Respiratory Panels: How Inclusivity and Cross-Reactivity Studies Differ

Start with the result each study needs to explain

For a US multiplex respiratory panel, inclusivity asks whether each claimed target is detected across relevant strains and variants. Cross-reactivity asks whether a different organism can produce an incorrect result for that target. The evidence is deliberately asymmetric: FDA's 2009 respiratory viral panel guidance recommends inclusivity near the assay cutoff or limit of detection (LoD), while cross-reactivity challenges use the highest anticipated clinical level. An organism may therefore belong in both studies, depending on which assay result is being evaluated. FDA guidance, section 6.B.

Keep separate objectives, challenge rationales, expected results and conclusions for these two evidence sets. They may be managed within one validation plan or in separate study documents; the cited FDA texts do not establish a blanket rule requiring two standalone protocols. What matters is whether the submission distinguishes the questions and supports the claims. A single percentage labeled 'specificity' does not show which targets, strains or non-target challenges were evaluated.

Consider influenza A in a panel that also detects respiratory syncytial virus. Influenza material supports inclusivity when the question is whether the influenza assay detects its intended target. It supports intra-panel cross-reactivity assessment when the question is whether that material produces an erroneous RSV result. Expected influenza type and subtype detections must still follow the panel's claimed reporting logic. 'No cross-reactivity' means no incorrect non-target result; it does not mean that every other optical signal or related reported analyte must be absent.

This article concerns manufacturer evidence for the two named US device types. It does not establish an assay-specific acceptance threshold, a laboratory verification procedure, or a method for interpreting an individual patient's result. The numerical examples below belong to the cited guidance or the identified historical labeling version.

Check the classification before borrowing study expectations

21 CFR 866.3980 classifies the respiratory viral panel multiplex nucleic acid assay as class II. It describes a combination of influenza, RSV, parainfluenza, human metapneumovirus, rhinovirus and adenovirus targets. Its special controls name the respiratory viral panel guidance and, where applicable, companion guidance for human metapneumovirus and influenza A subtype differentiation. Those additional controls matter when the claimed analytes bring them into scope. 21 CFR 866.3980.

21 CFR 866.4001 covers a different device type: a multiplex respiratory panel detecting emerging respiratory pathogens together with common respiratory pathogens. The Federal Register order identifies the BioFire FilmArray RP2plus De Novo classification as its basis. The classification was applicable on November 24, 2017; the order codifying it became effective August 21, 2025. The later date is not the start of all the underlying special controls. The order also states that devices of this type are subject to premarket notification requirements. 2025 classification order.

For the latter type, subsection 866.4001(b)(2)(ii) requires labeling performance characteristics that include inclusivity and cross-reactivity separately, alongside other analytical attributes. Subsection (b)(1) concerns intended use; it is not the analytical-performance list. The distinction is useful when building the evidence index: each named characteristic needs a supportable account, even when several studies share materials or administrative records.

Do not infer classification from a marketing phrase such as 'syndromic panel' or from the presence of one bacterial target alone. Compare the proposed intended use with the device identification and applicable special controls. Likewise, the 2009 viral-panel recommendations are not automatically the binding numerical criteria for every emerging-pathogen panel. Use them within their scope, and distinguish guidance recommendations from requirements printed in the regulation.

Inclusivity: show the breadth behind the claimed target

The 2009 guidance recommends data for relevant strains of every detected virus, with confirmed identities and titers. The number of strains can vary with the target. It recommends temporal and geographic diversity near LoD, including at least ten strains for each claimed influenza subtype and 10–15 adenovirus serotypes as examples. These are attributed recommendations, not universal minimum sample counts for every molecular assay. FDA guidance, section 6.B.ii.

The reason for reviewing diversity is the breadth of the label's target claim. A result from one well-characterized isolate cannot, by itself, describe all relevant variation within that target. The evidence review should explain which variants the tested material represents, which parts of the claim rely on other evidence, and where coverage remains uncertain. Merely listing strain names without linking them to the claimed analyte leaves that question unanswered.

The guidance's strain tables date from 2009. Treat their named isolates as historical examples rather than a current shopping list. The text itself notes that influenza vaccine strains change between seasons and recommends discussing a study with FDA when sufficient strains are difficult to obtain. It does not promise that a particular substitute material or computational analysis will always be accepted. A sourcing difficulty should remain visible in the evidence rationale instead of being converted into an unsupported equivalence claim.

The September 2023 BIOFIRE RP2.1 De Novo instructions for use, BFR0000-8579-03, illustrate the separation in practice. They describe inclusivity using sequence analysis plus isolates tested in triplicate near LoD, with detection within ten times LoD. Their cross-reactivity section separately describes on-panel and off-panel challenges: generally at least 10^7 units/mL for bacteria and fungi and 10^5 units/mL for viruses. This is a historical, single-device example, not proof of the latest labeling or a generic ten-times-LoD acceptance rule. RP2.1 IFU, pages 40–45.

Record the relationship between the reported challenge concentration and the assay's own LoD evidence. A broad 'near LoD' description becomes difficult to assess if it omits the analyte or specimen context. Avoid importing an arbitrary one-to-three-times-LoD rule from another study. The reader needs the basis of the conclusion, including any limitations, rather than a number that appears precise but has no support for the proposed device.

Cross-reactivity: assign non-target status to each assay result

The FDA guidance distinguishes cross-reactivity with other viruses inside the panel from cross-reactivity with organisms outside it. Within-panel assessment asks whether one target is incorrectly recognized by reagents intended for another. Off-panel assessment considers related organisms, respiratory pathogens not claimed by the device, and organisms relevant to the collection site. The guidance explicitly mentions Epstein–Barr virus and cytomegalovirus among the latter considerations. FDA guidance, cross-reactivity recommendations.

For cross-reactivity, the same guidance recommends medically relevant levels, usually 10^6 CFU/mL or higher for bacteria and 10^5 PFU/mL or higher for viruses. Its language is recommendation-based; these figures are not universal statutory floors. Preserve the original concentration units. A value expressed in infectious units, colony-forming units or genome copies is not an interchangeable label for the same quantity without evidence establishing the relationship.

The guidance also recommends evaluating analytical specificity through the complete system, from extraction through amplification and detection, in matrices equivalent to each claimed specimen type. A result from an isolated component answers a narrower question than a result from the claimed workflow. The evidence file should identify that boundary explicitly. It should not describe a component finding as established complete-system performance or assume that one specimen matrix represents every labeled matrix.

Cross-reactivity and interference belong to related analytical-specificity work, but their conclusions differ. A cross-reactivity assessment asks whether a non-target produces an incorrect positive result. An interference assessment asks whether another substance or organism alters detection of the intended target. Competitive interference within a multiplex panel is another reason not to treat a negative non-target result as proof that all interaction risks have been addressed. The FDA guidance discusses these questions separately.

Use a challenge-role matrix to review the evidence

The matrix below is an editorial synthesis of the FDA guidance's study distinctions. It organizes the evidence review rather than prescribing a complete protocol or new acceptance criteria. Assign each material relative to the assay result under review. An organism that is off-panel for one product can be a claimed target for another; a classification based only on its name will lose that distinction.

The last row has a specific regulatory boundary. Under 866.4001(b)(3)(ii), when the intended use includes an emerging respiratory pathogen for which an FDA-recommended reference panel is available, design verification and validation must include results from that analytical study. It is not a universal obligation to use every reference panel, nor evidence that a regulator-supplied panel replaces inclusivity and cross-reactivity work for all the device's other targets. Reference-panel special control.

A detected cross-reaction needs assessment, not an automatic verdict

The 2023 RP2.1 labeling reports predicted and observed cross-reactivity separately. Examples include predicted SARS-CoV-2 assay reactivity with selected animal coronavirus sequences and Bordetella-related cross-reactions affecting particular reported targets. Its table and footnotes distinguish predictions, tested findings and reporting consequences. Those distinctions must survive any summary; an in silico prediction is not an observed patient false positive. RP2.1 IFU, Table 45.

This published example shows why an assertion of perfect exclusion can be misleading. It does not establish that any reaction occurring at a high concentration is acceptable, or that adding a warning guarantees authorization. A finding still needs a device-specific account of the affected result, supporting evidence, clinical significance and proposed mitigation. Authorization is a regulatory determination, not a threshold that this article can derive from one vendor's labeling.

Keep unresolved findings visible in the study conclusion. If a result changes the apparent target breadth or specificity, reconcile the proposed claim, reporting interpretation and limitation text with that evidence. Do not infer FDA's reason for accepting a device from the existence of a labeled limitation. Similarly, do not transfer a finding between product generations without checking the corresponding device and labeling version.

Plan the recurring obligations under 21 CFR 866.4001

For devices within 866.4001, the annual influenza reactivity provisions are explicit. They should be tracked separately from the initial analytical evidence. The classification's 2017 applicability and its 2025 codification are different dates; neither supports calling the annual work a newly invented 2025 requirement. The following schedule summarizes the cited order's text, rather than asserting that any named manufacturer has or has not complied. Annual and emergency testing controls.

Both emergency clocks begin with FDA notification that characterized samples are available for evaluation. They do not begin when the manufacturer receives isolates, starts testing or completes a report. The provision concerns the specified influenza emergency circumstances; it should not be generalized into a 30-day duty for every new respiratory organism or every public health announcement. Maintaining the notification date and its associated deliverables is a practical way to make the actual trigger auditable.

For public disclosure, the order permits results directly in the accompanying labeling or a prominent hyperlink there to the manufacturer's public website. With the website route, the home page and primary product-information portion must also prominently link to the results, and access must be unrestricted. An internal report or a login-protected customer portal does not match that public-access option. Newer devices report all available annual results since authorization rather than waiting until three years have accumulated.

Read the mutation provision carefully. Subsection (b)(3)(iii) requires an appropriate documented risk-mitigation strategy for postmarket identification of genetic mutations and novel isolates or strains. It gives regular literature review and annual in silico analysis as examples and requires documentation of results and findings. Calling one example the regulation's only computational requirement would obscure the broader strategy and the separate software-description provisions.

These controls do not state an equivalent annual cross-reactivity testing schedule. That observation is limited to the cited provisions. It does not exempt a device from reassessing specificity when evidence, risks or changes warrant it, and it does not establish that every device under 866.3980 has only static premarket duties. Check the applicable companion controls and device-specific conditions before drawing lifecycle conclusions. There is also no prescribed 'Q2 testing window' in the cited order.

Translate terminology into a reviewable study plan

ATCC's white paper defines inclusivity as the proportion of target strains or DNA samples giving a correct positive result, and exclusivity as the proportion of non-targets giving a correct negative result. It also discusses representative variants, related organisms and shared specimen niches. This supplier terminology is useful, but the white paper is not FDA policy and does not establish acceptance criteria for a respiratory-panel submission. ATCC terminology overview.

A percentage can summarize observations. It cannot identify the diversity represented, the challenge level, the tested matrix or the untested boundary by itself. Translate a supplier's 'exclusivity panel' into the claimed-target and non-target roles in the matrix before using it as evidence. Avoid claiming that FDA rejects all percentage reporting: the problem is an unsupported conclusion, not the use of arithmetic.

Before approving the analytical evidence package, review the following decisions. These are editorial review prompts drawn from the distinctions above, not a regulator-issued checklist or an additional named test.

1. Identify the device type and applicable controls. Record why the intended use fits the classification, including any companion guidance that applies to the claimed analytes.

2. Map each claimed result to its inclusivity evidence. Identify the represented diversity, supporting study context and remaining limitations without implying that an isolate list proves exhaustive coverage.

3. Map each non-target challenge to the assay results it could affect. Separate other panel targets, off-panel organisms and specimen-relevant background organisms, while preserving expected type/subtype reporting.

4. Retain concentration units and the rationale for the challenge level. Distinguish the 2009 recommendations from a particular manufacturer's tested values and from device-specific acceptance criteria.

5. Separate cross-reactivity, interference and reference-panel conclusions. Record predictions and observations distinctly, and explain where a component or specimen boundary limits the conclusion.

6. Reconcile study conclusions with the labeling. Carry relevant limitations forward; do not turn a characterized reaction into an automatic approval rule or an unqualified specificity claim.

7. For 866.4001 devices, assign responsibility for the recurring reactivity results, the public labeling route and the FDA notification clocks. Retain the broader mutation-risk strategy and its findings as part of the lifecycle evidence.