Monday, November 16, 2026

FDA Downclassifies Again: IHC for CDx

I posted a blog on September 3, 2026, that FDA had recently downclassified oncology cDx with ISH or FISH from  Class III (PMA) to Class II (e.g. 510k or de novo 510k).  At the time, the news was two weeks old but had gotten very little press.

Now, in November, similar news for a much larger area, IHC, immunohistochemistry.

This should be the auto AI that Google itself gives you: 

https://share.google/aimode/O5TrMaWcf2m8Ki6J1

Here's the FDA reclassification proposal:

https://www.federalregister.gov/documents/2026/11/12/2026-23005/hematology-and-pathology-devices-reclassification-of-immunohistochemistry-test-systems-for-use-with

FDA accepts public comment now and will finalize in a year or two.  You have until January 11, 2027.


Pubmed tells us FDA wrote about the topic on June 3, 1998, PMID 10179865.   Here.  63fR30132.

Read about the new publication at Sharon Liang at linked in here.

IHC is hardly cut and dried, see debates about PDL1 staining and about Enhertu in pan cancer here.

###

AI CORNER

###


FDA Proposes Easier Market Entry for Oncology IHC—and a Regulatory Gem from 1998

The proposed move from PMA to 510(k) has a revealing historical precedent.

The 2026 Proposal

On November 12, FDA proposed reclassifying oncology immunohistochemistry (IHC) test systems from Class III to Class II, replacing premarket approval (PMA) with 510(k) clearance and special controls. Comments are due January 11, 2027.

The scope includes companion diagnostics essential to treatment selection, plus certain tests that inform a drug’s benefits or risks without being prerequisites for treatment. FDA proposes a new regulation, 21 CFR 864.1895.

Substantial Evidence Requirements Remain

The proposed controls require analytical and clinical performance evidence, validated scoring thresholds, and reproducibility—including between readers and near clinical decision cutoffs.

Former FDA leader Sharon Liang, in her recent LinkedIn commentary, highlights why this matters for PD-L1: different antibody clones, scoring systems, thresholds, and reader variability complicate reliable treatment selection. Her discussion places IHC within FDA’s broader effort to reclassify higher-risk IVDs.

The Historical Treat: FDA Already Revisited IHC in 1998

The June 3, 1998 final rule, 63 FR 30132–30142, is a fascinating companion read. FDA established three categories:

  • Class I, exempt from 510(k): adjunctive stains supporting conventional histopathology, such as keratin.

  • Class II: certain established prognostic or predictive tests, including estrogen and progesterone receptors.

  • Class III: uses falling outside those categories.

Public Comments Changed the Outcome

The final rule was substantially less burdensome than the proposal. FDA broadened Class I, exempted those devices from premarket notification, and moved ER/PR into Class II.

The reasoning is especially interesting. FDA recognized that pathologists integrate adjunctive stains into a diagnosis, while tissue context and laboratory controls help reveal reagent failures. The rule explicitly discussed CLIA-required positive and negative controls.

Commenters also questioned whether reviewing each antibody would meaningfully improve interpretation, and warned about costs and barriers to introducing useful reagents. Reading these exchanges nearly three decades later feels surprisingly contemporary.

What Connects the Two Decisions?

My reading is that both documents show FDA adjusting oversight as experience makes risks—and the means of controlling them—better understood.

The distinction remains consequential: routine adjunctive staining, independently reported predictive information, and drug-linked treatment selection carry different evidentiary demands. The 2026 proposal would bring a defined group of oncology IHC systems into Class II while retaining explicit performance requirements.

For developers, that could make commercialization and subsequent product changes more manageable. For policy readers, the 1998 document offers something rarer: a detailed record of how today’s familiar regulatory categories were negotiated into existence.

##

##


Part Two: 

What FDA Says—and Where Industry Will Push for Clarity

FDA’s Argument: Experience Supports a Different Regulatory Approach

FDA describes a technology whose risks are now sufficiently understood to manage through general and special controls. Its experience includes 22 original PMAs and approximately 270 supplements. For this reclassification, FDA relies on 14 original PMAs across five product codes to support a category covering 12 codes.

That distinction matters. Under the statutory six-year rule, FDA excludes recently approved PMAs and supplements from the information used to support this action. FDA nevertheless concludes that differences in antibodies, biomarkers, tumor types, and corresponding drugs do not create unique risks requiring separate classifications. Its principal concerns are false results, test-system failures, and incorrect interpretation.

FDA also cites limited postmarket safety signals. Those reports support its reasoning, although passive reporting cannot establish the true frequency of diagnostic errors.

See FDA’s proposal, Sections II and VI.

What FDA Would Actually Change

FDA would establish a single Class II category under 21 CFR 864.1895, encompassing both essential companion diagnostics and certain drug-linked tests providing additional benefit-risk information.

Previously approved products could remain on the market without another authorization. For significant subsequent modifications, FDA anticipates 510(k) submissions. It also explicitly encourages considering predetermined change control plans (PCCPs), which can permit specified, preauthorized changes without a separate submission each time.

The proposed special controls reach beyond antibody specificity. They address specimen stability, robustness, scoring rules, clinical thresholds, reader reproducibility, and clinical performance. Labeling must summarize performance and remain consistent with the corresponding drug labeling.

FDA proposes an effective date 30 days after publication of a final order. See Sections IV, VII, and X, plus footnote 10.

Where I Expect the Most Industry Comment

The following are my anticipated discussion points; they are not a list of questions expressly posed by FDA.

1. How Much Clinical Bridging Is Enough?

FDA expressly contemplates clinical evidence from either the therapeutic trial’s enrollment assay or a method-comparison study against an appropriate FDA-authorized device. Developers will want clearer criteria for when comparison studies suffice—and when new outcome-linked evidence is necessary. Agreement near a treatment cutoff will be particularly consequential.

2. How Predictable Will Validation Requirements Be?

The controls require testing across reagent lots, operators, instruments, days, readers, and intended-use sites, with attention to borderline specimens. Industry will likely seek practical expectations for sample sizes, acceptance criteria, rare specimens, and the permitted use of surrogate samples. Repeated references to FDA determining what is appropriate provide flexibility, but leave important development costs uncertain.

3. What Does One Broad Category Mean for New Entrants?

A shared classification does not establish that different assays are clinically interchangeable. Manufacturers will want clarity about suitable predicates and the evidence needed when changing antibody clones, platforms, scoring methods, or indications. The breadth of FDA’s risk-based grouping may attract particular scrutiny from developers of newer biomarkers.

4. Where Do Digital Scoring and AI Fit?

FDA identifies software failure as a possible source of error, but this proposal does not provide a detailed pathway for every digital or AI-assisted scoring configuration. Developers will want to understand the assay–software boundary, the necessary validation, and the treatment of algorithm updates.

5. How Much Will Product Maintenance Improve?

PCCPs could be commercially important. Industry will likely request concrete examples of changes that can be planned in advance, plus instructions for moving existing PMA products and pending supplements into the new framework. With a proposed 30-day transition, operational clarity deserves attention.

Comments are due January 11, 2027, under FDA-2026-N-12508. The full proposal and submission instructions are here.