Notes
PRIMARY-SOURCE UPDATE: Contrary to the wave1 hint ("aggregate % not publicly reported"), ClinicalTrials.gov posted FULL results for NCT04622774 on 2025-01-15. Study = Phase 1 FIH (of Phase 1/2), 56 patients: Schedule A dose escalation 0.5-7.0 mg/kg IV Q3W (21-day cycle, DLT window 21d), a fractionated Schedule B (2.0 mg/kg Days 1,8,15 then 3.0 mg/kg Days 1,8 of 28-day cycle, n=3), and dose expansion at RP2D 6.0 mg/kg Q3W in NSCLC (n=13) and TNBC (n=6). RP2D = 6.0 mg/kg Q3W (protocol explicitly conducts expansion "at the RP2D"; CT.gov shows expansion at 6.0 mg/kg). I tagged the NSCLC expansion (largest single RP2D cohort, n=13) as the primary RP2D ocular line.
OCULAR (why this ADC qualifies): Dominant, dose-dependent corneal/ocular-surface toxicity (classic off-target maytansinoid). Vision blurred rose 33.3% (1.0 mg/kg) -> 60% (6.0) -> 77.8% (7.0). At RP2D 6.0 mg/kg (NSCLC n=13): vision blurred 61.5%, dry eye 38.5%, photophobia 30.8%, punctate keratitis 23.1%, visual impairment 23.1%, keratitis 15.4%, keratopathy 15.4%. Other corneal PTs across cohorts: corneal epithelial microcysts and cornea verticillata (Schedule B). Sponsor (MGC028 paper PMID 41166694) states IMGC936 was "affected by dose-limiting ocular adverse events, including blurred vision and keratopathy... no other prominent safety signals." oae_g3plus_pct left blank: CT.gov posted AE module is all-grade (not CTCAE-graded) and no ocular event was serious; grade-specific ocular incidence is not publicly disclosed.
CLEAN-ELSEWHERE PROFILE (documented absences via CT.gov threshold-0 AE table, all 56 pts): NO neutropenia/neutrophil count decreased anywhere; peripheral neuropathy only 2 isolated low-dose cases (none at RP2D). I recorded these as pct=0 RP2D rows because the threshold-0 table documents the absence (corroborated by the "no other prominent safety signals" statement) — not an imputed 0.
PAYLOAD DISCREPANCY FLAG: ADCdb lists ADC synonym "AEX6003 (S442C)-DGN549" yet its payload field is "DM21-C" with a "Microtubule" mechanism. DGN549 is an unrelated ImmunoGen IGN (DNA-alkylating) payload; ALL authoritative sources (MacroGenics, AACR 2019 abstr 3898/538, AACR 2021 abstr 1841, the protocol's "Maytansinoids are known to cause ocular toxicity," and the MGC028 paper) confirm IMGC936's payload is DM21, a next-generation maytansinoid (microtubule inhibitor), DAR 2.0, site-specifically conjugated via engineered cysteine S442C with a cleavable Mc-Ala-Ala-Ala tripeptide linker; major catabolite DM51. Treated the DGN549 token as a likely ADCdb data error.
PAYLOAD PHYSCHEM: DM21 is proprietary and not in PubChem; exact MW/xLogP/logD/TPSA/pKa/charge are not publicly disclosed, so left blank (DM21 is NOT one of the standard shared payloads, so DM1/DM4 values were deliberately NOT substituted). Bystander = yes (designed for efficient hydrophobic catabolite release vs DM4).
SOURCE CHECK: Of the wave1 best_source list, the Oncologist 2024 ADC ocular review (academic.oup.com/oncolo/article/29/11/e1435) was fetched and verified to NOT mention IMGC936; the authoritative quantitative data come from CT.gov posted results (used here). Antibody isotype not disclosed by ADCdb (left blank); antibody is a humanized mAb (MGA021) bearing an engineered S442C cysteine in CH3 for site-specific conjugation. Development discontinued 2024 (MacroGenics/AbbVie) for not meeting pre-set safety/efficacy benchmarks (ocular toxicity constrained the therapeutic window); follow-on ADAM9 ADC MGC028 switched to an exatecan topoisomerase-1 payload specifically to avoid the maytansinoid ocular toxicity. Note: ADAM9 is expressed in cornea/retina (ADAM9 loss-of-function causes cone-rod dystrophy), so a minor on-target corneal contribution cannot be fully excluded, but the toxicity pattern matches generic off-target maytansinoid corneal epithelial toxicity, hence ocular_surface_subtype = Corneal (off-target). All toxicity_rows are all-cause TEAEs (CT.gov), all-grade, tier B (primary posted results). Saved local copies: protocol text /tmp/imgc936_prot.txt; CT.gov JSON /tmp/imgc936.json.