November 20, 2026
In Exelixis, Inc. v. MSN Laboratories Private Ltd., No. 2025-1236 (Fed. Cir. Aug. 31, 2026), the Federal Circuit affirmed that claims encompassing crystalline forms of a pharmaceutical salt satisfied the written-description requirement of 35 U.S.C. § 112(a), although the specification specifically described only two polymorphs. The precedential decision illustrates how common structural characteristics can support a chemical genus claim without a showing that the disclosed species are representative of the genus.
Exelixis holds the New Drug Application for Cabometyx®, which contains cabozantinib (L)-malate. Cabozantinib (L)-malate can exist in amorphous or crystalline forms. The crystalline form may further exist as different polymorphs.
Exelixis owns three patents relevant here: U.S. Patent Nos. 11,091,439, 11,091,440, and 11,098,015 (collectively, the “Malate Salt Patents”), directed to crystalline cabozantinib malate salts, pharmaceutical formulations containing the salts, and methods of treating cancer using them, respectively. The patents shared a specification that specifically described two crystalline polymorphs, N-1 and N-2.
Importantly, the claims of the Malate Salt Patents were not limited to either N-1 or N-2. Instead, they broadly encompassed crystalline cabozantinib (L)-malate without specifying a particular polymorph.
MSN sought approval for generic cabozantinib (L)-malate tablets containing a different crystalline polymorph, Form S, for which MSN had obtained its own patent. Exelixis sued MSN for patent infringement. Although MSN conceded infringement of the asserted claims of the Malate Salt Patents, it challenged their validity for lack of adequate written description under § 112(a). Specifically, MSN argued that the specification failed to identify structural features common to all crystalline forms of cabozantinib (L)-malate sufficient for a skilled artisan to recognize the members of the claimed genus. Although all polymorphs share the same chemical name and formula, they have different three-dimensional crystal structures and may exhibit different physicochemical properties, including melting points, solubilities, and stability. MSN contended that the properties of the disclosed N-1 and N-2 polymorphs could not be used to predict those of other polymorphs, and therefore the disclosure did not demonstrate possession of the broader claimed genus.
The district court rejected MSN’s written-description challenge, and MSN appealed. MSN also appealed the judgment on a fourth patent, U.S. Patent No. 11,298,349, but the Federal Circuit dismissed that part of the appeal as moot and vacated the district court’s judgment on the asserted claim of that patent.
Applying Ariad Pharmaceuticals, Inc. v. Eli Lilly & Co., the Federal Circuit explained that written-description support for a genus may be shown by either (1) a representative number of species or (2) structural features common to genus members that allow a skilled artisan to recognize them. Because the issue of written-description support is one of fact, the court reviewed the district court’s factual findings for clear error.
The Federal Circuit found the district court did not clearly err in the determination that the specification identified common structural features, particularly the chemical name and formula of cabozantinib (L)-malate and its crystalline character. The specification also disclosed processes for preparing the invention, although that evidence was not dispositive.
MSN emphasized that polymorphs may differ in melting point, solubility, hygroscopicity, and stability, making the properties of undisclosed forms unpredictable. The court found that argument insufficient because the claims did not require those properties, and MSN had not explained why the unclaimed differences were relevant in claims that did not include these features or shown that the district court clearly erred in its fact finding. In particular, the court distinguished this case from AbbVie Deutschland GmbH & Co. v. Janssen Biotech, Inc., where specific evidence showed substantial structural differences between the disclosed antibodies and other antibodies within the claimed genus. In contrast, MSN failed to establish the relevance of these properties in the structural analysis here. Moreover, the district court did not rely on the properties of the N-1 or N-2 polymorphs to identify other polymorphs.
The court also noted an unchallenged district court finding that the maximum potential size of any pure polymorph genus is fourteen forms, distinguishing the dispute from cases involving potentially vast genera. Because common structural features sufficed, the court did not decide whether N-1 and N-2 constituted a representative number of species.
For chemical and pharmaceutical applicants, the decision underscores the value of describing the structural characteristics shared by the members of the claimed genus, rather than relying exclusively on examples of individual species. It also shows that differences in unclaimed physicochemical properties need not defeat written-description support when the claimed genus is adequately identified by structure. While adding certain properties to a claimed genus may help with novelty or nonobviousness during prosecution, including such properties may limit the breadth of the claimed genus and raise concerns about the written-description support.
The holding is fact-specific. It does not establish that two disclosed polymorphs always support claims to every crystalline form or resolve whether such claims meet the separate enablement requirement. The size of the genus, the specification’s disclosures, and the actual claim limitations remain critical to the analysis of written description.