Abstract: Objective To systematically analyze the clinical characteristics of ocular adverse events (OAE) associated with anti-cancer agents based on data from the FDA adverse event reporting system (FAERS), and to clarify the differences among different drug classes in terms of affected sites, risk intensity, and temporal patterns of OAE, so as to provide evidence for clinical risk-stratified monitoring. Design Retrospective study. Participants Data reports from the FAERS database spanning 89 quarters, from the first quarter of 2004 to the first quarter of 2026. Methods After applying FDA-standard deduplication strategies, keyword stem matching, all OAE-related reports in the FAERS database were included. Disproportionality analyses, including reporting odds ratios and information components, were performed on the signal intensities of system organ classes (i.e., affected sites/organs) and preferred terms (i.e., signs/symptoms of OAE manifestations) for various anti-cancer agents, to quantitatively assess the intensity and severity of OAE associated with each drug class. Descriptive analysis was conducted on the time to onset (TTO) of OAE associated with anti-cancer agents. Main Outcome Measures Signal intensity of system organ classes and preferred terms, reporting rates, distribution of vision-threatening OAE, and TTO. Results Antibody drug conjugates (ADC) were associated with a higher overall risk of OAE, being the most prominent at both the single agent and preferred term levels, and exhibited marked payload subtype heterogeneity; ADC with the monomethyl auristatin F (MMAF) payload demonstrated the most extensive corneal toxicity. Subgroup analysis of tyrosine kinase inhibitors (TKI) revealed that the fibroblast growth factor receptor (FGFR) inhibitor subclass carried the highest risk of ocular toxicity, suggesting that the overall ocular risk of TKI is primarily driven by the FGFR-targeting subclass. At the preferred term level, the strongest signal was belantamab associated keratopathy (information component value=12.49). Time to onset (TTO) varied across anti-cancer drug classes: the median TTO for ADC was relatively long (228 days), suggesting that long term safety monitoring is warranted in clinical practice. Conclusion OAE associated with anti-cancer agents exhibit distinct class-specific characteristics. ADC demonstrated particularly prominent ocular toxicity signals and showed clear payload-dependent differences, with MMAF-containing ADC being characterized by a high risk of corneal disorders. TTO analysis suggested that ocular safety monitoring for ADC should be extended long time.

Key words:  FAERS database, Ocular adverse events, Anti-cancer drugs, Disproportionality analysis, Antibody-drug conjugate