Modafinil is a candidate compound for the treatment of cocaine addiction that binds to the dopamine transporter (DAT) in healthy humans, as observed by positron emission tomography (PET). This mechanism, analogous to that of cocaine, might mediate a putative therapeutic effect of modafinil on cocaine dependence, though the binding of modafinil to DAT has never been assessed in cocaine-dependent patients. We aimed at quantifying the DAT availability during a controlled treatment by modafinil, and its clinical and psychometric correlates in cocaine-dependent patients at the onset of abstinence initiation. Twenty-nine cocaine-dependent male patients were enrolled in a 3-month trial for cocaine abstinence. Modafinil was used in a randomized double-blind placebo-controlled design and was administered as follows: 400 mg/day for 26 days, then 300 mg/day for 30 days, and 200 mg/day for 31 days. Participants were examined twice during a 17-day hospitalization for their DAT availability using PET and [(11)C]-PE2I and for assessments of craving, depressive symptoms, working memory, and decision-making. Cocaine abstinence was further assessed during a 10-week outpatient follow-up period. Baseline [(11)C]-PE2I-binding potential covaried with risk taking and craving index in striatal and extrastriatal regions. A 65.6% decrease of binding potential was detected in patients receiving modafinil for 2 weeks, whereas placebo induced no significant change. During hospitalization, an equivalent improvement in clinical outcomes was observed in both treatment groups, and during the outpatient follow-up there were more therapeutic failures in the modafinil-treated group. Therefore, these results do not support the usefulness of modafinil to treat cocaine addiction.