BPI Europe 2022: Statistical Modeling, Process Analytical Technology and QbD

BioProcess International took place in Vienna this year, with talks centred around six main streams including cell line development and engineering, cell culture, recovery and purification, viral safety, cell and gene therapy as well as smaller sessions for manufacturing strategy and Industry 4.0 and vaccine manufacturing.  

IDBS: Sponsor and presenter 

IDBS – a sponsor of the conference – had several associates attend the conference to reconnect with customers in person for the first time since the pandemic began. Alberto Pascual gave a keynote speech titled “Building a Digital Data Backbone to Improve BioPharma Lifecycle Management (BPLM)”, which brought many attendees to our booth to learn more about Polar and PIMS. Data contextualization, data visualization, our data backbone and continued process verification (CPV) reporting were the major areas of interest.  

Statistical modeling in process development 

Several speakers addressed the need for statistical modeling and mechanistic modeling in process development. Statistical modeling uses big data to predict trends and patterns in data sets and builds a model through experimental experience. The model improves over time as more data is generated within the specified parameters to which the model is calibrated. Mechanical modeling is based on natural laws of science rather than experimental data. Both physical and biophysical principles are used to calibrate the model and determine unknown parameters. These models can also extrapolate beyond the process design space which is beneficial when testing parameter variations and coping with unexpected deviations. To advance process understanding, mechanical and statistical models are being used to develop digital twins, an in silico approach, as it enables scientists to easily change process setup and parameters without the need for large data sets.  

Is the industry ready to adopt simulation and modeling tech?

During a keynote address, Per Lidén of Cytiva, shared his vision of in silico process development and his belief that the industry is ready to adopt simulation and modeling technology to work more efficiently. The desired outcome of simulation in industry is to help improve robustness, process understanding and reduce the number of experiments for each stage of development. The yields and cost are secondary to this trend as in silico process development can have an enormous impact on each stage of process development including process definition, optimization and troubleshooting, continuous improvement and tech transfer. 

Mechanistic modeling of chromatography was a common example to demonstrate an in silico experiment. Using the mathematical interpretation of the biophysical and physicochemical interactions on the chromatography column, an in silico experiment will take approximately five to ten seconds to generate copious quantities of data in contrast to design of experiment (DoE)-based experiments, which take four to eight hours for a single experiment. This can save scientists up to 75% of lab experiments once the calibration experiments are completed to determine the constants of the simulated experiment. With constant pressures to develop and transfer processes faster and more predictably, the industry is ready to adopt simulation and modeling technology to work more efficiently.  

Process analytical technology is gaining pace 

Process analytical technology (PAT) was another hot topic, which also uses modeling and in silico simulations. PAT is the robust sensor technologies in situ for the equipment and the conditions during process development activities. For example, during upstream processes, PAT sensors can measure cell density and cell viability in real time, which provides 100% of batch monitoring, reduced risk of contamination, less hands-on activity for operators and increased quality and productivity.

Sartorius used the conference to launch their new PAT sensor toolbox, BioPAT which can be used with the Biostat RM and STR bioreactors. BioPAT is a predictive analytics tool for bioreactor scaling to identify the more intense seed train for maximum product/cell yield and reduce the overall cost of goods. This approach will help customers achieve quality by design (QbD) which will reduce waste, meet compliance criteria and get products to market faster.  

QbD is the future 

Continuing the topic of QbD, a presentation given by Isabelle Lequeux from BioPhorum highlighted a best practice approach for the registration of innovative and complex raw materials for proper product lifecycle management. The process begins with a target material profile (TMP) to summarize the characteristics of the raw material followed by a summary of its performance attributes. Next, a control strategy is developed to assure process performance and product quality, and finally critical material attributes (CMA) are defined to ensure product quality and safety.

A case study of a virus retentive filter was presented as the first of many types of raw materials to go through this process using the collaborative effort of the BioPhorum members. This method can help biopharma to simplify their regulatory filings as this approach also has the backing of national health authorities. A round table which followed this presentation highlighted the need for standardized methods in many areas of process development and how the industry can work together to provide a best practice for everyone because, at the end of the day, patient safety is still the number one priority.  

The industry is constantly innovating. What will we see next?

Our time in Vienna showed that the industry is looking to move faster and smarter. A lab at Technische Universität Berlin has developed a fully automated bioprocess development device including a mobile lab assistant. QbD principles ensure appropriate registration of starting materials and risk assessments are used as a key repository for critical knowledge for submission of regulatory filings. Process intensification delivers higher quality products to the market faster while lowering the cost of goods. Analytical tools and data modeling applications, delivered by several different companies, are the future for process understanding and robustness.

IDBS, the leading innovator in BPLM, is primed to help our customers with their development needs from start to finish.  

IDBS attends and organizes events regularly. Check out our events page!