Analyzing Queueing Applications in Chimeric Antigen Receptor (CAR) T Cell Therapy Supply Chains

Abstract

This thesis sets out to investigate the effect of queuing on a Mixed Integer Linear Program (MILP) representing a CAR T therapy supply chain. CAR T-cell therapies are innovative genetic cancer treatments that require a unique dose to be developed for every patient. Due to their high manufacturing cost and “made-to-order” distribution process, researchers have attempted to understand how the end-to-end manufacturing and distribution supply chain can be optimized to minimize cost and/or delivery time to patients. Current work has used MILP’s to model the end-to-end CAR T supply chain, but fail to address the potential of queueing as a mechanism of optimizing cost. In this thesis, a queue mechanism for incoming patient T-cell samples was implemented at the beginning of therapy manufacturing in a baseline MILP developed by Triantafyllou et al. (2021). The tradeoff between total production cost and average turnaround time was analyzed under operational stresses such as clustered patient demand and manufacturing utilization caps. The results demonstrated that across multiple model scenarios, queuing provides outsized manufacturing cost savings (often ~35-50%) with a disproportionately minimal turnaround time sacrifice (~10.5%). Further, it was shown that queueing allows the MILP to use available manufacturing site capacity more productively under utilization constraints. This work is the first comprehensive analysis of the applications of a queue in literature formulating the CAR T supply chain as an MILP. It shows that queuing is a valuable strategic tool that can significantly reduce production costs, and thereby expand CAR T access whilst protecting supply chain feasibility across diverse operational conditions.