Understanding the facility's original capabilities
Although reasonably flexible, B633 had been originally designed to manufacture only one to two commercial products per year.
With MedImmune's changing business focus and robust pipeline, the company now had to target multiple late-stage clinical candidates
and fixed commercial requirements on an annual basis. By analyzing the long-term clinical and commercial demands on the facility,
not just manufacturing capacity but also technical transfer and postmanufacturing support, a team was able to identify immediate
areas of opportunity to increase capacity. Additionally, longer-range initiatives were identified.
First, the company modeled the plant and processes to identify potential bottlenecks. Because the manufacturing plant was
now required to run multiple products, archetypical processes were developed that reflected the range of titers that could
potentially occur with the product portfolio. Large-scale manufacturing plants such as B633 are designed with a certain amount
of throughput. The throughput of the plant is a function of run rate. The capacity of the plant is driven by a combination
of run rate and plant idle time. The idle time of the plant basically resides in two areas, one being the time to evaluate
and characterize products that have been first introduced to the plant and scaled to 15k bioreactor scale and the other being
plant changeover from one product to the next. In case of product evaluation, most commercial manufacturing facilities do
not require the time to evaluate products between batches; therefore, this switch is not a significant factor influencing
capacity. However, this element represents greater influence in a late-phase clinical program where scaling the process beyond
the pilot plant level is common practice.
Likewise, product changeover can be fairly influential in determining overall plant capacity. For example, if a changeover
process requires 10 days, then one has to consider the impact on available overall plant capacity if it has two products going
through the facility per year versus five products. In the former case, one could attribute plant changeover to a 20-day reduction
in plant capacity. In the latter case, 50 days of capacity are lost to changeover time. Some key questions emerge: What does
capacity look like in the current state versus the optimized state? And how aggressively can changeover time be improved to
MedImmune examined ways to reduce the second component of idle time—that is, product evaluation. In this case, a team developed
a risk-based approach to evaluate and characterize products with the goal of minimizing the time between batches. The ability
to model and demonstrate consistency of scalability from bench to commercial scale for the bioreactors provided a high degree
of assurance that product quality would not be impacted as the process is scaled.
During the processing of the initial batch from a campaign at scale, the team gathered further evidence by conducting laboratory
scale purification studies to evaluate harvested material from the production reactor for product compared with material produced
earlier in the development cycle at the pilot scale. This analysis typically occurs simultaneously with the large-scale purification
process and data from the "scouting study" is available prior to the end of the large-scale purification.
Lastly, the team looked at ways to minimize differences in the overall production process required for multiple products.
Rather than investing time and money focusing on changing the facility itself because the manufacturing processes are different,
the team evaluated the option of approaching each new product candidate employing a common process that can be broadly used
for a range of products in the development portfolio. The use of this platform-process approach by process development enables
the site to meet the goal of minimizing changes to the facility to accommodate new products.