Is PGx Interpretation Software FDA-Cleared? Vendor vs. Lab Liability in 2026

No—most PGx interpretation software is not FDA-cleared, and that is by design. Whether PGx interpretation software is FDA-cleared depends on where it sits: interpretation and reporting platforms run downstream of variant calling, ingesting already-called genotypes under your CLIA lab's medical director authority. The director reviews and signs out every report, keeping the lab—not the vendor—as final decision-maker.

Short answer: no, most PGx interpretation software is not FDA-cleared—and here is why

The question "is PGx interpretation software FDA-cleared?" usually rests on a category mismatch. FDA clearance—typically a 510(k)—applies to medical devices, including in vitro diagnostic tests that generate a clinical result from a patient specimen. A pharmacogenomic interpretation platform is not that. It never touches a specimen, aligns a read, or calls a variant. It ingests genotypes your assay already produced and maps them to published guideline recommendations for your medical director to review.

That places most interpretation software outside the diagnostic-device pathway and inside the clinical decision-support and reporting layer that CLIA laboratories have always operated. The absence of a clearance number is therefore not a gap to explain away—it reflects the regulatory category the software belongs to. Much software that surfaces sourced evidence for a qualified professional to weigh, rather than driving a result on its own, has historically fallen under clinical-decision-support considerations rather than the device pathway.

For your lab, the practical takeaway is that a clearance number is not the right yardstick. What matters is whether the tool is transparent about its logic, validated in your own hands, traceable to cited sources, and structured so that your director remains the person who releases every report.

What happened to the FDA LDT rule in 2025 and 2026?

The regulatory ground shifted twice in two years, and lab directors should understand the sequence before drawing conclusions about their own obligations.

On May 6, 2024, the FDA issued a final rule asserting authority to regulate laboratory-developed tests (LDTs) as medical devices, with compliance phased in through 2028. That rule would have pulled many lab-built tests—potentially including some PGx assays—into the device framework, layering registration, listing, and premarket expectations onto tests that had long been overseen under CLIA.

It did not survive. On March 31, 2025, the U.S. District Court for the Eastern District of Texas, in ACLA v. FDA (consolidated with AMP v. FDA), vacated the rule nationwide, holding that FDA lacked statutory authority to regulate LDTs as devices. HHS and FDA did not appeal, and on September 19, 2025, FDA issued a further final rule formally rescinding the 2024 framework and reverting the regulatory text to its pre-2024 state.

As of August 2026, LDTs are governed primarily by CLIA (CMS), not FDA's vacated device rule, with FDA retaining its narrower historical enforcement-discretion posture. This is not settled ground: Congress is weighing new legislation—including the Enhancing CLIA Act of 2026 and a reintroduced VALID Act—that could change the picture again, as regulatory trackers such as Arnold & Porter's one-year review have documented. Treat this section as background, not legal advice, and confirm your obligations with counsel.

How is PGx interpretation software different from FDA-cleared diagnostics?

The dividing line is variant calling. An FDA-cleared pharmacogenomic diagnostic is an end-to-end assay: it accepts a specimen, runs the chemistry, aligns or genotypes the sample, and calls star alleles or variants. That combined wet-lab and bioinformatics pipeline is the regulated device, and the FDA's LDT and IVD guidance is written around that kind of result-generating system.

Interpretation software begins after that work is finished. SignalPGx, for example, ingests already-called genotypes—VCF, PharmCAT output, Agena MassARRAY, or CSV—and translates diplotypes into phenotype assignments and guideline-based recommendations across 50+ pharmacogenes and 950+ medications. It does not align reads, call variants, or assign star alleles, and it is not a diagnostic test. It is the reporting and interpretation layer that sits on top of whatever calling pipeline your lab has already validated, connecting to it through your existing integrations rather than replacing it.

That downstream position is exactly why the FDA-clearance question lands differently here than it does for an assay. The software's job is to make published evidence consistent, auditable, and reviewable for your director—not to generate the underlying genetic result. Getting that boundary right is the whole point of the platform: the calling stays where your lab already validated it, and the interpretation layer adds structured, cited guidance on top without touching the primary result.

Vendor vs. lab: who is liable when PGx interpretation is wrong?

Under the CLIA model, clinical responsibility rests with the laboratory and its medical director, not with the software vendor. The lab validates the tool in its own environment, the director reviews and signs out each report, and the treating physician makes the actual prescribing decision. The vendor's role is to supply a transparent, deterministic engine and a complete audit trail—not to practice medicine, interpret as a service, or release results. No credible interpretation vendor should position itself as assuming the director's sign-out duty, and no lab should accept a tool that blurs that line.

The Translational Software episode illustrates why the boundary matters in practice. GenomeWeb reported that FDA found the company's PGxPortal "not substantially equivalent" to 23andMe's authorized Personal Genome Service—the predicate cited in its 510(k), first filed in February 2020 with a final submission in October 2022. In October 2023, GenomeWeb and Precision Medicine Online reported that CEO Don Rule announced the company would withdraw the PGxPortal decision-support service by year-end; GenomeWeb reported that roughly 98 customers, mostly labs, then needed to find new reporting options. Reported industry and company commentary attributed the outcome partly to FDA's use of a consumer-facing predicate and to PGxPortal's reliance on CPIC recommendations, which FDA does not formally recognize as an evidentiary standard—context to read as commentary, not a published FDA rejection letter.

The lesson is not that one vendor stumbled. It is that keeping interpretation under your own CLIA authority, on a pipeline you control, reduces single-vendor dependency and keeps accountability where it legally belongs. When you evaluate contracts, look closely at how validation duties, audit access, and indemnification are allocated, and pair that review with a clear internal standard for what a clinically defensible PGx report requires. None of this is legal advice.

CLIA validation: what labs must audit before adopting interpretation software

Because your lab owns the validation, evaluate interpretation software the way you would any component that touches a reportable result. A defensible validation and ongoing audit should, at minimum, cover the following.

• Genotype-to-phenotype logic: confirm that diplotype-to-phenotype assignments and the resulting recommendations match the guideline sources—such as CPIC and DPWG—you intend to follow, across the genes and drugs in your menu.
• Guideline provenance and version: know which knowledge-base version produced each recommendation and how the vendor tracks and communicates updates, so a mid-year guideline change does not silently alter outputs.
• Output accuracy: run known samples and deliberate edge cases—no-calls, rare alleles, ambiguous diplotypes, multi-drug scenarios—and confirm each report matches expected results before you rely on it.
• Determinism: the same input should always produce the same output. Reproducibility is what makes a result defensible if it is ever questioned.
• Audit trail: every recommendation should trace to a cited source and to a logged reviewer action, so you can reconstruct exactly why a report said what it said.

SignalPGx is built to support this work with a deterministic engine, source-level citations spanning 16 evidence sources (including CPIC, DPWG, FDA, DailyMed, PharmVar, and ClinVar/ClinGen), and a complete audit trail. The validation itself, though, is performed by your lab, under your CLIA license—the software fits your workflow rather than certifying it. You can review how the platform handles data integrity, access control, and audit logging on the security page.

CLIA medical director sign-out: the director is the gatekeeper, not the software

No interpretation platform signs out a report. Under CLIA, that authority belongs to your qualified medical director, and well-designed software should reinforce that role rather than erode it. Every SignalPGx report is a draft for human review: the director inspects the called genotypes, the mapped phenotypes, the cited recommendations, and any flags before releasing anything to a clinician. The tool structures and sources the evidence; the director decides.

This is where guardrails earn their place. SignalAI, the platform's reviewer assistant, is built to cite-or-refuse—it surfaces sourced evidence to support the person reviewing a case and declines when it cannot ground an answer in a citation, rather than fabricating one. It never acts autonomously, never overrides the director, and never signs out. The human-in-the-loop is not a marketing phrase here; it is the control that keeps interpretation consistent with how your lab already practices and defensible if a report is later scrutinized.

That model also keeps the downstream boundary intact: the report explains the path from called genotype to guidance, and the director owns each step of the release. If you want to see how that pipeline is structured from intake to reviewed output, the walkthrough of the genotype-to-guidance workflow traces where the software assists and where the director stays firmly in control.

How to evaluate PGx software for compliance and sign-out validation

FDA clearance is the wrong first filter for interpretation software. These questions are better predictors of whether a platform will hold up in your CLIA program and survive an audit or a challenge.

1. Is the engine deterministic and transparent, so every output is reproducible and explainable rather than a black box?
2. Does each recommendation cite its underlying source, and can you trace it to a specific guideline and version?
3. Does it integrate with your existing calling pipeline—VCF, PharmCAT, Agena, CSV—rather than forcing you to rebuild or replace validated infrastructure?
4. Is there a complete audit trail of inputs, interpretive logic, and reviewer actions?
5. Does the workflow require human review and sign-out by your director on every report, with no autonomous release path?
6. Can you white-label and validate it in your own environment, under your own license and quality system?
7. Is the vendor candid about regulatory posture—reporting and interpretation software, downstream of calling, not a diagnostic test, and not FDA-cleared—rather than implying an endorsement it does not have?

A vendor that answers these plainly is one you can validate and defend. A vendor that dodges them, or that markets a clearance it does not hold, is a liability regardless of feature list. If you are weighing a commercial platform against an in-house build, work through the maintenance, guideline-drift, and validation-burden trade-offs deliberately—the build-versus-buy analysis lays out where each path tends to cost more than teams expect.

The bottom line for lab directors in 2026

For most PGx interpretation software, "not FDA-cleared" is the correct and expected status—because these platforms are reporting and interpretation tools that operate downstream of variant calling, under your CLIA lab's authority, not diagnostic devices. The 2024 LDT rule that might have reshaped this was vacated in March 2025 and formally rescinded that September, leaving CLIA (CMS) as the governing framework as of August 2026, with congressional proposals still pending and unresolved.

That means the accountable party has not changed. Your medical director validates the tool, reviews each report, and signs out; the treating physician makes the prescribing decision; the vendor supplies transparent logic, cited evidence, and an audit trail. The right diligence question is not "is this FDA-cleared?" but "is this transparent, deterministic, auditable, integrated with my pipeline, and does it keep my director in control?" Because oversight in this space is still moving, confirm the current LDT status and your specific obligations with qualified regulatory counsel—this article is background for that conversation, not legal advice.