Notes
Zanidatamab zovodotin (ZW49): biparatopic HER2 ADC = zanidatamab/ZW25 (humanized IgG1, Azymetric heterodimer binding HER2 ECD2 + ECD4) + ZymeLink Mc-PEG3-Val-Cit cleavable linker (random interchain Cys/maleimide, DAR 2) + proprietary N-acyl sulfonamide auristatin payload ZD02044 (tubulin inhibitor). Sponsors Zymeworks/BeiGene; NCT03821233 (ZWI-ZW49-101). PROGRAM DISCONTINUED (Phase 1 terminated); NO peer-reviewed full publication exists - all clinical data are conference-level (AACR-NCI-EORTC 2023 poster B130, the most detailed/final cut, data cutoff 10 Jul 2023, n=67; ESMO 2022 oral abstract 460MO, earlier cut n=76). Hence all clinical/toxicity cells are tier D.
OCULAR (dominant, dose-limiting toxicity): corneal keratitis is the single most common TRAE. RP2D = 2.5 mg/kg Q3W (n=31): combined keratitis/punctate keratitis 45% any-grade (14/31), 3% Gr>=3 (1/31). Pooled DE+DX (n=67): 49% any-grade (33/67; 11 max G1, 22 max G>=2), 3% Gr>=3 (2/67). Grade>=2 ophthalmic AESI breakdown (pooled): keratitis 28%, punctate keratitis 4%, vision blurred 4%, dry eye 4%, visual impairment 1% (composite Gr>=2 ophthalmic events 37% pooled / 39% at RP2D). The single DLT at RP2D was a Grade 2 keratitis lasting >14 days (resolved to <=G1); other DLTs (1.5 mg/kg QW) included keratitis and a Grade 3 blurred-vision DLT (1.25 mg/kg QW). Keratitis is reversible (improved/resolved with dose reduction/delay + steroid and lubricating eye drops). MANDATORY ocular prophylaxis implemented: prednisolone + tetrahydrozoline/naphazoline (or equivalent) eye drops + cooling eye masks during infusion. 20/67 (30%) had dose reductions (14 [21%] for ophthalmic events: keratitis 17%, dry eye 3%, blurred vision 3%); 8/67 (13%) discontinued for TRAEs (mostly keratitis). ocular_surface_subtype = Corneal (off-target).
OAE_ANY_PCT decision: used the RP2D (2.5 mg/kg Q3W) cohort value of 45% per instruction ('max across ocular PTs at RP2D'), consistent with the RP2D-cohort convention used for existing rows (e.g., disitamab-vedotin). Alternative documented cuts: 49% (AACR 2023 pooled DE+DX n=67); 42% (ESMO 2022 pooled n=76); ~44% appears in an interim abstract. ae_dry_eye_pct (6%) and ae_blurred_vision_pct (3%) are RP2D Grade>=2 AESI values; any-grade for those PTs was not separately tabulated, so they may understate any-grade.
OTHER ORGAN SYSTEMS: Diarrhoea 24% pooled / 29% at RP2D (GI; nausea/vomiting below the >=20% threshold). IRR 19% pooled (Gr>=3 4%, incl. 1 Grade 4 IRR) / 16% at RP2D. One Grade 4 neutrophil count decreased (the only documented heme event). Alopecia 25% (skin off-target; not a tracked target-organ system). DOCUMENTED ZERO ILD/non-infectious pulmonary toxicity and 0 treatment-related deaths. Hepatic, hematologic (any-grade), and peripheral-neuropathy rates are NOT tabulated (poster reports only TRAEs >=20% plus Gr>=3 in >=3 pts), so those master per-organ columns must be left BLANK (not 0). CTCAE v5.0; overall Gr>=3 TRAE rate 21% (14/67).
PAYLOAD PHYSCHEM (ZD02044): NOT a standard payload, so values were DERIVED. The released free auristatin ZD02044 = C37H64N6O8S, MW ~753, computed from the MT-VC-ZD02044 linker-payload (CAS 1800462-96-7, C61H101N11O17S, MW 1292.58; Chemsrc/MedChemExpress SMILES) minus the Mc-PEG3-Val-Cit linker (ADCdb LIN0KNQWV, C24H39N5O10) + H2O; RDKit gave TPSA 180.7 and Crippen MolLogP 2.72 (vs RDKit MolLogP 3.87 / PubChem XLogP3 4.1 for MMAE), i.e. less hydrophobic than MMAE - corroborating Liu et al. (XB002) Mol Cancer Ther 2025 PMC11791478 ('zovodotin less hydrophobic than vedotin'). Net charge ~0 (zwitterion: +1 N,N-dimethyl tertiary amine, -1 acyl sulfonamide pKa~4-5). payload_logd_7_4, payload_pka, and payload_bystander left blank (no public data; structure predicts limited bystander vs MMAE). hydrophobicity_masking_strategy = Discrete-PEG-spacer (the discrete PEG3 in Mc-PEG3-Val-Cit).
DATA-QUALITY NOTE: the ADCdb linker DISPLAY SMILES for LIN0KNQWV encodes a guanidine/arginine side chain ('CCCNC(N)=N'), but the linker is Val-CITRULLINE; ADCdb's own computed formula (C24H39N5O10, 5 N), MW (557.601), TPSA (215.69) and XLogP (-1.49) all match the correct citrulline (urea) structure exactly (RDKit-verified). The reported linker_smiles here is corrected to citrulline ('CCCNC(N)=O'); the numeric ADCdb values are retained as-is (correct). Derived columns (dna_damage subtype, stability conditional/unconditional, FcgammaR/C1q, severity, HCA/HPA) intentionally left for downstream derivation per instructions.