In Reply: Application and Safety of Externally Controlled Metronomic Drug Delivery to the Brain by an Implantable Smart Pump in a Sheep Model

To the Editor: In their letter,1 Mingsheng Huang and Yiheng Liu refer to our article,2 describing the performance of an implantable smart pump (ISP) in an animal model. In that study, the ISP was implanted into the shoulder region of several sheep and delivered methotrexate to the brain ventricles through a connected catheter. We showed that the ISP functioned as expected over the course of the study; i.e., it remained responsive to the wireless control that enabled external adjustments of metronomic methotrexate output even after pump implantation and while the sheep were kept outdoors in a pen.2 Although generally commending our work, Huang and Liu1 are noting several potential limitations of the ISP, in particular in view of the device's possible long-term patient health in future clinical applications. Indeed, we thank the authors for their thoughtful comments and do not disagree with the valuable points they raise. We already are investigating some of these important aspects. For example, our work pursues the development of a modified device with a dual-lumen catheter that will have retrograde flow to allow sampling of the cerebrospinal fluid for the determination of in vivo drug concentrations. An additional unit with a refrigerated drug compartment is in the works as well. Both features will expand the functionality of the ISP. Regarding the authors' comment1 concerning the lack of long-term clinical follow-up data that hinders the evaluation of the ISP's clinical feasibility: Our study2 extended over 12 weeks and did not reveal any detrimental safety issues in the sheep model during this time, and the lack of the emergence of inflammatory markers could be an indicator that much longer time frames should be possible. However, further experimental and clinical evidence is needed to establish the long-term biocompatibility of ISP. Although 12 weeks of intervention might suffice for applications in the context of malignant brain cancer treatments, certain other envisioned applications, such as those aimed at neurological diseases for example, might require substantially longer interventions. In addition, as Huang and Liu have pointed out, the results from a specific animal model may not always be generalizable for clinical applications. In all, our study2 was designed as a preclinical proof-of-concept study, and now that we have successfully achieved its specific goals, we have set the stage from which to pursue the next critical steps in ISP's translational development and clinical testing, including aspects that were constructively indicated by Huang and Liu.