Some of the most damaging commercial problems in advanced therapies appear late in development. Their origins, however, can often be traced back years earlier.
That was the warning at the heart of this discussion: decisions that seem perfectly reasonable during research or early clinical development can become extremely difficult — and expensive — to reverse once a therapy progresses.
Margot Pont pointed to early cell therapy programmes as an example. Many pioneering processes were rooted in academic research, where the primary aim was to answer scientific questions and demonstrate clinical potential.
But an academically interesting process is not automatically a commercially optimal one.
Once patients are being treated and a product progresses through clinical development, changing manufacturing becomes increasingly difficult. What initially appeared to be a small process choice can therefore become locked into the product.
The implication is clear: developers need to think much earlier about the type of manufacturing and quality-control system that a future product will require.
Núria Gavaldà sees a similar issue when academic projects transition towards GMP manufacturing.
Researchers may assume that a process working successfully in a laboratory environment can simply be transferred into GMP. In reality, that transition often exposes gaps in process understanding.
Technology transfer and process development therefore deserve far more attention than they sometimes receive.
Developers need to understand critical parameters, quality attributes and their manufacturing strategy before entering clinical development. At the same time, they must avoid designing processes that are so rigid they cannot accommodate future evolution.
Finding that balance is difficult because funding often creates pressure to move quickly.
For academic developers and smaller organisations in particular, reaching a clinical milestone can be essential to unlocking the next source of financing. The temptation is therefore to move into the clinic as quickly as possible and address manufacturing optimisation later.
But moving fast is not always the same as moving efficiently.
Jacob Smith argued that the industry now has substantially more manufacturing knowledge than it did 15 or 20 years ago. Many failure points are no longer mysteries. Experienced manufacturing teams increasingly know where processes are likely to struggle and which elements can be standardised.
Yet the pressure to pursue biology and accelerate towards clinical proof of concept remains powerful.
This creates a risk that organisations generate positive safety or efficacy data around a process that is poorly suited to later scale.
Esteve Trias described the underlying requirement simply: developers need a product that is sufficiently mature.
That does not mean innovation stops once development begins. Advanced therapies evolve in a fast-moving scientific environment, and new opportunities inevitably emerge. But continuously modifying an immature product can become extremely difficult within existing regulatory frameworks.
The regulatory landscape itself can change during development too.
That makes early engagement with regulators particularly important. Regulatory strategy cannot be an activity left until the end of development. It needs to form part of the development process from the beginning.
The lesson from this discussion is not that developers should spend years attempting to create a perfect process before treating their first patient.
It is that speed needs to be balanced with foresight.
Early development should create enough product and process understanding to know where flexibility exists, where risks lie and which decisions could become extremely difficult to change later.
In advanced therapies, the costliest decisions may not be the obviously bad ones.
They may be the decisions that work perfectly well in Phase I — until the company discovers they do not work nearly as well in Phase III.
)