NewsArticles
Articles
NEW DEVELOPMENTS IN COMPOUND PATENT INVALIDATION EXAMINATION OBSERVATIONS ON CNIPA PRACTICE AND LEGAL ISSUES, 2025—2026
2026-07-29
Author: CAO Lili, Stephen Zou

Executive Summary

This newsletter examines key developments in China's compound (including salt/prodrug) patent invalidation practice, based on landmark decisions issued by the China National Intellectual Property Administration (CNIPA) during 2025-2026. The analysis covers three dimensions:

I. Scope of Stereoisomer Claims: CNIPA clarified that an unmarked bond ( “plain line” ) in a chemical structure encompasses all stereochemical configurations, not merely the racemate.

II. Advances in Inventiveness Assessment: From the evidentiary characterization of foreign conference slide decks, to substantive eligibility review of the closest prior art, to holistic consideration of structure-activity relationships (SAR).

III. Sufficiency of Disclosure for Compound Salts: An integrated evaluation framework anchored on the parent compound's activity and supplemented by lawfully cited reference documents.

Editor's Note

This newsletter analyzes recent decisions from CNIPA's Reexamination and Invalidation Department(the "Collegial Panel"or "Panel") concerning pharmaceutical compound patents. In China's patent system, any entity may file a request for declaration of patent invalidation after grant. The Panel examines such requests and issues invalidation decisions, which are appealable to the courts. The decisions discussed below are identified by their official decision numbers. For ease of reference, Chinese statutory provisions are cited with their English translations; readers should consult the original Chinese text for authoritative interpretation.

Introduction

Invalidation proceedings for compound (including salt/prodrug) patents have long been the most contested and technically advanced battleground in pharmaceutical intellectual property. As innovative drug R&D capabilities rise and generic/modified drug market competition intensifies, the adversarial dynamics of invalidation proceedings continue to escalate, and every adjustment in examination standards may reshape industry expectations. Through several key invalidation decisions issued by CNIPA during 2025-2026 — particularly the ruxolitinib salt, enzalutamide, finerenone, desloratine citrate disodium hydrogen salt complex, and sofosbuvir prodrug series cases — one can clearly observe a shift toward an examination model that places greater emphasis on structure-activity relationships, technical motivation, and substantive reasoning. This article analyzes this new examination trend and its legal implications, layer by layer, with reference to the above landmark cases.

Part I: Scope of Stereoisomer Claims — Interpretation Boundaries

In compound patent invalidation examination, when a claim defines its scope of protection through a chemical structural formula, the manner in which the bond at a chiral center is drawn — whether using a wedge bond with explicit stereochemical orientation or merely a plain line — gives rise to disputes over the scope of protection that concern not only the basis for inventiveness determination in prosecution but also constitute the core battleground between parties in infringement disputes. Two landmark decisions from the 2025-2026 period, namely the sofosbuvir prodrug case (Decision No. 588818) and the finerenone case (Decision No. 586870), addressed the debate over whether an unmarked configuration is equivalent to a racemate. CNIPA established a clear interpretation rule for chiral compound claims: a “plain line” in a structural formula encompasses all possible stereochemical configurations.

Sofosbuvir Prodrug Case  —  Decision No. 588818

Claim 1 of the patent at issue contained multiple chiral centers, most of which were explicitly marked with wedge bonds for stereochemical configuration, except for the phosphorus atom — a key chiral center — which was connected only by a plain line “-” without any stereochemical notation. The petitioner argued that this line should be interpreted as a limitation to a “racemate” (i.e., a 1:1 mixture of R- and S-configurations at the phosphorus atom, as in the product of Example 25, PSI-7851), thereby excluding sofosbuvir (PSI-7977, with S-configuration at phosphorus) from the scope of protection.

The Panel rejected the petitioner's argument through a four-layered analysis:

First, starting from the text and graphical expression of the claim itself, the Panel noted that a plain line in a chemical structural formula merely represents a covalent bond between atoms and does not inherently possess the function of indicating stereochemical configuration. To specifically denote a racemate, additional nomenclature identifiers such as “(±)” or “rac-” would typically be required. Even if certain textbooks use plain line notation in specific contexts, this is done on the premise that the compound is already known to be a racemate, rather than serving as a general definition of the term.

Second, the Panel examined the specification's explanation of the structural formula. Paragraph [0072] of the specification explicitly recorded the chirality of the phosphorus atom and the possibility of R/S designation, while emphasizing that the applicant had considered both the racemate and resolved enantiomers for use. Paragraph [0818] more directly stated that when the phosphorus atom in the general formula is connected by a plain line, it encompasses “all reasonable stereochemical configurations.” These explicit textual descriptions provided a solid internal basis for a broad interpretation of the claim.

Third, the Panel clarified the methodological relationship between "claims"and"examples". The petitioner attempted to use the identity of the structural formulae of Claim 1 and Example 25 to narrow the claim's scope to the specific product of Example 25. The Panel rejected this view, stating: “Based on the information recorded in the specification, one can only determine that the compound of Example 25 is a mixture of R- and S-configurations at the phosphorus center; it is difficult to conclude that it is a 'racemate.'” The Panel further explained that a claim is a generalization based on the specification, while an example is typically a description of an exemplary technical solution within the claim's scope. Even if the compound of Example 25 were indeed a “racemate,” it would merely be one specific embodiment falling within the claim's scope of protection and cannot retroactively narrow the claim's generalization.

Fourth, the Panel examined the prosecution history to confirm the patentee's true intent. Although the patentee had deleted the phrase “or stereoisomers thereof” during prosecution, the file history showed that the patentee consistently maintained specific claims covering stereochemical variants at the phosphorus center and explicitly stated in observations that the general formula already covered relevant diastereomers, never having subjectively abandoned protection for sofosbuvir. This consistent interpretation further confirmed the position that “a plain line merely represents an unmarked configuration.”

Based on the above analysis, the Panel established the core adjudicative rule in the sofosbuvir prodrug case: when a chiral center in a chemical structural formula is not marked with a wedge bond, hashed bond, or other stereochemical notation, the chiral center should be understood to encompass all possible stereochemical configurations (including R-form, S-form, and mixtures thereof in any ratio), rather than being limited to the racemate.

Finerenone Case  —  Decision No. 586870

The subsequent finerenone case confirmed the application of this rule. In that case, the compound of Claim 1 had one chiral center that was likewise not marked with a wedge bond. The petitioner argued that, because Claim 1 did not limit the specific stereochemical configuration, it should be understood as an equimolar mixture of different stereoisomers (i.e., a racemate) and correspondingly matched to the specific experimental compound in the specification's examples. The Panel directly cited and followed the adjudicative logic established in the sofosbuvir prodrug case, stating that a plain line merely represents a covalent bond and does not indicate stereochemical configuration. Combined with the fact that the specification likewise recorded that a single enantiomer could be obtained through chiral resolution, the Panel again confirmed that the scope of protection of Claim 1 was not limited to the racemate but encompassed all possible stereochemical configurations at that chiral center.

Key Takeaways: Interpretation Rule for Stereoisomer Claims

When a chiral center in a chemical structural formula is not marked with a wedge bond or hashed bond, it encompasses all possible stereochemical configurations (R-form, S-form, and mixtures thereof).

A plain line merely represents a covalent bond and does not indicate stereochemical configuration; denoting a racemate requires additional identifiers such as (±) or rac-.

An example is an exemplary technical solution within the claim's scope and cannot retroactively narrow the claim's generalization.

Consistent statements by the patentee in the prosecution history may serve as supplementary evidence for interpreting claim boundaries.

These two decisions jointly established the interpretation rule for the scope of protection of chiral compound claims, providing important guidance for determining the scope of protection of chiral compound patents in the pharmaceutical chemistry field. At the prosecution level, this rule formally affirmed the feasibility of seeking protection for chiral compound families using general formulae or Markush claims, encouraging innovators to maximize the protective radius of core compounds through high-quality claim design and comprehensive specification support during patent drafting.

Part II: Advances in Compound Inventiveness Assessment

The core challenge in compound patent inventiveness assessment lies in delineating a reasonable boundary between “structural similarity” and “non-obviousness.” The 2025-2026 invalidation decisions indicate that the examination focus is shifting from one-dimensional structural comparison to a comprehensive analysis of the overall teaching of the prior art and the compound's structure-activity relationship (SAR). Furthermore, the application of procedural and evidentiary rules also profoundly affects case outcomes, particularly regarding the evidentiary characterization of academic conference presentation slides.

(1) Evidentiary Characterization of Foreign Conference Presentation Slides: Analogous to “Public Use” Rather Than “Publication”

Enzalutamide Series Cases  —  Decision Nos. 582014, 582044, 582111

The core dispute in this series of cases concerned a presentation slide deck displayed by the inventor, Charles L. Sawyers, at the Prostate Cancer Foundation Scientific Retreat held in Scottsdale, Arizona, USA, from September 29 to October 1, 2005. The petitioner sought to establish that this slide deck constituted prior art under patent law, relying on Information Disclosure Statement (IDS) submission records, evidence of the conference's open nature, and findings from UK counterpart litigation. However, the Panel rendered highly consistent negative assessments in all three decisions, with the core logic resting on a precise characterization of the mode of disclosure.

1. Distinction Between Physical Form and Legal Characterization

The Panel first distinguished the physical form of the slide deck from its legal characterization. Although the slide deck recorded specific compound structures (e.g., Compound 162) and experimental data, its display process — projecting slides onto a screen via a projector — more closely resembled "public use"or"other means of disclosure"as defined in the Patent Examination Guidelines (2006), specifically “information materials and visual materials placed on display stands or in shop windows for public perusal” (such as posters, drawings, and photographs), rather than “publication.” This classification was critical because it directly triggered the territorial limitation provision.

2. Application of the Territorial Limitation

Under the then-applicable 2001 revised Implementing Regulations of the Patent Law, Article 30, the territorial scope of prior art disclosed through publication was worldwide, while prior art disclosed through use or other means was strictly limited to within China. The conference took place in the United States and was a foreign event. Therefore, even if the content was acknowledged as having been displayed, as long as it did not constitute "publication", it would not constitute prior art under Chinese patent law.

3. Distinction Between Fact-Finding and Legal Application

In determining the factual question of "whether the slide deck was actually displayed", the Panel invoked the principle of good faith and the burden-of-proof allocation mechanism. Given that the slide deck was created by the inventor affiliated with the patentee, and that the patentee had submitted the document to the USPTO via an IDS in a counterpart U.S. application, the patentee was closer to the factual truth and possessed a stronger evidentiary capacity than the petitioner. In the absence of rebuttal evidence from the patentee, the Panel applied the standard of high probability and found that the slide deck was publicly displayed at the 2005 conference. However, this factual finding did not affect the ultimate legal conclusion — based on the characterization of the disclosure mode and the territorial limitation, the slide deck still could not serve as prior art for assessing the inventiveness of the Chinese patent.

In summary, the enzalutamide series cases established a clear adjudicative approach: for presentation slides displayed at foreign academic conferences, unless the petitioner can prove that the content was formally published (e.g., included in conference proceedings bearing an ISBN/ISSN number) or was publicly disclosed within China, the slides are merely treated as evidence of foreign "public use" and cannot serve as prior art for assessing the inventiveness of a Chinese patent. This approach both respects the tradition of free academic exchange and adheres to the territoriality principle of patent law.

(2) Substantive Review of Closest Prior Art Eligibility: From "No Right or Wrong in Selection"to"Eligibility Review"

Finerenone Case  —  Decision No. 630569

China's three-step method for patent inventiveness examination, though derived from the European Patent Office (EPO) problem-solution approach, evolved a markedly different logic in the long-standing practice of the first step — "determining the closest prior art". For an extended period, Chinese judicial and administrative practice tended to hold that there is no absolute "right or wrong" in the selection of prior art, but only whether it is"appropriate". Under this lenient approach, even when the patentee argued that a certain reference document was unsuitable as the starting point for modifying the invention, the Panel would typically accept the petitioner's selection and proceed directly to the analysis of distinguishing features and teaching. By contrast, the EPO imposes a strict eligibility threshold for the"closest prior art"— under the EPO Guidelines, the selected document must be the single reference that provides the “most promising starting point” for the invention's development.

However, the 2025-2026 finerenone case signals a significant change in this lenient attitude. In that case, the petitioner insisted on using Evidence 2 (WO2007009670A1) as the closest prior art, and the compound disclosed in Evidence 2 was likewise an MR(mineralocorticoid receptor) antagonist (same use/mechanism). However, the Panel did not accept this selection as a matter of course but instead conducted a substantive eligibility review of the "first step". The Panel stated in its decision: “The determination of the closest prior art aims to enable the person skilled in the art to determine the starting point of the invention in a manner as close as possible to the inventive process. For compound inventions, if the patented compound and the closest prior art differ too greatly in structure, the person skilled in the art would typically have no motivation to select it as the starting point for the invention based solely on the same mechanism or use.” Upon comparison, the patented compound and Evidence 2 had "obvious differences in ring unit type, inter-ring connectivity, and ring substituents", and the Panel found that the person skilled in the art would not select it as the starting point for the invention.

The significance of this adjudicative logic lies in its break from the old inertia of “no right or wrong in selection; only teaching needs to be addressed.” It represents a refined reconstruction of the first step of the three-step method, aimed at ensuring that subsequent inventiveness determinations are based on genuinely meaningful comparisons rather than hypothetical inferences.

(3) Technical Teaching: From Structural Comparison to Holistic SAR Analysis

Historically, petitioners were accustomed to arguing that the patented compound was obvious based on structural similarity to a compound disclosed in the prior art. However, the 2025-2026 invalidation decisions indicate that the examination focus is shifting from one-dimensional structural comparison to a comprehensive analysis of the overall teaching of the prior art and the compound's structure-activity relationship (SAR). The following three sets of cases demonstrate this shift from different angles:

1. Different Mechanism of Action — Finerenone Case (Decision No. 586870)

The petitioner cited Evidence 3 (EP0234516A1) and Evidence 4 (US4760081A) as the closest prior art. Both disclosed structurally similar 1,4-dihydro-1,6-naphthyridine compounds, but explicitly stated that the invention's purpose was “highly potent L-type calcium channel (Ca²⁺) antagonists,” whereas the patented compound was ]intended to provide “selective mineralocorticoid receptor (MR) antagonists with low affinity for the L-type calcium channel.” The Panel stated: “If the mechanism of action of the patented compound differs from that of the closest prior art compound, and the evidence on file is insufficient to motivate the person skilled in the art to use the closest prior art compound as a research basis for structural modification, then it cannot be considered that the prior art as a whole provides a technical teaching for arriving at the patented compound invention.” Evidence 3 and 4 provided no teaching pointing to MR antagonist activity, and the person skilled in the art would not“reverse” reduce calcium channel activity while simultaneously pursuing MR antagonist activity.

2. Scaffold Differences — Finerenone Case (Decision No. 630569)

Even where the prior art (Evidence 2) was likewise an MR antagonist with the same target/activity as the patent, its core scaffold was a 4-chromenone-1,4-dihydropyridine carbonitrile system, which differed significantly from the patent's 1,4-dihydro-1,6-naphthyridine amide system in terms of scaffold, ring connectivity, and substitution pattern. The Panel found that the person skilled in the art would not use it as the starting point for group substitution. Cross-scaffold “splicing and substitution” without a structure-activity relationship (SAR) transitional bridge does not fall within the routine experimentation reasonably expected of the person skilled in the art.

3. Sensitive Structure-Activity Relationship — Enzalutamide Series Cases

The enzalutamide case ((2022) Supreme People's Court Administrative Judgment No. 287; Decision No. 581944) supplemented the other side of the same logic from the perspective of sensitive structure-activity relationships (SAR): even where the core scaffold is the same and some groups are common, if the prior art as a whole does not provide directional teaching toward site-specific modification to solve the technical problem at issue — particularly where the specification discloses that the site has a sensitive SAR (minimal change causing drastic activity variation) — the existence of a technical teaching cannot be found merely because the “groups are common.” The enzalutamide patent specification explicitly stated that its compounds have a “sensitive structure-activity relationship” — seemingly minimal structural changes on the right-side aromatic ring could cause a drastic decline in activity (from Class I to Class V compounds). Under this premise, even if the prior art disclosed the existence of certain groups, if its overall teaching did not point toward introducing those groups into the specific scaffold to solve the actual technical problem, a technical teaching could not be found.

Key Takeaways: Three Advances in Inventiveness Assessment

Foreign slide decks: Classified as “public use,” subject to domestic territorial limitation; must be proven as formally published or publicly disclosed in China to serve as prior art.

Closest prior art: Shift from “no right or wrong in selection” to substantive eligibility review; references with excessive structural differences may not serve as the starting point.

Technical teaching: Shift from pure structural comparison to holistic SAR analysis; different mechanisms, scaffold differences, or sensitive SAR can each negate technical teaching.

Part III: Sufficiency of Disclosure for Compound Salts — Ruxolitinib Salt Series

Ruxolitinib Salt Series Cases  —  Decision Nos. 580173, 584867, 589066, 601005

In the ruxolitinib salt series invalidation cases, the Panel systematically addressed the issue of "sufficiency of disclosure for compound salts". The dispute focused on three levels: (1) Did the cited document disclose the JAK inhibitory activity of ruxolitinib? (2) Was the technical effect of the priority document sufficiently disclosed? (3) Did the patent's disclosure of three salts meet the sufficiency of disclosure requirement?

The Panel analyzed these issues layer by layer, with its reasoning organized into four progressive levels:

Level 1 — Confirmation of the parent compound (free base) activity. The Panel noted that although the cited document, Example 67, did not provide specific IC₅₀ values, through the separation of R/S enantiomers, the salt formation process, and the statement that they were “found to be active JAK inhibitors,” combined with the general knowledge in the field that "salt formation typically does not cause a drug compound to lose activity", the person skilled in the art could directly and unambiguously confirm that ruxolitinib (R-enantiomer) possesses JAK inhibitory activity. This finding established the factual basis for "the salt inheriting the core pharmacological activity of the free base".

Level 2 — Priority validity and citability of the reference document. The Panel adopted a pragmatic and first-to-file-consistent position. Where the cited document was published after the priority date but before the PCT international publication date, and was the applicant's own earlier application, the Panel applied the principle of "retroactive application of the more favorable law"and adopted the amended provision in the Patent Examination Guidelines (2023) that relaxed the time limitation for citing foreign patent documents, recognizing that the cited document could be referenced in the patent specification to explain the background art and technical effects.

Level 3 — Substantive content of the priority document. Although the priority document did not record specific assay data for JAK inhibitory activity, combined with the ruxolitinib JAK inhibitory use already disclosed in the cited document and general knowledge in the field, the person skilled in the art could understand the technical problem to be solved (providing pharmaceutically acceptable salts of ruxolitinib to improve formulation properties) and the expected technical effect (inheriting JAK inhibitory activity). Therefore, the patent and the priority document related to the same subject matter, and the priority was valid.

Level 4 — Preparation and confirmation of the salts themselves. The patent specification Examples 1-3 provided a general salt formation method (free base + corresponding acid, conventional crystallization and separation), described ¹H-NMR showing 1:1 salt formation, confirmed crystallinity through XRPD, and provided DSC melting peak and TGA data, all using detection means that are conventional confirmation methods in the field. Since the parent compound's activity had been confirmed and salt formation does not alter that activity, the person skilled in the art could reasonably expect that the three salts (maleate, sulfate, phosphate) would all possess corresponding JAK inhibitory activity, and the specification had provided a clear and complete explanation of the preparation, confirmation, and core effects of the salts.

Key Takeaways: Sufficiency of Disclosure for Compound Salts

Where the preparation, confirmation, and core activity of the parent compound (free base) have been substantiated by the patent specification itself or a lawfully cited document,And the patent specification provides conventional salt formation methods, basic physicochemical characterization, and an activity expectation consistent with the parent compound for the specific salt form,Then the compound salt claim meets the sufficiency of disclosure requirement under Article 26, Paragraph 3 of the Patent Law.

Conclusion

Reviewing the series of invalidation decisions issued by CNIPA during 2025-2026, compound patent invalidation examination is undergoing a transformation toward deeper substantive adjudication. This transformation spans multiple dimensions — claim interpretation, inventiveness assessment, sufficiency of disclosure, and application of procedural principles — exhibiting a clear and systematic evolutionary trajectory.In the area of claim interpretation, the sofosbuvir prodrug case and the finerenone case jointly established the interpretation rule for general formula claims of chiral compounds: a plain line encompasses all stereochemical configurations, not merely the racemate, providing innovators with stable institutional expectations for constructing scope of protection using Markush claims. In the area of inventiveness assessment, examination practice is shifting from the lenient stance of “no right or wrong in selection” to substantive eligibility review of the closest prior art, and emphasizing comprehensive analysis of structure-activity relationships, mechanisms of action, and the overall teaching of the prior art in determining technical teaching. In the area of sufficiency of disclosure, the ruxolitinib salt series cases established an integrated evaluation framework anchored on the parent 
compound's activity and supplemented by lawfully cited documents, avoiding unreasonable disclosure barriers caused by overly demanding data completeness.