Researchers from the Indian Institute of Technology Delhi, led by Anurag S. Rathore, PhD, are using novel real-time soft sensors based on spectroscopy, machine learning, and other tools to look at critical quality attributes during upstream and downstream continuous processing.
“People have done lots of work on the enabling technologies [for continuous processing],” explains Garima Thakur, a PhD student at the Institute’s Center of Excellence for Biopharmaceutical Technology (CBT), who will be presenting her work at BioProcess International in September. “But one of the challenges that remains is what happens when something goes wrong.”
In particular, she explains, there has been limited research into strategies to handle process deviations, including expected process variabilities such as changing titers, and unexpected errors or breakdowns.
To investigate this, she says the team has incorporated real-time process analytical technology tools, including spectroscopic sensors in-line in the process streams. For example, she explains that spectroscopic sensors have been integrated in chromatography columns to monitor the product capture and ensure that the resins don’t get over- or under-loaded. The team has also used spectroscopic probes to look at final product formulation, including the concentration of potentially harmful salts.
Integrated control system
The spectroscopic tools can take measurements in a few seconds by throwing light onto the flowing process streams and acquiring multivariate spectral information, according to Thakur, who adds that her group has developed a cutting-edge integrated control system that can flag deviations and deploy various control strategies based on the real-time inputs from the spectroscopic and process sensors.
“We’re able to detect during processing a lot of things that usually need offline analysis, such as charge variants, excipients and product concentrations,” says Thakur. “Traditionally, during a batch process, in a lab, you need to tick off boxes before the next operation, but here it’s all in real time.”
“Our vision is to demonstrate a real-time control system that can make sure that each unit of biotherapeutic produced in a continuous manufacturing facility meets the high product quality expectations of a biotherapeutic product and is safe to inject into the patient,” notes Rathore. “This would truly bring the promise of continuous processing, including lower capital costs and enhanced productivity, into the manufacturing plant, and ultimately allow more people to access the otherwise expensive biotherapeutics.”
