Biomarker Cutoff Harmonization: A Practical Guide for Labs and Trials

Biomarker cutoff harmonization is the work of making sure that one result means the same thing, whichever laboratory, platform or trial site produced it. Done well, a value of 4.2 at one site leads to the same clinical decision as a 4.2 at another. Done poorly, the same patient can be “positive” in one lab and “negative” in another, and a trial can enroll or exclude people for reasons that have nothing to do with their biology.

This guide is for physicians, practice managers, laboratory directors and clinical operations teams who send samples to more than one lab, or who run studies across several sites. It covers what harmonization includes, where cutoffs drift, the checks to run before two labs share an interval, and what to ask a laboratory partner.

TL;DR

• Harmonization means results are comparable regardless of the measurement procedure used and where or when the test was done. It covers the sample, the measurement and the report, not only the instrument.
• Two labs on the same platform can still disagree clinically if they use different reference intervals or cutoffs.
• Before sharing an interval, verify it locally. A common approach uses 20 reference individuals; too many outliers means building a site-specific interval.
• In multi-site studies, lock down the assay and the cutoff rules before the first patient is enrolled, and revisit static eligibility limits as evidence grows.
• Ask any lab partner how it documents cutoffs, calibrators and changes, and how it delivers results to your system.

What harmonization actually covers

A cutoff is the value that separates one clinical category from another: normal and abnormal, low risk and high risk, eligible and ineligible. A reference interval is related but not identical: it describes the range expected in a defined healthy population. Both depend on how the sample was collected, how it was measured and how the result is reported.

The Harmonization: the Sample, the Measurement, and the Report review makes this three-part view explicit, spanning pre-analytical, analytical and post-analytical steps. That framing is useful for a practice manager, because it shows where to look when two labs disagree. Often the problem is not the analyzer.

The National Cancer Institute describes the goal of assay harmonization as making clinical laboratory results comparable even when test conditions differ. It also notes that variability comes from independent protocol development and different measurement systems, which is exactly what happens when a practice or a sponsor uses several labs.

Harmonization is not the same as standardization

Standardization usually means traceability to a reference method or material, so that every measurement lands on the same scale. Harmonization is the broader, more practical goal: comparable results even where a true reference standard does not exist. Alzheimer disease cerebrospinal fluid markers are a well-known example, where researchers had to align assay platforms without an absolute reference standard.

Where cutoffs drift between laboratories

Most disagreements come from a short list of causes. Knowing them lets you ask better questions.

• Different platforms. Even with traceable calibrators, LC-MS/MS and immunoassay methods can give fundamentally different results, so a simple transfer of one lab’s interval to another is not always valid. Platform-specific intervals may be required.
• Different reference intervals on identical platforms. The review of harmonization points out that labs often invest heavily in verifying the analytical method while paying less attention to choosing evidence-based reference intervals, yet the variation between intervals can be larger than the analytical error of the measurement itself.
• Different populations. Age, sex and, for some hormones, menstrual cycle phase change what is normal. An interval built on one population may not fit your patients.
• Different pre-analytical handling. Collection tube, time to centrifugation, temperature in transit and storage all affect some analytes. A specimen courier schedule is part of harmonization, not a separate topic.
• Different reporting. Units, decimal places and flag logic (for example, a result shown as high versus a numeric value only) change what the clinician sees.

A worked example: two labs, one result

Imagine a primary care group that sends most work to one laboratory and urgent or specialized orders to another. A patient’s result comes back at the upper edge of one lab’s interval and is flagged high. At the second lab, with a slightly wider interval, the same value carries no flag. The clinician sees two different messages about one number. Nothing is wrong with either instrument. The harmonization gap is in the interval, and it changes whether the patient gets a repeat test, a referral or reassurance.

The practical fix is to compare the two labs’ intervals and methods side by side, decide which one fits your patient population, and document the decision. Our own walkthrough, Reference Range Harmonization: 4 Checks Before Labs Share an Interval, covers that comparison in more detail.

Four checks before two labs share an interval

Use this sequence whenever you plan to adopt another lab’s interval or merge data from several sites.

1. Compare methods and calibration. Confirm the platform, the principle of measurement and the calibrator traceability. If the methods differ in kind (for example mass spectrometry versus immunoassay), expect that the interval may not transfer.
2. Compare the source population. Check how the original interval was derived: who was included, how many, and whether results were split by age or sex.
3. Verify locally. A published approach to reference interval transference uses 20 reference individuals at the receiving site. If more than 3 of them fall outside the limits, the laboratory should establish its own site-specific interval instead of adapting the existing one.
4. Build from scratch when you must. When a new interval is required, especially for markers that need stratification by age, sex or menstrual cycle, standard practice calls for a minimum of 120 participants per subgroup to keep the estimate statistically reliable.

The numbers matter for planning. A stratified interval with four subgroups needs at least 480 well-characterized reference individuals, which is a real project, not an afternoon task. That is one reason many practices choose to verify and adopt an interval rather than build one.

Harmonizing cutoffs in multi-site clinical trials

In a trial, the cutoff is often the protocol. It decides eligibility, dose adjustment, safety stopping and, for biomarker-selected studies, which patients count as positive.

Lock the assay before the first patient

Guidance on integral assays in trials states that analytical and clinical validation should be completed before patient accrual begins, so the assay’s performance is locked down and results are reliable. A cutoff that is chosen or changed mid-study is hard to defend and harder to analyze.

Choose the cutoff on purpose

Many biomarkers are continuous, and turning them into high/low groups is a statistical decision as well as a clinical one. A review of methods for determining cutoff values in clinical studies shows how classification into risk groups depends on the method used to select the threshold. Pre-specify the method and the cutoff in the statistical analysis plan, and record which assay version produced each result.

Align results across labs, or test centrally

There are two broad ways to keep multi-site data comparable. One is to route specimens to a single central laboratory. The other is to let sites use local labs and then adjust the data. A study of ten laboratories found that mathematical adjustment using commutable samples produced more comparable results, though it requires suitable samples and added analysis. Central testing removes much of that burden, at the cost of shipping logistics.

Oncology offers a large-scale example. Friends of Cancer Research’s tumor mutational burden project had eleven labs take part in Phase 2, found variability between panel and whole-exome values, and developed calibration approaches to reduce the spread between tests. The lesson applies far beyond oncology: the same biomarker name does not guarantee the same number.

Do not let eligibility limits go stale

The ASCO-Friends working group found that laboratory eligibility requirements often stay static and can be unnecessarily strict, delaying enrollment and narrowing who the results apply to. When you set or copy lab-based eligibility thresholds, ask whether they reflect current clinical experience and whether a lab’s reference range, rather than a fixed number, is the better anchor.

If you are earlier in the process, our guide to lab data harmonization lays out three first steps for labs and research teams.

Practical cases outside of trials

Harmonization is not only a research issue. Several everyday situations depend on it.

• Trending a patient over time. If a patient’s labs move between facilities, a change in the result may reflect a change in method or interval, not a change in health. Bone turnover markers such as PINP and β-CTX-I are monitored over months, so a consistent assay and interval matter when clinicians compare serial values.
• Skilled nursing and home health. Residents and patients often move between a facility, a hospital and home. Using one laboratory for routine orders, with consistent reporting, reduces conflicting flags.
• Telehealth and wellness programs. Programs that draw samples across many locations need the same cutoffs applied everywhere, plus consistent collection and transport.
• Imaging and blood together. When a blood biomarker and an imaging finding are used together to triage, each has its own threshold. See Imaging vs Blood Biomarkers: Two Thresholds Shape Clinical Triage for how those thresholds interact.

Our perspective

We see harmonization as a service problem as much as a statistical one. Most disagreements between labs are found late, when a clinician or study coordinator notices two results that do not fit. By then the cost is a repeat draw, a delayed enrollment or a conversation with a patient that did not need to happen.

Our view is that a laboratory partner should make its cutoffs visible before you ask. That means telling you which platform and reference interval apply to each test, what changes when a method or interval is updated, and how results are formatted when they reach your EHR or study database. We built our work around direct communication and customized laboratory solutions, because a practice with three physicians and a sponsor with twenty sites have different harmonization needs, and a one-size-fits-all report rarely serves either.

For trials, we think the best time to talk about cutoffs is during protocol design, not after the first data transfer. For practices, it is when you add a second laboratory or change what you order. In both cases, the habit that pays off is the same: write down the interval, the method and the reason.

— Kohealth Labs

What to ask a laboratory partner

Whether you are choosing a primary laboratory or adding one for specialized testing, these questions surface harmonization risk early.

1. Which platform and method do you use for each test I order, and has either changed recently?
2. What reference interval or cutoff do you report against, and how was it established or verified?
3. How will you notify us if an interval, calibrator or method changes?
4. Can results be delivered electronically and integrated with our EHR or LIS, with units and flags consistent across orders?
5. How are specimens collected, labeled, stored and transported, and who is accountable if a pre-analytical problem affects a result?
6. For studies: can you support central testing, specimen management and biomarker analysis against the protocol’s thresholds?

A good answer is specific. A vague answer is information too.

How KoHealth Labs supports consistent results

KoHealth Labs is a full-service diagnostic laboratory and clinical research partner. For practices, we offer clinical laboratory testing, molecular and PCR diagnostics, chemistry, microbiology, pathology support, specimen pickup and transportation, mobile phlebotomy, electronic results reporting and LIS integration. For sponsors and CROs, we act as a centralized laboratory partner for specimen collection and management, biomarker analysis and study-specific diagnostic solutions.

If you are planning a study, start with our clinical trials laboratory page or see how it works. If you run a clinic or group practice and want a primary or supplemental lab, our laboratory services overview is the place to begin. For molecular work, see our PCR testing for clinics and our guide to PCR testing for clinical trials.

FAQ

What is biomarker cutoff harmonization?

It is the process of making results and decision thresholds comparable across laboratories, platforms and sites. It addresses sample handling, the measurement itself and how results are reported and interpreted. The aim is that the same patient receives the same clinical message wherever the test is run.

Can I use another laboratory’s reference interval?

Sometimes, but verify it first. One published approach tests 20 reference individuals and calls for a new site-specific interval if more than 3 fall outside the limits. If the other lab uses a different kind of platform, the interval may not transfer at all.

How many people do I need to build a new reference interval?

Standard practice calls for at least 120 participants per subgroup when results must be stratified by factors such as age, sex or menstrual cycle. Several subgroups multiply that number quickly, so many labs verify an existing interval where that is appropriate.

Why do two labs report different results for the same patient?

Common reasons are different measurement platforms, different reference intervals, different pre-analytical handling and different reporting formats. Differences in intervals can be larger than the analytical error itself. Your physician should interpret any discrepancy in the context of the patient’s care, and your laboratory director can help trace the cause.

When should a trial lock its biomarker cutoff?

Before patient accrual begins. Validation of the assay and the clinical threshold should be complete so that results across the study are reliable and analyzable.

Sources

• Harmonization: the Sample, the Measurement, and the Report (PMC)
• Assay Harmonization Activities and Resources (NCI)
• Alzheimer Disease Biomarker Testing in Cerebrospinal Fluid (Clinical Chemistry)
• A Guide to Harmonisation and Standardisation of Measurands Determined by LC-MS/MS (PMC)
• Reference interval transference of common clinical biomarkers (Scandinavian Journal of Clinical and Laboratory Investigation)
• Development and Use of Integral Assays in Clinical Trials (PMC)
• Determination of cutoff values for biomarkers in clinical studies
• Harmonization of laboratory results by data adjustment in multicenter clinical trials
• Friends of Cancer Research: Phase 2 of the TMB Harmonization Project
• Modernizing Clinical Trial Eligibility Criteria: ASCO-Friends Work Group (PMC)

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.