How is spatial biology used in clinical trials?

Most commonly as an exploratory endpoint, generating data reviewed after the trial without informing enrollment or treatment decisions in real time. As evidence accumulates, a spatial signature can graduate toward informing patient stratification or enrichment, and spatial data can also validate computational models that simulate trial outcomes, a distinct application from direct biomarker use.

Can spatial assays be trial endpoints?

Yes, most commonly as exploratory endpoints early in a program, which carry a lower validation burden than a primary or decision-driving endpoint since they do not yet inform which patients enroll or how the trial’s main result is interpreted. A spatial endpoint can graduate toward a stratification or enrichment role once it demonstrates consistent, biologically meaningful behavior.

What is a spatial biomarker in a trial?

A biomarker defined by spatially resolved measurements, such as cell proximity or tissue architecture, used within a clinical trial at varying levels of formality, from a purely exploratory endpoint collected for hypothesis generation to a validated signature used to stratify or enrich enrollment. The appropriate level of validation and standardization scales with how much decision-making weight the biomarker is expected to carry.

Spatial biology in clinical trials

Highlights in this article

Where spatial fits in a trial
Exploratory spatial endpoints
Stratification and enrichment
Mechanism-of-response readouts
Operational and standardization challenges

Building spatial biology into a clinical trial protocol is rarely an all-or-nothing decision. A spatial assay typically enters a trial quietly, as an exploratory endpoint collecting data without yet informing any real-time decision, well before it is trusted enough to help decide which patients enroll or which arm they are assigned to. Understanding that progression, and where a given spatial assay actually sits along it, matters more for a trial-design decision than a general enthusiasm for spatial data.


Key takeaways The utility of spatial biology data in accelerating drug programs requires balancing its exploratory, hypothesis-generating value against the practical costs of building it into a trial protocol formally. Most spatial assays enter a trial as exploratory endpoints, collecting data without yet informing enrollment or treatment decisions in real time. Spatial data can validate and calibrate computational pharmacology models capable of simulating trial outcomes, a distinct application from direct biomarker discovery. The full trial-design framework for using a validated spatial signature to stratify or enrich enrollment is a separate, later-stage question covered elsewhere in this cluster. Operational and standardization requirements for a spatial endpoint scale with how much weight the endpoint is expected to bear, from purely exploratory to decision-driving.

Key takeaways

The utility of spatial biology data in accelerating drug programs requires balancing its exploratory, hypothesis-generating value against the practical costs of building it into a trial protocol formally.
Most spatial assays enter a trial as exploratory endpoints, collecting data without yet informing enrollment or treatment decisions in real time.
Spatial data can validate and calibrate computational pharmacology models capable of simulating trial outcomes, a distinct application from direct biomarker discovery.
The full trial-design framework for using a validated spatial signature to stratify or enrich enrollment is a separate, later-stage question covered elsewhere in this cluster.
Operational and standardization requirements for a spatial endpoint scale with how much weight the endpoint is expected to bear, from purely exploratory to decision-driving.

Where spatial fits in a trial

A 2024 Cancer Discovery piece frames the central tension a trial sponsor actually faces: the utility of spatial biology data in accelerating drug programs requires balancing its exploratory value, the hypotheses it can generate about mechanism, resistance and patient selection, against the practical costs of collecting and analyzing it within a real trial, additional tissue burden, analysis time, and the standardization work covered elsewhere in this cluster’s biomarker development guides. That framing, balancing rather than assuming unambiguous benefit, is the honest starting point for deciding where in a specific trial a spatial assay actually belongs.

In practice, that balance plays out along a progression. A spatial assay typically starts as a purely exploratory endpoint, generating data reviewed after the trial without influencing any decision made during it. As evidence accumulates that the assay’s output is reproducible and biologically meaningful, it can graduate toward informing stratification or enrichment decisions, the trial-design questions covered in full in Predictive Spatial Signatures for Patient Stratification. This spoke concentrates on the earlier, more common stage: what a spatial endpoint is actually doing in a trial before it has earned that later, more consequential role.

Exploratory spatial endpoints

An exploratory endpoint’s job is specifically not to bear the statistical or regulatory weight a primary or secondary endpoint carries. It exists to generate a hypothesis, characterize a mechanism, or establish that a specific spatial measurement behaves consistently enough across patients to be worth developing further, all without that data yet informing which patients enroll or how the trial’s primary result is interpreted.

That lower bar carries a specific, practical advantage worth stating plainly.

Lower validation burden than a primary or decision-driving endpoint. An exploratory spatial assay does not need the same locked-down, cross-site harmonized protocol a companion diagnostic eventually requires, since it is not yet the basis for a clinical decision.
Room to iterate on the assay itself during the trial. Because an exploratory endpoint’s results are not feeding a real-time decision, a sponsor has more latitude to refine the spatial method between patient cohorts than a locked, decision-driving assay would allow.
A natural proving ground before a heavier investment. Demonstrating that a spatial signature behaves consistently as an exploratory endpoint is precisely the evidence a program needs before committing to the harmonization and validation work a stratification-grade or companion-diagnostic-grade assay requires.

Stratification and enrichment

Once an exploratory spatial signature has demonstrated consistent, biologically meaningful behavior, the next question, whether and how to build it into enrollment or randomization decisions, is a substantial trial-design question in its own right, addressed in full detail elsewhere in this cluster rather than repeated here.

The full framework, including the standard trial design categories a biomarker can be built into and the specific conditions that justify prospective, decision-driving use over a simpler exploratory or retrospective approach, is developed in Predictive Spatial Signatures for Patient Stratification. What is worth stating here is only the connecting point: a spatial signature graduates from this section’s exploratory role into that framework once it has accumulated the kind of reproducible, mechanistically grounded evidence exploratory use is specifically designed to generate.

Mechanism-of-response readouts

Beyond biomarker discovery and patient selection, spatial data serves a third, distinct trial application: grounding and validating computational models that simulate trial outcomes, a use case that treats spatial measurements as inputs to a predictive pharmacology model rather than as a biomarker read directly off a patient.


Spatial data calibrating a virtual clinical trial A described modeling effort developed a spatial quantitative systems pharmacology model to simulate tumor-immune dynamics at the organ scale while capturing a tumor’s spatial heterogeneity, aimed at supporting dosing regimen design and biomarker identification for combination immunotherapy. The model was validated against spatial multi-omics data from a neoadjuvant hepatocellular carcinoma trial combining a checkpoint inhibitor with a multitargeted tyrosine kinase inhibitor. Imaging mass cytometry data showed that closer physical proximity between CD8-positive T cells and macrophages correlated with non-response, and that spatial proteomics-derived relationship was used to validate and calibrate the computational model, which was further compared against spatial transcriptomics profiling of post-treatment samples. This is a genuinely different use of trial-derived spatial data than direct biomarker discovery: rather than reporting a spatial signature as a candidate biomarker in its own right, the spatial measurements serve as ground truth that calibrates a computational model capable of simulating dosing regimens and outcomes across a virtual patient population, potentially reducing how much a program needs to learn through further live-patient iteration alone.

Spatial data calibrating a virtual clinical trial

A described modeling effort developed a spatial quantitative systems pharmacology model to simulate tumor-immune dynamics at the organ scale while capturing a tumor’s spatial heterogeneity, aimed at supporting dosing regimen design and biomarker identification for combination immunotherapy. The model was validated against spatial multi-omics data from a neoadjuvant hepatocellular carcinoma trial combining a checkpoint inhibitor with a multitargeted tyrosine kinase inhibitor. Imaging mass cytometry data showed that closer physical proximity between CD8-positive T cells and macrophages correlated with non-response, and that spatial proteomics-derived relationship was used to validate and calibrate the computational model, which was further compared against spatial transcriptomics profiling of post-treatment samples.

This is a genuinely different use of trial-derived spatial data than direct biomarker discovery: rather than reporting a spatial signature as a candidate biomarker in its own right, the spatial measurements serve as ground truth that calibrates a computational model capable of simulating dosing regimens and outcomes across a virtual patient population, potentially reducing how much a program needs to learn through further live-patient iteration alone.

Operational and standardization challenges

The operational burden a spatial endpoint imposes scales directly with how much decision-making weight it is expected to carry, which is precisely why matching that burden to the endpoint’s actual role, rather than over-engineering an exploratory measurement or under-validating a decision-driving one, is a genuine trial-design skill.

Three practical considerations recur across trials incorporating spatial endpoints at any stage.

Tissue budget has to be planned across every planned use, not just the primary endpoint. A trial collecting biopsies for a spatial exploratory endpoint alongside standard pathology and any companion diagnostic assay needs a tissue collection plan that accounts for all three uses from the protocol’s design stage, not added as an afterthought once the primary endpoint’s tissue needs are already set.
Analysis turnaround time should match the endpoint’s actual role. An exploratory endpoint analyzed after database lock can tolerate considerably longer turnaround than a stratification-grade assay expected to inform a real-time enrollment decision, and over-investing in rapid turnaround for a purely exploratory measurement wastes resources a decision-driving assay would need more.
Standardization investment should track the endpoint’s intended graduation path. A spatial measurement intended only to generate a publication-worthy hypothesis does not need the cross-site harmonization rigor covered in this cluster’s biomarker development guides; one intended to eventually support regulatory decisions does, and planning that investment early avoids redoing validation work later under time pressure.

The detailed harmonization and validation standards a spatial assay needs once it moves beyond exploratory use are covered in full in Spatial Biomarkers and Companion Diagnostics: The Next Frontier.

For the broader convergence argument and clinical infrastructure this spoke builds on, see Translating Spatial Biology into the Clinic: Digital Pathology and Beyond, and for where this fits within the full spatial biology pipeline, Spatial Biology in Drug Discovery: From Target Discovery to Translational Medicine.

This article was produced under Drug Discovery News’s AI editorial policies.