Assay Validation in Clinical Trials: A Practical Guide

Assay validation in a clinical trial is the documented proof that a laboratory method measures what you need it to measure, reliably enough to support the decision the data will inform. The standard depends on that decision. Data that supports dosing, safety or labeling needs full validation. Data used for internal go/no-go calls can be validated more flexibly. Either way, the work should be finished and written down before the first participant’s sample is run.

This guide is for clinical operations leads, study directors and lab vendor managers. It covers what regulators expect, which performance characteristics to test, where trials most often go wrong, and how to judge a central or specialty lab.

TL;DR

• Validate for the intended use: full validation is expected when biomarker data supports pivotal regulatory decisions, and a lighter, justified approach can fit exploratory use.
• Test accuracy, precision, sensitivity, selectivity, linearity, range and reproducibility, and include pre-analytical conditions such as collection, storage and shipping.
• Complete validation before enrollment. If rare samples limit you, use contrived samples first and confirm with early-trial specimens.
• Testing on human specimens in the U.S. falls under CLIA, so the lab’s certification and its validation records matter as much as the method.
• Centralizing testing with one qualified lab reduces site-to-site variability.

What assay validation means in a trial setting

An assay is a laboratory procedure that measures an analyte: a drug concentration, a protein, a genetic variant, a viral load. Validation is the structured set of experiments showing the procedure performs as claimed. It is not the same as verification. Verification confirms that a lab can reproduce the performance of an already-cleared method. Validation establishes performance for a method that is new, modified or used outside its original claim.

Three questions frame every plan:

1. What decision will this result drive? Eligibility, dose escalation, endpoint, safety signal or exploratory research.
2. Who will rely on it? A sponsor’s internal team, an IRB, or a regulator reviewing a submission.
3. What happens if the result is wrong? A misclassified participant or a false safety signal costs far more than a noisy exploratory marker.

The answers set the depth of validation. The literature on integral assays makes the same point: when a biomarker determines eligibility, treatment assignment or outcome, the assay must be robust before it is used.

What the FDA expects: fit for purpose, anchored in ICH M10

On 21 January 2025 the FDA issued its final guidance on bioanalytical method validation for biomarkers, which replaced the 2018 bioanalytical method validation guidance for these assays. Two ideas run through it.

Fit-for-purpose validation. When biomarker data supports pivotal regulatory decisions on safety, efficacy, labeling or dosing, sponsors should apply full validation aligned with ICH M10. When an assay supports only internal decisions, such as candidate selection or proof of concept, the extent of validation is left to the sponsor to determine and justify.

ICH M10 as a starting point, not a template. The ICH M10 guidance covers bioanalytical method validation and study sample analysis for assays that generate data supporting regulatory submissions. It was written around drug concentration assays. Biomarker platforms differ, so sponsors must document where an M10 criterion does not apply and why an alternative characteristic is more appropriate.

The practical consequence: write the justification down. A reviewer rarely objects to a thoughtful deviation. They object to an unexplained one.

The same principle appears in a December 2025 discussion of defining what “enough” validation looks like: the higher the consequence of the decision, such as patient selection or dose justification, the more completely teams validate key performance characteristics and pre-analytical requirements.

Research use only is not a shortcut

Reagents labeled for research use only are common in early assay development, but that label describes the product’s regulatory status, not the performance of your assay. If you are weighing RUO kits, read the FDA signals on research use only materials so that the documentation in your validation package reflects how the material is actually regulated.

The performance characteristics to test

The FDA’s guidance for industry on circulating tumor DNA states that assays should be analytically validated with sufficient performance characteristics before use in clinical trials. Although that document is oncology-specific, its logic applies broadly. Validation must be pre-planned in a technical protocol, and every acceptance criterion needs a justification tied to clinical use.

Validation should cover the whole system, from the collection tube to the assay output, not only the instrument run.

A marker such as bone turnover markers shows why. Small shifts across a few months can matter clinically, so an assay must be precise enough to separate true biological change from analytical noise. Precision and reproducibility limits should be set before you interpret change over time.

Pre-analytical variables decide more results than you expect

Many trial teams spend months on method performance and a few days on how samples reach the lab. That is backwards. A 2025 study of a phase 1 dose-finding trial found that improperly filled sample tubes prevented complete thawing and mixing and produced recovery deviations of up to 30% from nominal values, which showed up as aberrant pharmacokinetic profiles.

Build these into the protocol and the site manual:

• Tube type, fill volume and mixing instructions
• Time from draw to centrifugation, and centrifuge settings
• Aliquot volume and labeling
• Freezer temperature and maximum storage time
• Thaw procedure before analysis
• Shipping conditions, courier hand-offs and temperature logging

Stability data should back each limit. Our guide to sample stability guidelines that hold up in an audit walks through how to document them.

Validate before enrollment, and plan for scarce samples

Expert consensus on biomarker test validation is that it should finish before a trial begins. The goal is to show the result is reliable enough to keep the risks of its clinical use acceptable. If the target population is rare and positive samples are scarce, the suggested route is to validate with contrived samples first, then run a feasibility study on early-trial specimens to confirm the assay is adequate.

A workable sequence:

1. Define intended use and the decision the data drives.
2. Write the validation plan with acceptance criteria and their rationale.
3. Run analytical validation, including pre-analytical stability.
4. Freeze the method and lock the standard operating procedure.
5. Train sites on collection and shipping, then run a small pilot shipment.
6. Monitor performance during the study with quality controls and proficiency checks.
7. Document any change and assess whether partial or full revalidation is needed.

Special case: sequencing-based assays

Next-generation sequencing adds complexity. Published work on regulated NGS trial assays notes that current NGS technologies have higher error rates and novel error modes than traditional sequencing, which can cause variability in mutation reporting. That is why NGS validation needs to cover the full workflow, including library preparation, bioinformatics and reporting thresholds, not just the sequencer.

CLIA: who is allowed to validate and run the test

In the United States, laboratory testing on human specimens that generates results for clinical use falls under the Clinical Laboratory Improvement Amendments. NIH materials on IVDs, LDTs, FDA and CLIA describe how CLIA covers preanalytic, analytic and postanalytic systems, including verification and validation of methods. Tests are categorized as waived, moderate or high complexity, and labs show compliance through a Certificate of Compliance or Certificate of Accreditation.

Two practical points follow:

• Moderate and high complexity tests must be validated by qualified laboratory personnel before use. UCSF’s research guidance states that unlicensed research staff cannot perform these validations, however skilled they are.
• For laboratory-developed methods, the lab must establish and document performance specifications such as accuracy, precision, reportable range, reference interval and analytical sensitivity and specificity, as described for molecular assays.

If your study returns results that influence patient care, CLIA status is not optional. Building your own certified lab can take months, so many sponsors contract one instead. We cover that tradeoff in avoiding six month CLIA buildouts and explain what to check on a certificate in our CLIA certified laboratory guide.

Where trials most often go wrong

Using local hospital labs for pivotal data. Different sites with different methods produce discordant results. Published analysis of NGS trial assays points out that local labs often lack the validation protocols and specialized expertise complex assays need. Central testing keeps method, reagents and interpretation uniform.

Validating late. Discovering during enrollment that an assay misses its sensitivity target is expensive and can compromise data already collected.

Unjustified acceptance criteria. Borrowing limits from another assay without explaining why they suit this one invites questions.

Ignoring the sample journey. Courier delays, mislabeled tubes and temperature excursions erase good method performance.

Changing the method mid-study. A new reagent lot, instrument or software version needs a documented bridging assessment.

How to evaluate a lab for assay validation

Ask any central or specialty lab for specifics, not reassurance:

• CLIA certificate type and the list of tests it covers
• A sample validation report, with protocol, acceptance criteria and deviations
• How it handles ICH M10 alignment for biomarker work and documents justified departures
• Stability data for your analyte and matrix
• Specimen logistics: kit design, courier coverage, receipt checks, chain of custody
• Data delivery: formats, turnaround, and whether results reach your systems electronically
• Change control and who you speak with when something goes wrong

The last point is where service model matters. Study teams tell us they want a named contact who answers questions about a sample on the same day, not a ticket queue.

Our perspective

We see validation problems less often in the method itself than in the handoffs around it: a tube filled halfway at a site, a courier pickup missed on a Friday, a result delivered in a format the data team cannot load. That shapes how we work with research sponsors and CROs.

KoHealth Labs serves as a centralized laboratory and supplier partner for clinical trials, supporting laboratory testing, specimen collection and management, and biomarker analysis designed around study-specific requirements. We pair that with specimen pickup, electronic results reporting and LIS integration so that the sample journey and the data journey are planned together with the assay. Our view is simple: agree intended use first, document the acceptance criteria, treat pre-analytical steps as part of the method, and keep one accountable contact from onboarding to final data.

— Kohealth Labs

Working with a central lab on your next study

If you are planning a study with biomarker or molecular endpoints, bring us the protocol early. We can discuss specimen collection, courier pickup, test panels and data delivery before you lock your lab manual. See how our clinical trials laboratory services work, or read how the onboarding process runs. For studies built around molecular endpoints, our PCR testing page outlines that capability. Our primer on the role of biomarker panels in clinical research is a good companion read for endpoint selection.

FAQ

What is the difference between assay verification and assay validation?

Verification confirms that a laboratory can achieve the performance already established for an existing, cleared method. Validation establishes performance for a method that is new, modified or used differently from its original claim. Trials with novel biomarkers or laboratory-developed methods typically need validation.

Which regulatory standard applies to clinical trial sample analysis?

For bioanalytical methods supporting regulatory submissions, the FDA recognizes the ICH M10 guidance. Biomarker assays are covered by the FDA’s January 2025 biomarker guidance, which uses M10 as a starting point. In the U.S., the testing laboratory must also meet CLIA requirements when results are used for clinical purposes.

Does every assay in a trial need full validation?

No. The extent depends on the assay’s context of use. Assays that support pivotal decisions on safety, efficacy, labeling or dosing need full validation, while assays for internal decisions allow a justified, sponsor-determined approach. Document the reasoning either way.

Can validation happen after the trial starts?

Experts advise completing it before the trial begins. When rare samples make that hard, validate with contrived samples first and then confirm adequacy with a feasibility study on early-trial specimens.

Sources

• FDA bioanalytical method validation for biomarkers, 2025 commentary
• FDA: M10 Bioanalytical Method Validation and Study Sample Analysis
• R&D World: Defining what “enough” biomarker validation looks like
• Development and Use of Integral Assays in Clinical Trials
• FDA guidance for industry on ctDNA assays
• Aberrant pharmacokinetics and sample processing in immunoassays
• Expert consensus on biomarker test validation
• Regulated NGS-based clinical trial assays and companion diagnostics
• IVDs, LDTs, FDA and CLIA (NIH seminar)
• UCSF Clinical Research Resource HUB: Clinical Laboratory Testing
• Validation of Laboratory-Developed Molecular Assays for Infectious Diseases

For more information about our diagnostic testing, clinical research laboratory services, or partnership opportunities, visit https://kohealthlabs.com or contact KoHealth Labs at (469) 792-8439 and ask for Amy Patterson.