• Sponsored Content

Abel Hastings

December 14, 2018

4 Min Read

On 24 October 2018, BPI presented a free “Ask the Expert” webinar with Abel Hastings, director of process sciences at Fujifilm Diosynth Biotechologies. He discussed the use of systematic tools to expedite process characterization and maximize reliability of process validation campaigns.

Hastings’s Presentation
As a project moves from clinical manufacturing toward process validation — and ultimately toward preapproval inspections — project timelines can become hypervisible at all levels of an organization. Missteps can be costly. The commercial viability of a program can be at stake.

It’s important to maintain focus. Fujifilm Diosynth Biotechologies (FDB) uses systematic tools to strike a balance between science and speed to product launch. Science is paramount, with risk-based focus on activities that are of greatest value to the project. The company has built a suite of such tools for process characterization and validation called Risk Assessment Process Template Application (RAPTA). It helps drive a systematic balance of discipline and flexibility to address a range of strategic directions and indications for different molecules. This system has been used for a many products, ranging from biosimilars to breakthrough molecules.

The first tools work in parallel: A process attribute map provides systematic process understanding to drive consensus about control and identify the new, unique, and difficult elements (NUDs) that make each molecule special. Meanwhile, a risk-based failure modes and effects analysis (FMEA) yields preliminary classification for parameters and enables data-driven decisions about equipment.

Those activities help FDB develop forward-looking process characterization for manufacturability. Next comes a systematic and quantitative assessment of acceptable and/or normal operating ranges (AORs/NORs) to predict future performance and process performance qualification (PPQ).

A process attribute heat map shows a consensus of the scientific logic behind the process itself, with unit operations on its horizontal access and critical quality attributes (CQAs, the release panel) on its vertical. The heat map vormat shows the strength of links between them.

Risk assessment is prepopulated based on FDB’s experience with more than 300 molecules and adapted to a given FMEA. This involves three valuable factors: risk-priority numbers for each unit operation (to determine the level of characterizations necessary based on risk tolerance), sorted occurrence and detection scores for each unit operation (to identify equipment mismatches or improvements that could reduce risk), and sorted severity scores by themselves (to predict key and critical process parameters, KPPs and CPPs, that drive the scope of process characterization).

Process characterization should focus on enabling biomanufacturing. Three assumptions make this a data-driven campaign: that the input scores are data driven (using NORs based on process-specific equipment capabilities), that the FMEA scores can be predictive of both KPPs and CPPs, and that the difference between the AOR and NOR enables reliability. FDB believes that that safety margin between those ranges should be wide for KPPs and wider for CPPs. Finally, process characterization should establish functional AORs for the manufacturing group’s needs.

Questions and Answers
When and how do I consider specifications in process validation? Specifications come first. Process performance qualification starts with the specification and works backward. Understanding where a specification comes from is pivotal. We like to bunch them into three categories — safety, efficacy, and manufacturing reliability — with the sources and implications on manufacturing considered from day one.

Should we broaden the operation range for downstream processing? Selecting ranges for process characterization should be based on expected equipment variability. We work backward from that to build our process characterization studies. So we consider widening those ranges for CPPs and/or KPPs based on scoring.

Do you recommend a particular design of experiment (DoE) approach or linkage experiments? It is very easy to jump into the statistical bandwagon with the latest or most complex DoE. FDB has built a decision tree that helps us choose a full-factorial design or response-surface when necessary. One-factor-at-a-time (OFAT) studies often are the most effective methodology. There isn’t one solution. Use the right study for a given application. Linkage studies provide a great opportunity to cover a design space broadly and look for secondary/tertiary interactions with a multifactorial or more complex DoE. If there are no linkages, then use an OFAT.

More Online
Find the full presentation of this webcast — with more questions and answers — on the BioProcess International website at the link below.

Watch and listen online at www.bioprocessintl.com/category/webinars.

You May Also Like