We previously showed that low-dose fenretinide (200 mg/d) had limited activity in retinoid-resistant oral leukoplakia (34% response rate) possibly because serum drug levels were insufficient to induce retinoid receptor-independent apoptosis. Therefore, we designed the single-arm phase II trial reported here to investigate whether higher-dose fenretinide would improve leukoplakia response over that of our previous study. Leukoplakia patients received fenretinide (900 mg/m(2) twice daily) in four 3-week cycles (1 week on drug followed by 2 weeks off). At week 12, clinical responses were determined and blood samples were collected for serum drug level assessments. A planned interim futility analysis led to early trial closure after the initial 15 (of 25 planned) patients because only 3 (20%) had a partial response (stopping rule: <or=4 responses in first 16 patients). Fenretinide was well tolerated--only one grade 3 adverse event (diarrhea) occurred. Serum fenretinide levels changed from 0 (baseline) to 0.122 +/- 0.093 micromol/L (week 12). In correlative in vitro studies, high-dose fenretinide inhibited the growth of head and neck cancer cells more and oral leukoplakia cells less than did lower doses of fenretinide. This result is consistent with our clinical finding that high-dose fenretinide did not improve on the historical response rate of lower-dose fenretinide in our previous oral leukoplakia trial.