Ivonescimab Biologics Sequence Review Report 2026: PD-1×VEGF Similarity and Patent-Risk Signals

This Ivonescimab Biologics Sequence Review Report was produced with PatSnap Biology Modality MCP workflows. The workflow connects patent-sequence retrieval, protein similarity search, sequence fetching, pairwise alignment and antibody–antigen evidence in one reproducible review. Explore the PatSnap MCP servers used in this report.

Review date: 21 July 2026. This report is a sequence-landscape screen for R&D and IP triage. It is not a legal opinion, freedom-to-operate conclusion or validity analysis. A sequence hit does not establish infringement; “claimed” describes the returned PatSnap record and still requires jurisdiction-specific claim construction, family consolidation, legal-status verification and prosecution-history review.

Executive sequence-risk thesis

Ivonescimab (AK112/SMT112) is a tetravalent PD-1×VEGF-A bispecific antibody. The PatSnap Biology Modality MCP returned an exact 719-amino-acid heavy-chain record explicitly linked to Ivonescimab and a 214-amino-acid light chain from the core PD-1/VEGF patent disclosure. The architecture places a bevacizumab-like anti-VEGF IgG region at the front of the molecule and appends a Penpulimab/AK105-related anti-PD-1 single-chain variable fragment to each heavy-chain C terminus through glycine/serine-rich linkers.

The initial screening signal is elevated diligence priority, not an infringement finding. The complete heavy-chain search returned an exact self-match, several full-length variants above 97% identity and named PD-1/VEGF bispecific constructs BsAbB7 and BsAbB8. At the same time, the VEGF variable domain and the light chain are widely reused in patent sequences related to bevacizumab, so those modules are weak discriminators when assessed alone. The most informative review unit is therefore the combined architecture: VEGF-binding variable domains, Fc substitutions, linker placement, PD-1 VH/VL sequences and their order in the 719-residue fusion chain.

Sequence scope and MCP workflow

The review used ls_patent_sequence_fetch on EP3882275B1 to recover the sequence set, including the 719-aa full heavy chain, 214-aa light chain, 118-aa PD-1 VH, 107-aa PD-1 VL, 123-aa VEGF VH and associated CDR/linker sequences. Full records were resolved with ls_sequence_fetch. Five protein searches were then submitted through ls_sequence_search_submit, monitored with ls_sequence_search_check_status and retrieved with ls_sequence_search_get_results.

Searches covered the PatSnap all-patent protein database with gaps enabled, E-value ≤1×10⁻⁵, identity ≥70%, query coverage ≥70% and a 500-result task cap. The full heavy chain, full light chain, PD-1 VH, PD-1 VL and VEGF VH were evaluated separately. Pairwise comparisons were executed with ls_sequence_alignment, while ls_antibody_antigen_search supplied target-specific patent evidence for PD-1 and VEGF-A.

Similarity search readout

Reading the counts correctly. The totals above are capped search records, not unique patent families, independent inventions or enforceable claims. A single sequence may appear in multiple publications, examples, comparators and jurisdictions. Deduplication by family, priority, applicant and claim status is mandatory before business or legal decisions.

Pairwise alignment: where the close variants differ

The closest non-identical full-length record aligned across all 719 positions with 716 identities, 717 positive substitutions and no gaps, equivalent to 99.58% identity. All three differences were localized rather than caused by insertions or deletions. That profile is consistent with an Fc or developability variant built on the same overall PD-1×VEGF architecture.

The named BsAbB7 heavy chain aligned across all 719 positions with 702 identities, 711 positive substitutions and no gaps, or 97.64% identity. Differences occurred in the Fc region and within the PD-1 variable domains while the bevacizumab-like VEGF-binding segment, linker logic and overall fusion layout remained strongly conserved. One ambiguous residue was present in the returned patent sequence, so the record should be checked against the filed sequence listing before a definitive claim chart is prepared.

These alignments show why a single percentage is insufficient. Identity in the Fc constant region has limited target specificity, whereas substitutions in CDRs or framework positions close to the antigen interface may have disproportionate functional and claim significance. The next-stage review should annotate each mismatch by domain, Kabat/IMGT position, CDR status, linker boundary and claimed SEQ ID.

Antibody–antigen evidence

The target-specific search returned 17,272 patent-sourced PD-1 antibody–antigen records and 3,398 patent-sourced VEGF-A records before family consolidation. These broad counts illustrate the density of the two target spaces; they are not counts of blocking patents.

• PD-1 module: a Penpulimab-filtered search returned a direct patent record in WO2025193962A1 with 118-aa heavy- and 107-aa light-variable sequences, matching the dimensions used in the Ivonescimab fusion-domain review.
• VEGF-A module: the Bevacizumab-filtered search returned 20 records, including US20250263675A1, WO2023240067A2, US9567403B2 and US10894823B2. The evidence supports widespread reuse of bevacizumab-derived sequences in mono-, bi- and trispecific formats.
• Combined construct: the exact 719-aa MCP record names Ivonescimab and describes an anti-VEGF IgG1 fusion to anti-PD-1 VH and VL domains through glycine/serine linkers. This full-chain record is substantially more asset-specific than either target-domain search alone.

Patent-family screen

EP3882275B1, titled “Anti-PD-1 and anti-VEGFA bifunctional antibody, pharmaceutical composition thereof and use thereof,” is shown as active, granted on 18 September 2024, with an anticipated 2039 expiration before adjustments or jurisdiction-specific events. Its claims describe the paired VEGF and PD-1 CDR sets, variable-region sequences, bispecific arrangement and linker options. The family also includes active or pending members in the United States, China, Japan, Korea, Canada and other jurisdictions.

A separate Paragon Therapeutics patent, US12673997B2, issued on 7 July 2026, discloses Ivonescimab heavy and light chains as SEQ ID NOs 128 and 275 and claims broad classes of PD-1/VEGF binding proteins using sequence-identity thresholds. Its presence is an adjacent-claim signal requiring counsel review; the mere use of Ivonescimab as a reference or comparator does not show that Ivonescimab practices the issued claims.

Diligence actions for an Ivonescimab or PD-1×VEGF program

1. Confirm the exact development construct and distinguish clinical, commercial, analytical and patent-listed sequences.
2. Create a domain map for VEGF VH/VL, Fc substitutions, both linker segments and PD-1 VH/VL; assign every alignment difference to that map.
3. Consolidate sequence hits by INPADOC/simple family, earliest priority, applicant and live claim status.
4. Build claim charts for active core and adjacent families, including EP3882275B1, relevant U.S. continuations and newly issued PD-1/VEGF claims.
5. Review background licenses covering Penpulimab/AK105, bevacizumab-derived domains, Fc engineering and multispecific platform rights.
6. Repeat the search using CDR-only, linker-junction and nucleotide queries to capture sequences hidden by full-chain similarity thresholds.

Bottom line

Ivonescimab has a clearly traceable patent-sequence architecture and sits in a densely patented PD-1/VEGF landscape. The MCP results identify an exact named 719-aa heavy-chain record, numerous highly similar full-length constructs, extensive reuse of the component domains and active patent families directed to the combined format. The appropriate conclusion is not “blocked” or “clear”; it is that the asset warrants prioritized, domain-aware claim analysis before licensing, design-around or development decisions.

Explore PatSnap Open Platform MCP Servers to produce sequence similarity and patent-risk reports for antibodies, proteins, peptides and nucleic-acid assets.

Sources and methodology notes

Core sequence and target evidence was retrieved through PatSnap Biology Modality MCP on 21 July 2026. Patent-status context was checked against EP3882275B1 and the published record for US12673997B2. Database coverage, legal status and annotations can change; rerun searches at the decision date.