Guides for diagnostic development
The label depends on what results will be used for and who runs the test. RUO is for research, never patient care. GLP is for nonclinical safety studies that support an FDA application. An LDT lets one CLIA-certified lab report patient results. An IVD is FDA-cleared or approved and can be sold to, or run by, other labs or consumers.
At a glance
Research
Results used for
Research only; no patient management
Who runs it
Any lab, for research
Oversight
None for the use; FDA polices RUO labeling
Typical use
Discovery, biomarker work, assay development
Nonclinical
Results used for
Nonclinical safety studies submitted to FDA
Who runs it
A GLP testing facility with a study director and independent QA unit
Oversight
FDA GLP regulations (21 CFR Part 58) and FDA inspection
Typical use
Bioanalysis and biomarkers in toxicology studies for an IND
Clinical
Results used for
Patient care, reported by the lab that developed it
Who runs it
One CLIA-certified lab
Oversight
CLIA (CMS), CAP accreditation, state rules such as New York's
Typical use
Specialty testing, early clinical launch
Product
Results used for
Patient care, wherever the cleared test is run
Who runs it
Any lab certified for its CLIA complexity; only the named lab if site-specific; consumers if cleared for home use
Oversight
FDA (510(k), De Novo or PMA), plus CLIA for the lab
Typical use
Kits, home tests, companion diagnostics
RUO is the right label for discovery work: finding biomarkers, building and optimizing an assay, or running exploratory samples. An RUO product carries no clinical claims and FDA does not review it. The common mistake is using RUO results to guide a patient's care: FDA treats RUO-labeled products marketed for clinical use as misbranded. Careful optimization and early precision work at this stage can carry forward into later validation.
GLP applies when an assay produces data for a nonclinical laboratory study, such as drug levels or biomarkers in a toxicology study supporting an IND. The rules are in 21 CFR Part 58. GLP governs how the study is run, not the test's label: a written protocol and single study director, an independent quality assurance unit, a method validated before study samples are run, SOPs and archived raw data, and readiness for FDA inspection. It does not cover human samples from clinical trials (GCLP) or patient testing (CLIA).
An LDT is designed, validated and run within a single CLIA-certified laboratory, the fastest route to reporting patient results. CLIA, administered by CMS, certifies the lab and inspects its validation; many labs also hold CAP accreditation, and New York's Clinical Laboratory Evaluation Program reviews tests offered to New York patients. FDA's 2024 rule to regulate most LDTs as devices was vacated by a federal court on March 31, 2025; FDA did not appeal and has rescinded the rule. Legislation and CMS action on CLIA were pending as of mid-2026, so build validations that would stand up to FDA-level review anyway.
You need IVD authorization to sell a kit to other labs, offer a test direct to consumers, or support a drug label as a companion diagnostic. Moderate-risk tests with a predicate use a 510(k); new moderate-risk tests use De Novo; high-risk tests, including most companion diagnostics, need a PMA. A clearance can be site-specific, covering the test as run in one named laboratory, as in our GlutenID case study.
How to choose
Sources: FDA RUO/IUO guidance · 21 CFR Part 58 · FDA Law Blog, April 2025 · Arnold & Porter, July 2026
We develop and validate assays at every level, run LDTs in our CLIA/CAP laboratory, and prepare FDA submissions.