Notes
IDENTITY / DISAMBIGUATION (important): SC-002 is a DLL3-directed ADC (Stemcentrx -> AbbVie) that uses the SAME antibody (rovalpituzumab) and the SAME class of PBD-dimer warhead as Rova-T, but a DISTINCT linker/conjugation: a plasma-stable Val-Ala dipeptide linker with SITE-SPECIFIC conjugation via 2 introduced cysteine residues giving a homogeneous DAR2 (~95% DAR2). Rova-T (rovalpituzumab tesirine, SC16LD6.5) uses tesirine/SG3249 with conventional conjugation. Do NOT merge SC-002 and Rova-T toxicity datasets despite shared antibody/payload and the 'rovalpituzumab' name.
CITATION CORRECTION: the wave1_input attributed the source to 'Mansfield et al.' The actual first author is Morgensztern D (Morgensztern D, Johnson M, Rudin CM, Rossi M, Lazarov M, Brickman D, Fong A. 'SC-002 in patients with relapsed or refractory small cell lung cancer and large cell neuroendocrine carcinoma: Phase 1 study.' Lung Cancer 2020;145:126-131. PMID 32438272; PMCID PMC8173700; NCT02500914).
OCULAR (qualifying signal): a SINGLE grade 2 photophobia in 1/35 patients (~2.9%), at 0.2 mg/kg Q3W (not the MTD), considered possibly treatment-related, reported only in the safety narrative (below Table 3 threshold). No corneal/conjunctival/dry-eye/blurred-vision events reported (left blank, not zero). No grade >=3 ocular AE. Reversibility not stated. CTCAE v4.03. Photophobia may be related to the prominent PBD photosensitivity phenomenon rather than ocular-surface toxicity, hence ocular_surface_subtype = Visual. This is a weak/low-magnitude OAE signal.
SAFETY-POPULATION CAVEAT: the only safety table (Table 3) is the POOLED phase 1a/1b full analysis set (n=35, all dose levels 0.025-0.4 mg/kg). There is no separate RP2D-cohort safety table (MTD cohort = 8 pts at 0.4 mg/kg Q9W). All toxicity_rows therefore reflect the pooled n=35 population; line_context records this and flags the MTD/RP2D. Treat per-organ percentages as pooled-across-doses, not RP2D-specific.
SIGNATURE TOXICITY: serosal effusions dominate (pleural 43%/20%, pericardial 26%/3%) with hypoalbuminemia (17%/6%) and peripheral edema (31%/3%) -> capillary-leak/serosal-effusion phenotype characteristic of PBD-dimer ADCs. Photosensitivity reaction 20%/3% (skin SOC; the grade 3 photosensitivity was the DLT/MTD-defining event at 0.4 mg/kg Q9W) - not captured as a toxicity_row because 'Skin' is not in the organ_system enum, but it is the headline cutaneous toxicity. Other notable AEs (not in organ enum): fatigue 31%/6%, pyrexia 23%, hypotension 17%, hyponatraemia 11%/9%, pulmonary embolism 6%/6%. Overall grade 3 in 21/35 (60%) and grade 4 in 2/35 (6%); SAEs in ~66%.
NOT REPORTED (left blank per convention, NOT zero): neutropenia; peripheral neuropathy (consistent with a non-tubulin DNA-crosslinker payload); infusion-related reactions; AST/ALT/bilirubin elevations; vomiting; diarrhea.
OUTCOME: development discontinued for systemic toxicity and limited efficacy (ORR 14.29% [PR only, no CR]).
ADCdb CROSS-REFS (DRG0UUEEH): antibody ANI0NBQGD (Rovalpituzumab); antigen TAR0HELFC (DLL3); payload PAY0LZRXU (SG3199); ChEMBL CHEMBL3990011; TTD D0D4OL. ADCdb lists DAR=2, conjugate type 'Undisclosed', status Phase 1 (Terminated), sponsor Stemcentrx/AbbVie. ADCdb does NOT disclose a linker LIN id or linker physchem (SMILES/formula/MW/TPSA/xLogP) -> adcdb linker fields left blank. Linker is qualitatively 'plasma-stable'; no numeric half-life or stability % reported.
PAYLOAD PHYSCHEM: standard SG3199 values from PubChem CID 90132565 applied per shared-payload convention (MW 584.7, formula C33H36N4O6, XLogP 2.8, TPSA 102). The actual named warhead is SC-DR002, an SG3199-class PBD dimer. logD7.4 and pKa not in the literature (blank); charge (neutral) and bystander (Yes) are tier-C inferences from PBD-dimer chemistry.
DERIVED COLUMNS (DNA-damage subtype, stability conditional/unconditional, FcgammaR/C1q, severity scores, HCA/HPA expression) intentionally left for downstream computation.