news and events

MEMP - Thesis Defense - Jingzhi An

Monday, May 6, 2019

Towards the Clinical Translation of a Physiological Closed-Loop Control System for Precise Control of Propofol-Induced Pharmacological Coma

Physiological closed-loop control (PCLC) is an emerging technology that can revolutionize the standard-of-care by ensuring timely, precise, and consistent delivery of therapy that is also cost-effective, distraction-free, and non-labor intensive. However, the clinical translation of PCLC systems has been impeded by the lack of real-world performance and safety guarantees, which are difficult to provide due to unique engineering and regulatory challenges, such as uncertain, nonlinear physiological systems and complex healthcare settings. We studied the delivery of propofol-induced pharmacological coma in patients with refractory status epilepticus (RSE) as a prototypical application to address this translational challenge for PCLC systems. To motivate the chosen application, we conducted a retrospective clinical study and showed that the clinical management of pharmacological coma in RSE patients may be improved with PCLC. We then made contributions in five areas to enhance the performance and safety of a PCLC system for delivering propofol-induced pharmacological coma. First, we built the PCLC system with explicit considerations of usability issues and regulatory requirements for preclinical evidence of safety. Second, we built a hardware-in-the-loop platform to evaluate the PCLC system under conditions mimicking real-world operation. Third, we devised new experimental methods to characterize the performance of clinical infusion pumps for PCLC applications, and performed experiments to characterize several commercially-available infusion pumps. Fourth, we systematically analyzed pharmacokinetic-pharmacodynamic models to better understand physiological response to propofol. Finally, we proposed a flexible chance-constrained stochastic optimal control framework for designing controllers with explicit performance and safety guarantees under real-world uncertainties for PCLC systems. Overall, we have taken concrete steps towards clinical translation of a PCLC system for precise control of propofol-induced pharmacological coma, and the methods developed can be applied to other PCLC systems.
Thesis Supervisor:
Emery N. Brown, MD, PhD
Edward Hood Taplin Professor of Medical Engineering, MIT
Thesis Committee Chair:
Thomas Heldt, PhD
W.M. Keck Career Development Professor in Biomedical Engineering, MIT 
Thesis Readers:
M. Brandon Westover, MD, PhD
Associate Professor, Harvard Medical School
David L. Trumper, PhD
Professor of Mechanical Engineering , MIT


Date and Time: 
Monday, May 6, 2019 - 12:00pm to 2:00pm

MIT E25-401