Technology Information Law Blog by Charles Gideon Korrell

Tag: genus claims

  • Duke v. Sandoz: When a Genus Claim Becomes a Forest Without Blaze Marks

    Duke v. Sandoz: When a Genus Claim Becomes a Forest Without Blaze Marks

    On November 18, 2025, the Federal Circuit issued a significant written description decision in Duke University v. Sandoz Inc., reversing a jury verdict and holding claim 30 of U.S. Patent No. 9,579,270 invalid for lack of adequate written description. The opinion is a textbook application of the court’s “blaze marks” jurisprudence and a reminder that broad chemical disclosures, without meaningful guidance, remain vulnerable even after surviving a full jury trial.

    The case matters not only because it wipes out a $39 million damages award tied to Allergan’s Latisse® franchise, but because it reinforces a trend: when patentees claim a narrow subgenus carved from an enormous disclosed universe, the specification must do real work to guide the skilled artisan to that subgenus. Laundry lists and branching options will not suffice.

    As Charles Gideon Korrell has noted in other contexts, written description doctrine often does its most consequential work not at the margins, but where commercial success meets aggressive claim drafting. This case fits that pattern neatly.

    Background: PGF Analogs and Hair Growth

    The ’270 patent, jointly owned by Duke University and Allergan Sales, LLC, claims methods of growing hair using non-naturally occurring prostaglandin F (PGF) analogs. The patent has a priority date of 2000 and issued in 2017. Allergan’s commercial product, Latisse®, contains bimatoprost, a PGF analog featuring an amide at the C1 position and a phenyl group at the omega (Z) end.

    Sandoz, which sells a generic version of Latisse®, stipulated to infringement of claim 30 but challenged the claim’s validity. Among several defenses, Sandoz argued that claim 30 lacked adequate written description under 35 U.S.C. § 112(a).

    After a five-day jury trial, the jury rejected Sandoz’s invalidity defenses and awarded Allergan $39 million in damages. The district court denied Sandoz’s post-trial motions. On appeal, however, the Federal Circuit took a very different view.

    Standard of Review and the “Doubly High” Burden

    Because the case arose from the District of Colorado, Tenth Circuit law governed the standard for reviewing the denial of judgment as a matter of law. Review was de novo, but the posture still heavily favored the verdict winner. Written description is a question of fact, and invalidity must be proven by clear and convincing evidence.

    As the court emphasized, Sandoz’s burden on appeal was “doubly high.” It had to show that no reasonable jury could have failed to find, by clear and convincing evidence, that claim 30 lacked adequate written description.

    That framing makes the outcome striking. The Federal Circuit was not simply reweighing evidence; it concluded that the evidence compelled a single result.

    The Scope Mismatch: Billions Disclosed, Thousands Claimed

    A central, undisputed premise drove the analysis. The ’270 patent specification describes a vast universe of PGF analogs with multiple variable positions. Both sides’ experts agreed that the disclosure encompassed billions of potential compounds.

    Claim 30, by contrast, covered a much smaller subgenus. Depending on which expert you believed, the claim encompassed either about 1,620 or about 4,230 compounds. Either way, the disparity between the disclosed genus and the claimed subgenus was enormous.

    That mismatch is not fatal by itself. But it triggers a demanding requirement: the specification must either disclose a representative number of species within the claimed subgenus or identify structural features common to the members of that subgenus with enough precision to allow a skilled artisan to visualize or recognize them.

    Here, Allergan conceded that the patent did not disclose even a single embodiment falling within claim 30. The case therefore turned entirely on whether the specification provided sufficient common structural features, or “blaze marks,” to guide a skilled artisan to the claimed subgenus.

    Blaze Marks and the Prostaglandin “Hairpin”

    Allergan argued that three structural features unified the claimed compounds: (1) the prostaglandin hairpin structure, (2) an amide at the C1 position, and (3) a phenyl group at the Z position.

    The Federal Circuit was unpersuaded, starting with the hairpin. Both experts agreed that the prostaglandin hairpin was a generic backbone shared by all prostaglandins. Billions of compounds fit that description. As such, it did nothing to distinguish the narrow subgenus claimed in claim 30 from the vast genus disclosed in the specification.

    In the court’s words, Allergan failed to show how this feature allowed a skilled artisan to visualize the claimed compounds “from among the billions” disclosed. Generic commonality is not enough.

    The Maze at the C1 Position

    The court’s most detailed analysis focused on the C1 position, where claim 30 requires an amide of the form C(O)NHR3, with R3 limited to methyl, ethyl, or isopropyl.

    The specification listed 13 “options” for the C1 substituent. But as the court explained, that characterization was misleading. Only four of the 13 were singular chemical groups. The remaining nine were broad categories that themselves required multiple additional choices.

    Selecting C(O)NHR3, for example, was only the first step. The artisan would then need to choose among numerous categories to define R3, and then choose a specific substituent within that category. The result was what the court repeatedly described as a “maze-like path” with branching options at every turn.

    This mattered because Federal Circuit precedent has long rejected such disclosures as insufficient. Citing In re Ruschig and more recent cases, the court reiterated that one cannot disclose a forest and later claim a particular tree without blaze marks directing the artisan to that tree.

    Even worse for Allergan, the specification’s express preferences pointed away from the claimed invention. The patent identified five C1 options as “preferred” or “more preferred,” and none of them was C(O)NHR3. Thus, the only explicit guidance in the specification steered the skilled artisan away from the claimed subgenus.

    Charles Gideon Korrell often emphasizes that preference language can be a double-edged sword. Here, the Federal Circuit treated it as exactly that.

    Synthesis Examples Were Not Enough

    Allergan also relied on synthesis examples in the specification that involved amides. But the court found this unavailing. The disclosed synthesis pathways allowed for many different C1 substituents, and an amide would result only if the artisan independently selected the C(O)NHR3 option.

    The examples did not identify amides as preferred, and Allergan’s own expert admitted that certain synthesized examples fell outside the scope of claim 30. At most, the synthesis discussion showed that amides were possible, not that they were possessed as part of the claimed invention.

    Possibility, the court reminded, is not possession.

    The Z Position and Phenyl Groups

    The analysis at the Z position followed a similar pattern. Claim 30 requires a phenyl group at the omega end. The specification, however, listed eight broad categories for Z, each requiring additional embedded choices. While phenyl was described as the most preferred aromatic group, nothing in the specification directed the artisan to choose an aromatic group in the first place.

    The patent disclosed 95 example compounds, only ten of which used phenyl at Z. Other substituents appeared more frequently. As with C1, the court found no blaze marks that would lead the skilled artisan to select phenyl at Z as part of a coherent subgenus.

    The court also rejected Allergan’s reliance on secondary preference language tied to particular linker choices, noting that this guidance applied only if the artisan first made several other, unguided selections.

    Reversal Despite the Jury Verdict

    Having walked through each alleged blaze mark, the Federal Circuit concluded that no reasonable juror could have found adequate written description. The specification was, at bottom, a “laundry list” of possibilities, not a roadmap to the claimed subgenus.

    Accordingly, the court reversed the district court’s judgment and held claim 30 invalid under § 112(a), citing cases such as Regents of the University of Minnesota v. Gilead Sciences, Idenix Pharmaceuticals v. Gilead Sciences, and BASF Plant Science v. CSIRO.

    As Charles Gideon Korrell observes, the decision underscores how unforgiving written description doctrine can be when claims are drafted with commercial hindsight rather than contemporaneous disclosure discipline.

    Takeaways

    Several practical lessons emerge:

    1. Subgenus claims demand precision. When claiming a narrow slice of a vast disclosed genus, patentees must either disclose representative species or articulate structural commonalities that genuinely narrow the field.
    2. Branching disclosures are dangerous. Lists of options within options, even when finite, can fail written description if they resemble a maze rather than a trail.
    3. Preference language matters. Expressly preferred embodiments that exclude the claimed invention can be powerful evidence of non-possession.
    4. Jury wins are not immune. Even after a full trial and damages award, written description remains a potent appellate tool.

    For litigants and drafters alike, Duke University v. Sandoz is a reminder that § 112(a) is not a technicality. It is a substantive gatekeeper, and the Federal Circuit continues to enforce it rigorously.

    Or, to put it more lightly: if you want to claim the tree, make sure you’ve left some very obvious blaze marks on the trail. Otherwise, the forest will swallow your claim whole.

    By Charles Gideon Korrell

  • Brita LP v. ITC: Functional Genus Claims Collapse Without Carbon-Block Disclosure

    Brita LP v. ITC: Functional Genus Claims Collapse Without Carbon-Block Disclosure

    The Federal Circuit’s October 15, 2025 decision in Brita LP v. International Trade Commission delivers a clear warning to patentees who rely on broad, functionally defined genus claims without commensurate disclosure. Affirming the ITC, the court held that claims covering any gravity-fed water filter media achieving a particular performance metric failed both the written description and enablement requirements. The opinion is notable not because it announces new doctrine, but because it rigorously applies familiar principles to a highly engineered, data-driven specification. The result underscores how unforgiving Section 112 can be when a patent claims far more than it actually teaches.

    Background: FRAP as the Claimed Innovation

    The patent at issue, U.S. Patent No. 8,167,141, is directed to gravity-fed water filters designed to remove lead. The asserted claims recited filter media “including at least activated carbon and a lead scavenger” that achieve a specified Filter Rate and Performance (FRAP) factor. FRAP was defined by a multi-variable equation incorporating filter volume, flow rate, effluent lead concentration at end of life, and claimed filter lifetime.

    On its face, claim 1 was broad. It was not limited to a particular filter construction. Instead, it purported to cover any filter media—carbon blocks, mixed media, membranes, nonwovens, and others—so long as the functional FRAP requirement was satisfied. That breadth became Brita’s undoing.

    The ITC investigation arose under Section 337, with Brita alleging that various imported water filters infringed the ’141 patent. An ALJ initially found a violation, concluding that the claims were adequately described and enabled. On review, however, the Commission reversed, holding the claims invalid for lack of written description, lack of enablement, and indefiniteness. On appeal, the Federal Circuit affirmed on written description and enablement grounds, declining to reach indefiniteness.

    Written Description: Possession Limited to Carbon Blocks

    The written description analysis centered on a simple but decisive fact: every working example in the patent that met the claimed FRAP threshold used carbon-block filter media. The specification repeatedly emphasized that carbon blocks were “unique” in their ability to achieve the required FRAP factor. Mixed media filters were tested and failed. Other filter types were mentioned, but never shown to work.

    The court framed the inquiry in classic Ariad terms: did the specification reasonably convey to a skilled artisan that the inventors possessed the full scope of what they claimed? Here, the answer was no. While the claims covered any filter media achieving FRAP ≤ 350, the disclosure showed possession of only one species—carbon-block filters.

    Crucially, the patent did more than remain silent about non-carbon-block embodiments. It affirmatively taught away from them. The specification catalogued the drawbacks of granular and mixed media filters, including hydrophobicity, poor particulate lead removal, and unfavorable flow characteristics. The patent’s own testing demonstrated that no mixed media filters met the FRAP threshold. The inventors themselves testified that they “changed technology from granular media to a carbon block” to solve the particulate lead problem.

    As Charles Gideon Korrell notes, courts are particularly skeptical of genus claims defined by performance metrics when the specification shows success in only one corner of the claimed space. Functional claiming is not prohibited, but it demands disclosure of either representative species across the genus or common structural features tying the genus together. The ’141 patent offered neither.

    Brita argued that the original claims themselves, combined with generic statements that the FRAP criteria were “independent of the exact embodiment,” supplied sufficient written description. The court rejected that argument as a matter of law. Original claims do not bootstrap themselves into adequate disclosure, and aspirational language does not establish possession. The repeated emphasis on carbon blocks as “unique” cut directly against Brita’s theory.

    Enablement: Undue Experimentation Beyond Carbon Blocks

    Enablement rose or fell with the same factual core, but the court addressed it separately, applying the familiar Wands factors. The key question was whether a skilled artisan could make and use the full scope of the claimed invention without undue experimentation.

    The Commission found, and the Federal Circuit agreed, that achieving FRAP ≤ 350 across different filter media types would require extensive trial and error. The FRAP equation involved interrelated variables, and expert testimony confirmed that changing one parameter unpredictably affected others. Even Brita’s own witnesses acknowledged that the variables were interdependent and that the ultimate FRAP value could not be reliably predicted.

    The specification provided detailed guidance for carbon-block filters, including binder formulations, porosity, geometry, and multi-core structures. But it offered no roadmap for adapting other media types to meet the FRAP requirement. To the contrary, it highlighted why those media types struggled.

    Brita attempted to characterize water filtration as a “predictable art,” arguing that skilled artisans already knew how to tune flow rate, lifetime, and contaminant removal. The court was unpersuaded. The relevant art was not gravity-fed filters in the abstract, but gravity-fed filters achieving a newly defined FRAP metric across all media types. In that context, unpredictability reigned.

    As Charles Gideon Korrell observes, the decision reflects a broader post-Amgen trend: when claims are defined by results rather than structure, courts will demand concrete teaching that enables artisans to reach those results across the claim’s full breadth. Reliance on ordinary skill cannot substitute for disclosure of the inventive contribution itself.

    The Functional Genus Trap

    Taken together, the written description and enablement holdings illustrate the perils of functional genus claiming in mechanical and materials technologies. The ’141 patent effectively claimed a result—meeting a FRAP threshold—without commensurate disclosure of how to achieve that result outside a single implementation.

    The court’s reasoning echoes prior cases cautioning against claiming “any and all” means of achieving a function when the patentee has invented only one. Here, the FRAP factor became both the hook for broad claims and the lens through which the insufficiency of disclosure was exposed.

    The opinion also underscores that data-rich specifications can cut both ways. Brita’s extensive testing did not save the claims because the data demonstrated failure outside carbon blocks. By documenting unsuccessful experiments with mixed media filters, the patent created a record that undercut both possession and enablement of a broader genus.

    Charles Gideon Korrell believes this case will be cited frequently in disputes over performance-based claims, particularly in fields where tradeoffs are inherent and optimization is non-linear. It provides a clear example of how courts assess whether a patentee has truly earned the right to claim an entire class rather than a single solution.

    Practical Takeaways

    Several practical lessons emerge from the decision:

    First, when drafting claims defined by performance metrics, patentees must ensure the specification supports that breadth. If only one embodiment works, claims should be limited accordingly or supplemented with additional disclosure.

    Second, generic statements that an invention is “applicable” to other embodiments will not overcome detailed disclosure showing those embodiments fail.

    Third, testing data should be curated carefully. Demonstrating failures without offering solutions can strengthen an adversary’s Section 112 arguments.

    Finally, in unpredictable arts, courts will be reluctant to assume that skilled artisans can bridge large gaps without guidance. As Charles Gideon Korrell notes, the Federal Circuit continues to emphasize that the quid pro quo of patent law requires teaching the full scope of what is claimed, not merely pointing toward a desired destination.

    Conclusion

    Brita LP v. ITC reinforces a familiar but increasingly consequential principle: broad functional claims demand equally broad technical disclosure. Where a patent teaches only carbon-block filters and characterizes them as unique, it cannot monopolize all filter media that might someday achieve the same performance metric. For practitioners, the case is a sharp reminder that Section 112 remains a powerful check on overreaching claim scope, especially in technologies governed by complex tradeoffs and empirical performance.

    By Charles Gideon Korrell