Purpose: Irofulven (6-hydroxymethylacylfulvene) is a novel agent, derived from illudin S, with potent apoptotic effects in preclinical models. In the Phase I trial evaluating intermittent weekly schedules, visual symptoms were dose limiting. The aim of this analysis was to better characterize the visual adverse events of irofulven and provide treatment guidelines. Experimental Design: Clinical data from 277 patients entered in single-agent Phase I to II clinical trials who received irofulven on days 1 and 15 every 4 weeks; days 1, 8, and 15 every 4 weeks; or days 1 and 8 every 3 weeks were included in this multiparameter analysis. Results: Overall, 74 patients (27%) experienced visual symptoms. The most frequently reported symptoms were flashing lights (12% of patients), blurred vision (9%), and photosensitivity (8%). Grade 3 toxicity was observed in 12 patients (4%). The incidence and severity of visual events were dose dependent, with no grade 3 visual events occurring at doses ≤0.50 mg/kg and grade 1 to 2 events in only 12% and 8% of patients, at doses of ≤0.50 mg/kg and ≤20 mg/m2, respectively. Grade 1 to 2 toxicity was reversible in most patients. Abnormal electroretinogram and abnormal visual fields were noted after irofulven treatment in 24 of 39 patients (62%) and 15 of 26 patients (58%), respectively. All but 1 patient who had electroretinogram assessment received doses >0.50 mg/kg. Clinical examination and visual field assessment were found to be better correlated with symptoms and appear to be more appropriate for surveillance of irofulven retinal symptoms than electroretinograms. Conclusions: On the basis of retained antitumor activity and reversibility of grade 1 and 2 visual symptoms at lower doses, it appears that an irofulven dose of ≤0.50 mg/kg or ≤20 mg/m2, not to exceed 50 mg in a single dose, given as a 30-minute infusion OH days 1 and 8 every 3 weeks or days 1 and 15 every 4 weeks minimizes the frequency and severity of visual symptoms.