If you’ve worked in CGT over the last several years, you will be aware how the mood has swung. Headlines have focused on manufacturing stumbles, high costs, failed launches, and payer friction. Arguably, the market is going through Gartner’s so-called “trough of disillusionment.” The question is whether it can climb onto Gartner’s Slope of Enlightenment. I would argue that it can, most definitely, and that the companies that understand their manufacturing process will be the ones to ascend the slope fastest.
The opportunity remains strong
There is no doubt that the market remains attractive in principle, as there are multiple unmet clinical needs and the potential to deliver life-changing therapies, sometimes cures, across a range of diseases. There is ever-growing scientific evidence that CGTs work, with some modalities almost being considered well-established and material advances being made in others.
However, commercial challenges remain, largely due to cost and revenue aspects of the economic model. While some CGT companies are showing signs they can deliver the valuations that justify the risk capital invested, others are struggling to establish sustainable, viable businesses. What has changed is the definition of “viable.” Clinical promise alone is no longer enough. Supply, manufacturability, and data-driven control now lie at the centre of viability and of valuation.
Viability depends on manufacturability
For too long, a major drawback of a hyper-focused pursuit of clinical data has meant that commercial manufacturing and the resulting cost of goods (COGs) end up as an afterthought, leading to baked-in COGs that are practically impossible to improve later.
There is growing evidence, however, that this ‘clinical only’ focus is no longer acceptable. Acquirers and late-stage investors now screen for a credible route to low COGs at commercial scale. A therapy can be clinically successful and still fail on valuation if it cannot demonstrably be made reliably, affordably, and repeatedly.
As a result, even at the pre-clinical stage, companies must already consider how to close, automate, and scale processes, identify necessary changes, and assess technology gaps. This does not necessitate large capital investments up front, but rather a clear, deliberate plan for when and how to scale.
This shift is visible in how big pharma evaluate assets, no longer prepared to buy any therapy that is successful in clinical trials or gains approval. Instead, they choose assets that either fit their existing manufacturing capabilities or have an in-house route to low COG commercial scale.
In today’s environment, manufacturing isn’t a “nice to have.” It’s intrinsic to valuation, to deal probability, and ultimately to a successful exit.
Closing, automating, and understanding: the pillars of scalable CGT
Closing and automating the manufacturing process is essential, but these aren’t the only levers for driving down the fully loaded cost of goods. A major cost driver in CGT manufacturing stems from manufacturing failures, whilst the primary focus of much process development remains on meeting clinical milestones. To the extent that formal process characterisation is undertaken, it remains largely empirical.
While the delivery of clinical data must continue take priority, we think there is scope to apply a more explicit process engineering approach informed by system-level thinking to cell therapy manufacturing. Digital twins are already used in mAb production, where industry 4.0 approaches are enabling a move to continuous manufacture, bioreactor scale-up and more productive facilities.
Cell therapy manufacturing needs to follow suit.
For biologists it will come as no surprise that, unlike therapeutic effect, manufacturing is not defined by any single molecular mechanism; it is an integrated system in which cells respond to their environment, process conditions, and to each other in complex, nonlinear ways. Accordingly, there is an opportunity for biotechs to develop process models, nascent digital twins, designed to capture the scaling relationships in these systems. This builds on an approach we at TTP have long taken to bioreactor design.
While such models will be physics-led, building on process parameters such as mixing, dO2, shear, E-field, there is a possibility to add in known biochemistry, e.g. metabolite cycles during cell expansion, leading to the identification of key measurable variables, such as nutrient flow, signalling markers, or metabolic rates, that meaningfully affect process outcomes. This is in line with what Kim Branson of GSK has recently argued about the application of machine learning to biological systems, “We can still optimize and steer these systems (cellular control theory), even if their full logic remains beyond our grasp.1”
These variables could be incorporated into predictive models that have the potential to explain performance trends and inform real-time process decisions, enabling predictive, closed-loop manufacturing that improves reproducibility, reduces batch-to-batch variability, and ultimately supports the cost-effective commercialisation of CGTs.
Such an approach may:
- Help accelerate development through a reduction of the number of experiments required (in comparison to empirical process characterisation), leading to faster process development, and hence increased funding runway, and reduced COGs through operational improvements.
- Identify new critical process parameters, which can drive the development of new technologies to bring what are currently at/off-line technologies in-line to enable in-process control for improved product quality.
Understanding how conditions scale can also build understanding of the impact on changing critical process parameters as a process evolves with increasing doses during clinical trials, and identify how automation should be applied as a tool to improve product consistency.
Process understanding unlocks scale, transfer, and valuation
Companies that invest early in building well-characterised processes can move faster, reduce technology transfer risk, reassure acquirers of reproducibility at scale, and will ultimately command higher valuations.
In today’s environment, you need to demonstrate that your process is reliable, scalable, and economically viable. A deliberate strategy that couples understanding with appropriate automation provides both the flexibility to partner and the speed to scale.
TTP works with CGT companies to de-risk tech transfer by developing well-characterised, automated, and scalable processes — giving partners and acquirers confidence and helping you get to market faster.
1https://www.linkedin.com/pulse/we-may-never-understand-biology-machine-learning-either-kim-branson-8r3oc/
