eliminate multiple sources of variability and maximise throughput without sacrificing data quality.7

The ZEISS seamless sample-to-report experience: reagents; automated imaging; workflow management; and AI-driven analysis in a single integrated tissue multiplexing workflow7

The choice of reagent system is fundamental to analytical performance and has implications across the entire workflow. Development and operating costs, label choice, reproducibility, image contrast, and even the selection of light source and filter set are all shaped by this decision. Within the ZEISS reagent ecosystem, multiple reagent systems have been co-optimised and pre-validated to address distinct application needs, from high-plex cyclic staining for biomarker discovery, to ready-to-use assay kits for quick experimentation, to highly reproducible and cost-efficient solutions for routine use.

Designed for routine spatial biology applications, the ZEISS Axioscan 7 spatial biology scanner combines a co-optimised LED light source and filter set to separate up to eight fluorescent channels simultaneously, without the need for spectral unmixing in highly optimised assays. The use of unmixing libraries (pre-optimised within the reagent ecosystem) remains optional and can be used to reduce assay optimisation time and help mitigate autofluorescence. With a 100-slide capacity and continuous hot-swapping, the system enables laboratories to process large cohorts of more than 100 slides per day in a fully automated mode.

Consistent, comparable results across instruments, operators and time points are supported by tight control of user inputs and system performance, including Fluorescence Intensity Normalisation and Evaluation (FINE) and controlled LED intensity output. Fully-linear 16-bit high dynamic range imaging adds a further layer of robustness by giving the system greater tolerance for variability in sample material, helping it generate high-quality images with constant scan settings even when tissue characteristics and signal intensities vary. Together, these capabilities provide an essential foundation for multi-instrument laboratories and multi-site studies.

The SlideStream workflow manager integrates the scanner into a seamless workflow through a single, simple, audit-ready

software interface. Using the slide label information, such as text recognition or QR code-based slide identification together with LIMS integration, it automates scan setup, metadata capture, and data routing to image analysis and management systems, reducing hands-on time by approximately 80% compared with traditional manual workflows. It also supports automation of cyclic staining and scanning protocols and provides a highly simplified interface for setting up new assay processing workflows. For laboratories new to spatial biology, this dramatically lowers the barrier to entry and expert microscopy knowledge is no longer required to generate high-quality multiplexed data. For laboratories already using mIF as a routine method, it further improves efficiency, standardisation and scalability.

The Mindpeak PhenoScout platform closes the loop by delivering cloud-based, fully automated tissue image analysis using their patented Pathology Frontier Model with foundational AI models optimised for mIF and brightfield IHC images. The complete analytical workflow can be performed within an hour rather than taking weeks or months, covering single-cell identification, phenotyping, region detection, spatial relationship exploration and standardised report generation. Built on Mindpeak’s experience with

CE-marked clinical algorithms for brightfield applications and benchmark accuracy that exceeds competing platforms, the system translates complex spatial data sets into actionable, quantitative outputs that researchers and clinicians can interpret immediately.

Together, these four components constitute a robust sample-to-report experience: one workflow; one data standard; one output format – regardless of whether samples originate from a single-site discovery project, multi-site study or a lab with multiple instruments.

The clinical breakthrough demonstrated by TissueCypher is just one example of spatial biology’s growing impact.

Across pharmaceutical and biotech research, the same capability to map cellular interactions, quantify biomarker expression and track treatment response within intact tissue is proving equally transformative – from target validation and mechanism-of-action studies, through pharmacodynamic assessments and toxicology, to clinical trial enrichment and companion diagnostics development. For organisations running multi-site studies under regulatory requirements, a reproducible, scalable, GxP-compatible workflow is not aspirational, it is essential.

CROs play a critical role in enabling pharmaceutical sponsors and biotech companies to run complex projects requiring mIF spatial biology analysis, especially where in-house infrastructure and experience are limited. Concept Life Sciences, a UK-based CRO with strong experience in the spatial biology domain, has decided to strengthen its current offerings in this space by incorporating a full end-to-end spatial biology service around the ZEISS tissue multiplexing workflow, to improve and scale their service offering for biomarker discovery, target validation, mechanism-of-action studies, drug efficacy and pharmacodynamic assessments, and regulatory-grade clinical trial support. Concept’s service portfolio covers the full tissue analysis spectrum from chromogenic and fluorescence-based staining through to AI-based image analysis and quantitative reporting.8

Concept achieved significant improvements to ease of use, turnaround time and reproducibility with the use of the standardised reagent kits, the ZEISS Axioscan 7 spatial biology imaging system, and the SlideStream workflow manager, demonstrated in a rigorous multi-site reproducibility study of the new workflow. The three sites, including Concept, were able to achieve excellent concordance across all measured cell populations and across multiple tissue types, expressed by coefficient of variation values significantly below the 20% acceptance threshold.8 This high degree of inter-site reproducibility profoundly impacts clinical trial success: spatial biology services that can map disease and therapeutic response at the single-cell level – from biomarker discovery through therapeutic assessment – are increasingly essential tools for sponsors seeking to understand the tumour microenvironment in trials involving immunotherapy, targeted agents, and cell and gene therapies. Reproducibility across sites is a prerequisite for pooling data, supporting regulatory submissions and advancing drugs through the development pipeline with confidence.

The story of spatial biology over the past five years is one of rapid maturation from an exciting but operationally challenging research tool to a clinically and commercially viable platform with demonstrated impact across oncology, immunology and beyond. The defining characteristic of that maturation has not been any single technological breakthrough, but rather the integration of standardised reagents, high-throughput automated imaging, intelligent workflow management and AI-driven analysis into coherent, reproducible, end-to-end workflows.

For biotech, pharma and CROs, the practical implication is clear. Spatial biology’s power to decode the tumour microenvironment, stratify patients, validate targets and support clinical trial endpoints is now accessible without deep specialist expertise, scalable to large cohorts, and reproducible across sites and time. The infrastructure exists. The clinical proofs are growing, as shown by Castle Biosciences’ prognostic test designed to predict the risk of progressing to oesophageal cancer in patients with Barrett’s oesophagus and Concept Life Sciences’ multi-site CRO studies meeting regulatory-grade reproducibility criteria. What remains is for organisations to integrate these capabilities into their research and development strategies, and to do so with the confidence that the workflow will perform as reliably in their hands as it has in the hands of those who pioneered it.

1. Visit: nature.com/articles/s41592-020-01033-y

2. Visit: pubmed.ncbi.nlm.nih.gov/39643689/

3. Visit: weforum.org/publications/top-10-emergingtechnologies-of-2023/

4. Visit: genengnews.com/topics/cancer/spatial-biology/

5. Visit: pubmed.ncbi.nlm.nih.gov/37622544/

6. Visit: castlebiosciences.com/tests/prognostic/tissuecypher/how-it-works

7. Visit: zeiss.com/microscopy/en/applications/lifesciences/pharmaceutical-and-biotech-industry/spatialbiology.html

8. Visit: conceptlifesciences.com/services/biology/spatialbiology

Sherry Derakhshani PhD, completed her doctoral and postdoctoral training in Cell Biology, with a focus on immunology and 3D tissue model development. As global product marketing and applications manager – Biopharma Solutions at ZEISS Microscopy, she bridges scientific expertise with commercial strategy, helping biotech and pharma organisations translate complex imaging-based workflows into scalable solutions for translational and clinical research. With over a decade of experience spanning academia and industry, Sherry is passionate about accelerating the path from biological insight to patient impact.