FDA Takes Close Look at Innovation

Industry 4.0. Artificial intelligence. Big data. Even continuous manufacturing. All of these new technologies will drive the future of pharmaceutical and biopharmaceutical manufacturing. Yet questions persist as to how the U.S. FDA and other global regulatory agencies will address these new technologies, leaving some companies reluctant to fully embrace these advances as early adopters.

Some answers, while not necessarily definitive, could be found at the 2019 PDA/FDA Joint Regulatory Conference in Washington, D.C., Sept. 16–18. From the opening to the closing talks, both regulatory and industry representatives expressed the need to integrate more advanced manufacturing technologies into everyday processes and the critical role of collaboration between industry and regulators.

In fact, some of the FDA speakers pointed to innovation as a way to improve quality. In her presentation, “Advancing Drug Innovation through Investment in Quality,” as part of the opening plenary, Patrizia Cavazzoni, MD, Deputy Director for Operations, CDER, specifically cited how innovation is necessary to create a “more modern manufacturing infrastructure, including advanced technology for manufacturing of drugs and biological products.” This, in turn, will help reduce the threat of drug shortages due to fewer drug recalls.

“Innovation should help establish the quality system,” she said. “FDA continues to be a strong advocate for, and to support, modernization of the pharmaceutical industry, and we recognize that adopting innovative approaches to manufacturing may present technical and regulatory challenges. For instance, companies may have a concern that using such technologies would result in delays while FDA reviewers familiarize themselves with the technologies.”

Cavazzoni pointed to CDER’s Emerging Technology Program as one way FDA is addressing this concern. This group seeks to “promote the adoption of innovative approaches to pharmaceutical product design and manufacturing.” Manufacturers looking to implement such technology can communicate their concerns directly to the FDA members comprising the Emerging Technologies Team (ETT). This early interaction is supposed to help address any regulatory concerns early on before the technology is used.

Augmented Reality in the Cleanroom

The ETT came up again later that day in the breakout session, “Augmented Reality and Artificial Intelligence: Conceptualization through Implementation.” Bob Bowden, PhD, Senior Director, Cell Collection, Advanced Therapeutics Supply Chain, Janssen, discussed his company’s approach to using augmented intelligence during manufacturing operations. Currently, in the beginning stages of implementation, his company plans to use what he refers to as “augmented intelligence”—a device or system that enhances human capabilities ”to augment the intelligence of our operators in a cGMP environment.”

In particular, the operators within the company’s cell therapy manufacturing facility wanted an approach that would let them view the batch record and batch record status without directly handling a paper batch record. In addition, the operators wanted to interact with only one interface instead of six to seven systems to complete a process.

Bowden’s team took these needs into account, developing an approach relying on an augmented reality headset. Similar to virtual reality, augmented reality overlays computer-generated graphics on physical space. Pokemon Go is an example of such a technology. In fact, augmented reality is already being used in other manufacturing sectors, such as agricultural equipment and aviation. The company is now conducting use cases involving an augmented reality headset outside of the cleanroom as the company wants to be absolutely certain of the capabilities of the technology before bringing it into the cleanroom.

But what about the regulatory side? According to Bowden, his company’s main regulatory concern about the headset device involves 21 CFR Part 11, the FDA regulations applying to electronic signatures. He believes that since the data passes through the device and this existing data is not altered means it does not fall under Part 11.

At this time, the device is being used on the production floor as some technical limitations still need to be overcome.

During the Q&A portion of the session, David Jaworski, Senior Policy Advisor, CDER, FDA, joined the panel. The first question was directed at him, asking for the Agency’s perspective on artificial intelligence and machine learning.

“We are evaluating a lot of the new technologies, usually on a oneon-one case right now,” he replied. “We do have the ETT process over in CDER and we have a similar process in CDRH, so there is the ability to come in and discuss these concepts in advance…in the development stage. We would encourage people to open up those communication channels early. Right now, it is more exploratory.”

Later, another question referenced Cavazzoni’s point about the potential for manufacturing innovation to reduce drug shortages.

“I think some companies tend to underestimate the impact of looking at some of this technology and innovation upon the quality of their products,” Jaworski said. “And I really do think you can make a business case for improving the quality that then is going to reduce the shortages…the technology may be technically very complex, but it is the design of the process, the design of the software that you are going to use, the design of the equipment that you are going to bring in to your facility, that really needs to be looked at at the beginning. Try to really concentrate on these things in advance instead of waiting for a problem to occur downstream.”

In fact, Jaworski explained that sometimes he has seen companies using older, existing technologies that result in quality issues when a more advanced piece of technology could have prevented the problems from occurring.

“All of these things if you look at them, there is always something that could have been done early in the design process to really reduce that potential, and it may be technically challenging on the upstream side in your workflow but on the downstream side you will have much better results. I think when that all plays out, it can help eliminate drug shortages… if you concentrate on the quality aspects, then the ability to provide medication to patients in the end will be satisfied.”

Continuous Manufacturing Keeps Up

The next day, Sharmista Chatterjee, Supervisory Chemist, CDER, FDA, provided more details on the ETT. The team is headed by Sau (Larry) Lee from the Office of Pharmaceutical Quality and consists of 22 members including herself. The team is cross-functional within CDER, featuring representation from the Office of Pharmaceutical Quality, Office of Compliance and Office of Regulatory Affairs. The ETT provided support for FDA’s approval of the first new drug product produced via continuous manufacturing and the first switch from batch to continuous manufacturing for a previously approved product (1).

She views continuous manufacturing as another way for manufacturers to reduce quality issues.

“A quality product of any kind consistently meets the expectations of the user,” Chatterjee explained. “Drugs are no different.”

FDA has seen applications for continuous manufacturing for both small and large batches. For small batches, this includes both drug substance and drug product. Chatterjee has also seen continuous aseptic spray drying. For large molecules, applications have included continuous manufacturing for downstream processes, end-toend integrated bioprocesses and a small manufacturing platform for continuous bioprocesses (i.e., pharmacy on demand).

Since launching the ETT program in late 2014, Chatterjee explained that continuous manufacturing requests sent to the ETT have included multiple face-to-face meetings or teleconferences between manufacturers and the Team. Many had site visits with ETT members prior to submitting the application.

Approvals for continuous manufacturing have included Vertex’s ORKAMBI, Janssen’s Pezista, Eli Lilly’s Verzenio, Vertex’s Symdeko and Pfizer’s Daurismo.

In February 2019, FDA released a draft guidance on continuous manufacturing. T he primary objectives of this guidance are to communicate that FDA supports continuous manufacturing, that it is consistent with the FDA Quality Initiative and there are no regulatory hurdles for implementation of continuous manufacturing. FDA received over 20 comments. T he comments included recommendations for further clarity on what has to be included in the regulatory submission concerning a site’s pharmaceutical quality system, concerns about “intermingling” of cGMP expectations with design expectations, further information with respect to requirements for process analytical technology data storage and the riskbased approach for model maintenance over the product lifecycle and requests to only include topics unique to continuous manufacturing and avoid topics raised in other guidances.

FDA plans to move forward with the guidance while taking the comments into account. Chatterjee said that the content of the draft guidance served as the basis for the Agency’s perspective when drafting ICH Q13: Continuous Manufacturing for Drug Substances and Drug Products.

FDA Center Leaders Talk Tech

During the “Center Updates” on the last day of the conference, advanced manufacturing technologies came up as a topic of discussion. In fact, Peter Marks, MD, PhD, Director, CBER, FDA, explained how his Center is even embracing some of the same technologies internally.

“A.I. is something that is increasingly being used in various spaces. We are using it at the Center in a variety of ways in terms of exploring what it can do for us. For instance, sifting through thousands of adverse events and helping us understand where there is a real signal,” he said. “If A.I. can help flag a high percent of them and basically triage 95% of them, and categorize them, that would actually free up staff and allow people to concentrate on other things.”

In addition, he said that some companies are now submitting applications to CBER containing big datasets, such as whole genome sequencing. This is an area where A.I. can really shine and Marks stated that a high performance computing group at CBER is working on it.

“It [A.I.] is clearly the wave of the future,” Marks said. “I do not think we are going to see a computer doing reviews of the applications but certainly it can help us tremendously with regulatory review, particularly in the post-market setting. It may turn out that A.I. is very useful in helping us figure out whether there are safety signals, particularly, as we build very large databases.”

Alonza Cruse, Director, Office of Pharmaceutical Quality Operations, Office of Regulatory Affairs, also weighed in on A.I. and other technologies.

“[It requires] a whole new skillset and understanding but I think it is one [technology] that is important because drug manufacturing companies are using A.I…they are using this to look for, and to rehab, issues.”

Cruse envisions FDA inspectors walking into a facility and using A.I. to decide which areas to focus on. He sees A.I. as just one way to help with decision-making.

“A.I. can help us selectively reach more informed decisions.”

Reference

1. “Impact Story: Regulatory Science is Strengthening U.S. Drug Product Manufacturing.” Regulatory Science in Action. FDA.gov (Feb. 5, 2019) tinyurl.com/yxopbqxf