Key Takeaways From The 2024 Bioprocessing Summit
By Tyler Menichiello, contributing editor
Last week, I spent three long, dizzying days at the 16th annual Bioprocessing Summit in Boston, surrounded by the brightest minds and biggest innovators in the industry. The notebook I travelled with is still hot from my pen’s furious scribbling, and it’s fuller than I know what to do with, but I’ve tried my best to capture the key takeaways — at least, MY key takeaways — from this year’s summit. Any session could very well demand its own editorial coverage, but for the sake of this review, I’m just hitting the high points. Expect more coverage to come in the following weeks.
Data, AI, And Automation
Data collection in biomanufacturing has grown significantly since the field’s inception, and it isn’t slowing down anytime soon. As Amgen’s SVP of process development, Jerry Murry, Ph.D., said at Monday night’s keynote, “Data is the new oil, the new gold — however you want to say it.” Our tools for data collection are constantly improving, and I would go so far as to say we have more data being collected than we know what to do with. This presents unprecedented challenges — not only in terms of IT infrastructure, but also in analytics. How do we, the industry, effectively capture and analyze this vast and complex biomanufacturing data? Well, AI of course.
Unsurprisingly, automation, artificial intelligence (AI), and machine learning (ML) underscored most discussions and presentations this year. Murry mentioned Amgen’s robotic filling systems, as well as its automated, machine-learned visual inspection line. You read that correctly — 95% of all of Amgen’s units (filled vials and syringes) are released through this automated visual inspection system. He went on to say that Amgen’s aspiration is to have predictive, in-silico models for every unit operation the company runs, as well as digital twins for everything from products to bioreactors. Murry also thinks AI use in regulatory filing isn’t too far off, either. He says that as long as data can be organized and saved appropriately, AI will be able to one day help write your BLA — or, at least, a draft BLA.
An interesting point to mention is the prevalent issue of unstructured data in biomanufacturing. It can include everything from the data in old, hand-written lab notebooks to that of scanned documents, PDFs, videos, and more. ML can help structure this data, but we’re not quite at the point where you can simply snap your fingers and “structure” it. That will take time, technological advancements, and a lot of model training.
Another key session regarding automation was led by Claire State from Bristol Myers Squibb. The session was titled, “Fully Automated Cell Therapy Manufacturing,” and discussed the promises of automation in cell therapy manufacturing — namely, a reduction to manufacturing COGs and scalability. While no 100% fully automated systems currently exist, State explained that a stepwise approach to adoption is critical, and that it’s early to prioritize these process changes early in an asset’s lifecycle to make implementation easier.
Join me and my expert panelists at 11 a.m. ET on Wednesday, Sept. 4, as we explore the current use and applications of AI/ML in biomanufacturing in the upcoming Bioprocess Online Live event, “A Realistic Look At AI/ML Use In Biomanufacturing.” I’ll be leading the conversation with rBIO’s CSO, Deenadayalan Bakthavatsalam, Ph.D., Bernard T. Ferrari Professor at the Johns Hopkins Carey Business School and co-chair of the Johns Hopkins Workgroup on AI and Healthcare, Tinglong Dai, Ph.D., and Tugce Martagan, Ph.D., Associate Professor of Industrial Engineering and Innovation Sciences at Eindhoven University of Technology.
Sustainability In Biopharma
There was a great series of talks about sustainability at this year’s BPS, culminating with a panel discussion featuring Lijuan Li, Ph.D., from Takeda, Matt McNulty, Ph.D., from Tufts University, and David Roush, Ph.D., from Roush Biopharmaceutical Panacea. A standout quote from Roush, is “Perfect shouldn’t be the enemy of today.”
Biomanufacturing, and indeed, medicine as a whole, is far from a green industry, but that doesn’t mean we shouldn’t try. The panelists agree: committing to sustainability is a lot like committing to lowering COGs. You shouldn’t try to optimize a million things at once — think about the five most impactful or important aspects of your process that can be improved and start there. Sustainability, at its core, is about the mindful consumption of resources. Along that vein, Li said, the industry needs to find a way to financially incentivize sustainability efforts. Sustainability should be simple, and it needs to start from the ground up. “Don’t expect senior leadership or the government to figure it out for you,” said Roush.
Improving Analytical Methods
Two great presentations stick in my mind regarding analytical methods — one from Novartis and one from Eli Lilly. Miha Vodnik, Ph.D., a senior expert in science and technology at Novartis, gave a talk discussing the differences in analytical development and goals between new biologics and biosimilars. There are more stringent analytical requirements for biosimilars because the focus is on demonstrating comparability — i.e., the target ranges are derived from originator products, and are much more specific than the generic, “good-enough” specifications that new biologics have to meet. To establish quality ranges for biosimilar development, Vodnik said, it’s important to compare the biosimilar to originator batches sourced from different markets, different lots, and different expiration dates, and land somewhere in the middle.
Claudia Gributs, Ph.D., senior director of R&D at Eli Lilly, gave a talk titled, “Evolution of Core Analytical Methods During The Development Lifecycle for mAb Products.” In her presentation, she discussed the reasons a company may change analytical methods for a given product and the subsequent “method-bridging” studies that may need to happen as a result. The reasons she listed to justify changing analytical methods were unsatisfactory performance (of the old/current method), technological advancements (i.e., new and improved methods and technologies becoming available), operational reasons (to reduce cost, complexity, and/or turnaround time), and material availability.
Gributs discussed the considerations when it comes to the need for a bridging study, the goal of which is to create a link between the old method’s data and the new. The key, she said, is to understand how the methods relate to one another. Whether or not a bridging study needs to be done depends on the risk posed by changing the method, as well as how much is known about the new method (i.e., how well-validated is it?). The bridging protocol is a good place to store this risk assessment, she said. She also referenced ICH Q14 as a valuable source in guiding your analytical development.
Continuous Manufacturing
From what I saw, Amgen is leading the industry in terms of efficiency and innovation on the continuous manufacturing (CM) front. Not only was the shift toward continuous manufacturing a significant part of Murry’s keynote address, but Amgen senior process development scientist Madiha Khurshid gave an in-depth presentation the following day, titled “Continuous Manufacturing Enables Productivity And Reliability.” CM not only offers greater flexibility (e.g., a single-use process can fit into a multi-use facility) and productivity, but the reduced need for space reduces a facility’s carbon footprint, making it more sustainable.
A critical part of continuous manufacturing is upstream and downstream automation, according to Khurshid. “I cannot imagine an upstream CM process without the automation of the biomass control,” she said. She went on to describe Amgen’s hybrid CM process, which is broken into two distinct stages. The first half is an automated, continuous bioreactor process, with continuous harvest and connected purification. The second half consists of discreet polishing and filtration steps. This hybrid approach reduces the CM complexity, because it doesn’t require the coordination of all unit operations together as one huge process. Keeping CM contained to the first half reduces the need for automation tools, thus increasing the speed of development. To quote Khurshid, “The first portion is where the most value added is for continuous, so that’s where we put it.”
POC Cell Therapy Manufacturing
Point-of-care (POC) CAR-T manufacturing was the focus of several presentations in the cell therapy manufacturing track. Caring Cross co-founder and executive director, Boro Dropulić, Ph.D., gave a keynote presentation on Wednesday that examined the ways a POC manufacturing model can improve the affordability and accessibility challenges of CAR-T cell therapy. Looking at some of the statistics he highlighted, it’s clear that the traditional, centralized model of manufacturing not only presents prohibitive costs of treatment, but significantly limits access. Currently, only 25% of eligible patients receive CAR-T in the US, and only 21% of these patients are non-white. Globally, less than 1% of patients in low-to-medium-income countries are receiving CAR-T.
Not only does centralized manufacturing present geographical and temporal challenges (delaying patient treatment), but the logistics of this model compound into unsustainably high prices. I won’t call any companies out, but the lowest CAR-T price shared in his presentation was $373,000, and the highest was $3.1 million.
POC manufacturing isn’t just a theory — it’s being done right now at some academic medical centers. Dropulić mentioned Spain as an example. Nirav Shah, MD, physician and professor at the Medical College of Wisconsin (MCW), shared success stories from his own work done at MCW, where autologous CAR-T products are made on-site for clinical research patients. He used a dinner analogy to explain the differences in cost between traditional and POC manufacturing — with centralized CAR-T manufacturing being akin to flying to New York for a dinner at a Michelin-star restaurant, and POC being more like buying all the ingredients to make it at home yourself. In the end, the meal is comparable, and significantly cheaper when you cut out the travel and service costs. He also referenced a study which showed that patients receiving fresher (i.e., cells that weren’t frozen and thawed at any point) CAR-T products had slightly better outcomes.
Perhaps the most important message from Shah’s presentation was the idea that traditional CAR-T manufacturing and POC manufacturing don’t have to compete — they should complement each other. As Dropulić said, “COVID taught us that a single supply chain is a weak supply chain.” Decentralizing manufacturing with sponsor oversight can really only improve costs and access of therapies to patients.
Tune in Tuesday, Aug. 27, at 11 a.m. ET as I discuss the advances in CAR-T manufacturing and debate the feasibility of wide-spread POC manufacturing in the Cell & Gene Live event, “Modernizing CAR-T Manufacturing: Is Point Of Care The Future?” I’ll be joined by expert panelists, Jason Bock, Ph.D., founder and CEO of CTMC, Emily English, Ph.D., SVP and head of manufacturing operations at Cartesian Therapeutics, and Tal Salz, Ph.D., former FDA CMC expert-turned consultant at Dark Horse Consulting Group Inc.