PURPOSE: To elucidate the mechanistic basis for efficacy of intrathecal rituximab. We evaluated complement activation as well as the pharmacokinetics of intraventricular rituximab in patients who participated in two phase 1 multicenter studies. EXPERIMENTAL DESIGN: We evaluated complement activation as a candidate mediator of rituximab within the central nervous system (CNS). Complement C3 and C5b-9 were quantified by ELISA in serial cerebrospinal fluid (CSF) specimens after intraventricular rituximab administration. We determined rituximab concentration profiles in CSF and serum. A population three- compartment pharmacokinetic model was built to describe the disposition of rituximab following intraventricular administration. The model was derived from results of the first trial and validated with results of the second trial. RESULTS: Complement C3 and C5b-9 were reproducibly activated in CSF after intraventricular rituximab. Ectopic expression of C3 mRNA and protein within CNS lymphoma lesions was localized to myeloid cells. Constitutive high C3 activation at baseline was associated with adverse prognosis. A pharmacokinetic model was built, which contains three distinct compartments, to describe the distribution of rituximab within the neuroaxis after intraventricular administration. CONCLUSIONS: We provide the first evidence of C3 activation within the neuroaxis with intraventricular immunotherapy and suggest that complement may contribute to immunotherapeutic responses of rituximab in CNS lymphoma. Penetration of rituximab into neural tissue is supported by this pharmacokinetic model and may contribute to efficacy. These findings have general implications for intraventricular immunotherapy. Our data highlight potential innovations to improve efficacy of intraventricular immunotherapy both via modulation of the innate immune response as well as innovations in drug delivery.