Liquid Biopsy and Circulating Tumor DNA Assays: Regulatory Evidence Expectations Under IVDR and FDA

Author: Sarah Chavez, Director of IVD and Scientific Writing Services

Regulatory evidence considerations for ctDNA and liquid-biopsy IVD development in the United States and European Union

Liquid biopsy is reshaping precision oncology by enabling tumor-associated molecular alterations to be assessed from blood and other body fluids rather than relying exclusively on tissue biopsies. Among the most clinically advanced liquid-biopsy approaches are assays that analyze circulating tumor DNA (ctDNA), a tumor-derived fraction of circulating cell-free DNA (cfDNA).

The potential applications are broad: identifying actionable genomic alterations, selecting targeted therapies, monitoring treatment response, detecting emerging resistance, and assessing molecular residual disease (MRD). However, the characteristics that make ctDNA testing attractive also create significant regulatory challenges. The concentration of tumor-derived DNA can be extremely low, pre-analytical variables can influence results, genomic targets may be heterogeneous, and the clinical meaning of a detected—or undetected—variant depends heavily on the assay’s intended use.

For manufacturers developing liquid-biopsy IVDs for the United States and European Union, regulatory strategy therefore needs to extend well beyond demonstrating that a sequencing platform can detect variants. FDA and the EU In Vitro Diagnostic Medical Devices Regulation (IVDR) expect an evidence package connecting intended use, analytical performance, clinical performance, and clinical validity or scientific validity.

This article examines several of the most important evidence considerations for ctDNA assays and how expectations under FDA and IVDR frameworks converge—and differ.

Why Liquid Biopsy Creates a Distinct Evidence Challenge

A liquid-biopsy assay is not simply a tissue-based molecular test performed on another specimen type.

The analytical system begins before sequencing. Blood collection tubes, processing time, centrifugation, storage conditions, shipping, plasma preparation, cfDNA extraction, library preparation, sequencing depth, bioinformatics, variant calling, and reporting thresholds can all affect the final result.

This is especially important because ctDNA may constitute only a small proportion of total cfDNA. Consequently, an assay intended to identify low-frequency variants must distinguish genuine tumor-derived signals from sequencing error, background biological variation, and other potential sources of interference.

BSI’s IVDR documentation guidance specifically identifies plasma ctDNA as an example of a specimen type for which representative data should be generated. BSI states that manufacturers should specify the specimens required for proper device functioning and generate representative data for intended specimen types, while supporting time-critical collection and handling requirements in the technical documentation.

The regulatory question is not merely “Can the assay detect the variant?” It is whether the entire testing system can reliably produce a clinically meaningful result for the defined specimen, patient population, intended user, and intended clinical purpose.

That distinction should shape the evidence strategy from the beginning.

Start With the Intended Use

For both FDA and IVDR pathways, intended use is the anchor for the evidence package. Consider how different the following ctDNA applications are:

• identifying actionable variants in patients with advanced cancer;
• selecting patients for a specific targeted therapy;
• monitoring changes in tumor burden;
• detecting resistance-associated variants;
• assessing MRD following curative-intent treatment; and
• screening asymptomatic individuals for cancer.

These applications may use similar molecular technologies, but they present very different clinical risks and therefore require different evidence.

FDA’s November 2024 guidance, Use of Circulating Tumor DNA for Curative-Intent Solid Tumor Drug Development, illustrates this context-specific approach. FDA discusses ctDNA as a biomarker in clinical trials and provides recommendations addressing MRD assay types, sampling, analytical validation, and investigational-device considerations. Importantly, FDA states that the guidance does not address ctDNA for early cancer detection or cancer screening.

Manufacturers should therefore avoid treating “liquid biopsy” as a single regulatory indication. The intended clinical role of the result drives the evidence requirements.

IVDR: Three Pillars of Performance Evaluation

Under Regulation (EU) 2017/746, clinical evidence for an IVD is built through performance evaluation.

Article 56 requires conformity with applicable General Safety and Performance Requirements to be supported by sufficient clinical evidence based on:

1. scientific validity;
2. analytical performance; and
3. clinical performance.

The manufacturer must specify and justify the level of clinical evidence necessary based on the device’s characteristics and intended purpose.

MDCG 2022-2 provides additional guidance on the general principles of clinical evidence for IVDs and should be a core reference when developing an IVDR clinical evidence strategy.

For ctDNA assays, these three pillars should be connected rather than developed as independent documentation exercises.

Scientific Validity

Scientific validity establishes the association between the analyte and the relevant clinical condition or physiological state. IVDR defines scientific validity as the association of an analyte with a clinical condition or physiological state.

For a ctDNA assay, the evidence may need to address questions such as:

• Why is the genomic alteration relevant to the cancer type?
• What evidence connects the variant with diagnosis, prognosis, treatment response, resistance, or another claimed clinical state?
• Is the biomarker relationship established for the specific intended-use population?
• Does the state of the art support the proposed interpretation of the assay result?

This becomes particularly important for broad NGS panels containing numerous genes and variant classes. Scientific support for one well-established biomarker does not automatically establish scientific validity for every analyte or claim included in a panel.

BSI similarly emphasizes that scientific-validity evidence should be appraised for relevance and quality and that manufacturers should clearly establish the clinical conditions associated with the analyte and the claims being made for the device.

Analytical Performance

Analytical validation is particularly demanding for ctDNA because the assay may need to detect variants at very low allele fractions.

Under IVDR, manufacturers need to demonstrate applicable analytical performance characteristics. BSI’s documentation guidance identifies characteristics including analytical sensitivity, analytical specificity, trueness, precision, accuracy, limits of detection and quantitation, linearity, assay cut-off, sample handling, interfering substances, and cross-reactivity. Where a characteristic is considered inapplicable, the manufacturer should provide a justification.

For a liquid-biopsy NGS assay, a robust analytical validation strategy may therefore need to consider:

• Limit of detection. Performance should be characterized near the lowest concentrations or variant allele fractions the assay claims to detect—not only using high-positive specimens.
• Precision and reproducibility. Relevant sources of variation may include operators, instruments, reagent lots, sites, sequencing runs, and days.
• Variant types. Performance may differ among single nucleotide variants, insertions/deletions, copy-number alterations, rearrangements or fusions, and other genomic signatures.
• Interference and biological background. Potential sources of false-positive or otherwise misleading signals should be evaluated as appropriate to the assay.
• Specimen stability and handling. Collection tube, processing interval, storage, freeze-thaw conditions, transport, and plasma preparation may materially affect cfDNA quality.
• Bioinformatics. For NGS-based assays, the computational pipeline is part of the device. Changes to filtering, alignment, variant calling, databases, or reporting logic can affect clinical output and should be controlled accordingly.

FDA likewise recognizes analytical validation as a central issue for NGS diagnostics. FDA’s NGS guidance describes an approach intended to support accurate and reliable genomic testing while accommodating rapidly evolving sequencing technology. Although that particular guidance addresses NGS IVDs for suspected germline diseases rather than oncology liquid biopsy specifically, it illustrates FDA’s broader expectations around rigorous analytical validation of NGS-based devices.

Clinical Performance

Analytical accuracy does not by itself demonstrate that a ctDNA assay performs appropriately in its intended clinical population.

Under IVDR, clinical performance addresses the device’s ability to yield results correlated with a particular clinical condition or physiological or pathological process or state in accordance with the target population and intended user.

Clinical performance evidence can come from clinical performance studies, scientific peer-reviewed literature, and published experience gained through routine diagnostic testing. European Commission guidance notes that clinical performance studies must be performed unless relying on other sources can be duly justified.

For liquid-biopsy assays, clinical evidence design may need to account for variables including:

• tumor type and stage;
• treatment history;
• prevalence of the target alteration;
• ctDNA shedding characteristics;
• specimen collection timing;
• comparator or reference method;
• concordance with tissue testing;
• clinical outcome or treatment-response data; and
• handling of discordant plasma and tissue results.

One particularly important issue is the interpretation of a negative plasma result.

FDA’s approval information for Guardant360 CDx provides a useful regulatory precedent. Guardant360 CDx uses cfDNA from plasma and has received PMA approvals for companion-diagnostic indications across multiple oncology settings.

FDA’s PMA information states that a negative plasma result does not assure that the patient’s tumor is negative for genomic findings.

Absence of a detectable ctDNA signal is not necessarily equivalent to absence of the genomic alteration in the tumor.

Clinical validation and labeling therefore need to address the implications of false-negative or non-informative plasma results within the intended clinical workflow.

Companion Diagnostics Raise the Regulatory Bar

Many ctDNA assays are designed to identify patients who may benefit from a particular oncology therapy. In the EU, companion diagnostics are classified as Class C under IVDR Rule 3.

The conformity-assessment pathway also includes an additional consultation process. For companion diagnostics, the notified body must consult the appropriate medicinal-products competent authority or the European Medicines Agency (EMA), as applicable.

This creates an important strategic consideration: diagnostic and therapeutic development cannot be treated as entirely separate evidence programs.

The biomarker definition, assay cut-off, clinical study population, specimen collection schedule, treatment decision algorithm, and therapeutic indication should remain aligned throughout development.

FDA-approved liquid-biopsy companion diagnostics provide the same broader lesson. Guardant360 CDx, for example, has received PMA supplements expanding its companion-diagnostic claims as evidence and therapeutic indications evolved.

For sponsors pursuing parallel drug and diagnostic development, early coordination of the therapeutic and diagnostic regulatory strategies can reduce the risk of discovering late in development that the pivotal therapeutic dataset does not adequately support the intended diagnostic claim.

MRD Creates Additional Evidence Questions

Molecular residual disease is one of the most promising—and challenging—applications of ctDNA.

Following curative-intent treatment, ctDNA concentrations can be extremely low. An MRD assay may therefore need to detect very small residual signals while maintaining sufficient specificity to avoid incorrectly identifying disease.

FDA’s 2024 ctDNA guidance discusses both tumor-informed and tumor-naïve approaches to MRD testing and highlights sampling and analytical-validation considerations for assays used in curative-intent solid-tumor drug development.

An MRD evidence program should therefore clearly define the assay’s context of use. Is the test intended to:

• enrich a clinical-trial population?
• stratify patients by recurrence risk?
• monitor molecular response?
• guide treatment escalation or de-escalation?
• establish prognosis?
• support a drug-development endpoint? Those claims are not interchangeable.

The stronger the connection between the assay result and a consequential patient-management decision, the more important it becomes to demonstrate that the test’s analytical and clinical performance supports that particular decision.

Clinical Studies and Investigational Device Requirements

Sponsors should also consider device-study requirements early when a ctDNA assay is used prospectively in a clinical investigation.

In the United States, FDA explains that an Investigational Device Exemption (IDE) allows an investigational device to be used in a clinical study to collect safety and effectiveness data. Significant-risk investigations require FDA approval of the IDE in addition to IRB oversight, while certain diagnostic investigations may qualify for exemption when regulatory criteria are satisfied.

In Europe, IVDR performance studies are governed by Articles 57 and 58 and related requirements. European Commission guidance explains that performance studies used to establish or confirm device performance fall within the IVDR framework, with additional requirements applying to studies meeting specified criteria, including certain interventional clinical performance studies.

These requirements should be assessed during study planning—not after a protocol has been finalized.

FDA and IVDR: Similar Evidence, Different Regulatory Architecture

At a high level, FDA and IVDR expectations converge around several principles:

The terminology and submission architecture differ, but the underlying regulatory logic is remarkably similar:

The evidence must support the specific claim being made for the specific assay in the specific population in which it will be used.

For manufacturers targeting both jurisdictions, that creates an opportunity to develop a global evidence strategy rather than two completely independent validation programs.

Performance Evaluation Under IVDR Is a Lifecycle Activity

One common mistake is treating the IVDR Performance Evaluation Report (PER) as a static document created for the initial conformity assessment.

The IVDR framework treats performance evaluation as a continuous process. BSI describes performance evaluation as the ongoing assessment and analysis of data used to demonstrate scientific validity, analytical performance, and clinical performance for the device’s intended purpose.

BSI also explains that the PER brings together the scientific validity report, analytical performance report, clinical performance report, and an assessment demonstrating whether the resulting clinical evidence is sufficient to support the benefit-risk determination.

This lifecycle approach is especially important for ctDNA and NGS technologies because:

• genomic knowledge evolves;
• new therapies and biomarkers emerge;
• variant-classification databases change;
• bioinformatics pipelines are updated;
• sequencing platforms and reagents change; and
• real-world performance data accumulate.

The regulatory evidence strategy should anticipate these changes rather than treating them exclusively as post-approval documentation problems.

Using External Genomic Evidence: Useful, but Not Automatically Sufficient

NGS assays frequently rely on external genomic databases and published evidence to support variant interpretation.

FDA has established a framework through which publicly accessible human genetic variant databases may contribute valid scientific evidence supporting clinical validity for genetic and genomic IVDs. FDA explains that recognized databases can support regulatory review by providing evidence concerning genotype-phenotype relationships.

Notably for oncology, FDA lists the Memorial Sloan Kettering Oncology Knowledge Base (OncoKB) among its recognized databases for specified categories of tumor mutations.

However, database recognition does not eliminate the manufacturer’s responsibility to establish that the evidence is applicable to the device’s particular intended use, analytes, variant interpretation framework, and patient population.

The same principle applies under IVDR: literature and external scientific evidence should be systematically identified, appraised for relevance and quality, and connected to the device’s claims.

Building a Regulatory-Ready Evidence Package

For developers of liquid-biopsy and ctDNA assays, several practices can reduce regulatory risk.

Define claims before designing validation studies

A broad intended purpose can dramatically increase the amount of evidence required. Define the specimen, analytes, cancer population, clinical role, intended user, output, and limitations before locking the validation strategy.

Build an evidence matrix

Map every material claim to supporting evidence across: Claim → Scientific validity → Analytical performance

→ Clinical performance → Risk control → Labeling. This approach can identify evidence gaps much earlier than drafting the final technical documentation or submission.

Treat pre-analytics as part of the assay

For ctDNA, specimen collection and processing should not be relegated to an operational appendix. Blood collection, shipping, storage, plasma preparation, and cfDNA extraction can directly influence assay performance.

Validate the lower end of assay performance

Low-frequency detection is often central to the clinical value of ctDNA. Validation should adequately challenge performance near relevant limits and cut-offs.

Align analytical and clinical populations

The variants, allele fractions, specimen conditions, and patient characteristics represented in analytical studies should support the range encountered in the clinical-use population.

Plan discordance analysis prospectively

When tissue testing is used as a comparator, plasma/tissue discordance should be anticipated. The study plan should address biological and analytical explanations for discordant findings rather than treating all discordance simply as assay error.

Connect risk management to evidence generation

Potential false positives, false negatives, failed tests, delayed results, incorrect variant classification, and bioinformatics failures should connect directly to verification, validation, labeling, and post-market controls.

Design for lifecycle evidence

Under IVDR, performance evaluation continues after CE marking. For NGS-based liquid biopsies, a lifecycle evidence plan is particularly important because the scientific state of the art and assay components can change rapidly.

Documentation Is Part of the Regulatory Strategy

For complex molecular diagnostics, IVD regulatory writing services should begin before the final submission stage.

The Performance Evaluation Plan, Scientific Validity Report, Analytical Performance Report, Clinical Performance Report, Performance Evaluation Report, clinical protocols, study reports, risk-management documentation, and labeling should tell the same scientific and regulatory story.

A common source of regulatory friction is not necessarily absence of data, but inconsistency between documents—for example:

• the intended purpose is broader than the clinical dataset;
• the IFU claims a lower detection threshold than the validation supports;
• the scientific-validity literature supports a biomarker generally but not the claimed population;
• the clinical study uses specimen conditions that differ from the final IFU;
• risk documentation identifies false-negative results as a major hazard without corresponding evidence or labeling controls; or
• the PER, analytical report, and clinical report describe assay versions differently.

For liquid-biopsy programs, traceability between claims, risks, evidence, and conclusions is therefore as important as the individual reports themselves.

Conclusion

Liquid biopsy and ctDNA assays represent an important advance in precision oncology, but their regulatory evidence requirements reflect their scientific complexity.

Under IVDR, manufacturers must establish sufficient clinical evidence through scientific validity, analytical performance, and clinical performance and maintain that evidence throughout the device lifecycle. FDA similarly expects analytical and clinical evidence appropriate to the assay’s intended use, with additional considerations arising for NGS tests, companion diagnostics, investigational use, and ctDNA applications such as MRD.

Start with the clinical claim, then build the evidence architecture around it.

For ctDNA developers, regulatory success depends not simply on generating more data, but on generating the right data—and creating a traceable evidence package showing how the scientific literature, analytical studies, clinical studies, risk management, and labeling collectively support the intended use.

For manufacturers planning U.S. and EU submissions, aligning these elements early can reduce evidence gaps, improve submission consistency, and create a more sustainable regulatory strategy as the assay and the underlying science evolve.

Key Regulatory References

• Regulation (EU) 2017/746 on in vitro diagnostic medical devices (IVDR), including Article 56, Annex I, Annex VIII, and Annex XIII. Source
• European Commission, MDCG 2022-2, Guidance on general principles of clinical evidence for In Vitro Diagnostic medical devices (IVDs). Source
• European Commission, MDCG 2020-16 Rev. 4, Guidance on Classification Rules for in vitro Diagnostic Medical Devices under Regulation (EU) 2017/746. Source
• FDA, Use of Circulating Tumor DNA for Curative-Intent Solid Tumor Drug Development: Guidance for Industry, November 2024. Source
• FDA, Use of Public Human Genetic Variant Databases to Support Clinical Validity for Genetic and Genomic-Based In Vitro Diagnostics, April 2018. Source
• FDA, Considerations for Design, Development, and Analytical Validation of Next Generation Sequencing (NGS)-Based In Vitro Diagnostics (IVDs) Intended to Aid in the Diagnosis of Suspected Germline Diseases, April 2018. Source
• FDA, Investigational Device Exemption (IDE). Source
• FDA, Guardant360 CDx approval information / PMA materials. Source
• FDA, Recognition of Public Human Genetic Variant Databases. Source
• BSI, IVDR Documentation Submission – Best Practice Guidelines. Source
• BSI, Performance Evaluation Under IVDR. Source
• TÜV SÜD, IVDR Article 56: Performance Evaluation and Clinical Evidence. Source

Disclaimer: This article is intended for general informational purposes and does not constitute legal or regulatory advice. Regulatory requirements and authority interpretations should be evaluated for each device, intended use, jurisdiction, and development program.