The regulatory outcomes for cell and gene therapy submissions have drawn increasing attention from developers and investors alike, but a recent analysis discovered that 42% of decisions in this space now result in complete response letters (CRLs), a rate that points to a recurring pattern rather than isolated setbacks.

The natural instinct is to assume safety is the problem. In practice, it almost never is. By the time a therapy has moved through clinical development, its risk profile is generally well understood, and reviewers rarely encounter meaningful new safety signals at the submission stage.

While many of the recent FDA CRLs related to cell and gene therapies are tied to late cycle findings around clinical effectiveness, Chemistry Manufacturing and Controls (CMC) and regulatory issues are common.

One review found that across many Biologic License Application (BLA) CRLs, most of letters are rooted in how the therapy was manufactured, measured, and documented rather than in what it’s capable of doing clinically. Understanding why this keeps happening requires looking at a handful of common mistakes developers make during the development process, often without understanding how much downstream impact they will have.

The gap between early-stage measurement and final product characterization

The methods used to test and characterize a cell therapy product are expected to evolve as development progresses, with regulators generally accommodating for that. In the case of cell-based advanced therapies, determination of meaningful critical quality attributes and potency are often difficult determine early and establishing specifications a key challenge.

Significant complications arise when the final validated assay differs meaningfully from the method used to characterize products used to generate earlier clinical data. At that point, the FDA must consider whether the product being submitted is truly the same product that produced the clinical results, or whether the shift in methodology has introduced uncertainty about what was actually tested and treated.

If bridging data to connect old and new methods doesn’t exist, the developer is left making an argument after the fact that is far harder to establish than the prospective work that would have resolved the question before it arose. This pattern appears consistently in regulatory review documents, and the core issue is almost always timing. Internally, assay transitions tend to be treated as technical improvements rather than formal process changes, which means the documentation and comparative studies regulators will eventually ask for are rarely in place when the question arises.

To avoid regulatory red flags down the line, any meaningful change in how a product is analyzed should be governed the same way a manufacturing change would be, bridging data generated, while both methods are still active and a documented rationale built into the development record before the gap becomes a problem.

The narrow window for process optimization

One of the less widely appreciated features of the cell therapy regulatory environment is that most approvals in this space are built on Phase II data rather than the larger Phase III trials required for traditional small molecule drugs. This has been a meaningful accommodation by regulators, reflecting the severity of the conditions being treated and the genuine clinical need for these products.

But this carries a consequence for process development that programs often encounter too late. If a therapy will be approved on Phase II data, the manufacturing process needs to be substantially final before that study begins. Changes made after the study opens raise questions about whether the product patients receive throughout the course of the trial is the same product the company intends to commercialize. Resolving those questions requires additional data, time, and resources that most programs can’t readily absorb.

What happens more often than outright late-stage changes is that developers recognize the process isn't optimal, weigh that against the pressure of near-term milestones, and decide to proceed as is. The logic is understandable in the moment. The process works well enough, a clinical readout is expected, and stopping to optimize feels riskier than moving forward, but that decision carries more weight than it appears to.

Whatever manufacturing process a therapy goes through in its pivotal study is, in practice, the process it will carry to market. There is no window that opens for optimization after these studies close. The limitations that seemed manageable during development, such as manufacturing success rates in the case of an autologous product, tend to resurface as critical constraints that define commercial viability.

Treating process readiness as a prerequisite for advancing rather than a problem to revisit afterward is the more reliable approach, and for the unit operations that tend to receive less internal scrutiny, that often means bringing in outside expertise well before Phase II enrollment begins.

What clinical manufacturing data reveals about approved ranges

The third issue is one I have encountered repeatedly in practice, and it is perhaps the least intuitive of the three.

Developers set quality specifications early in development, and at that stage the tendency is to draw them wide. The process isn't fully defined yet, batch-to-batch variability is expected, and tightening constraints before the data supports it creates more problems than it solves.

What sometimes catches programs off guard is that FDA doesn't approve the range a developer set on paper, it approves the range the clinical manufacturing data actually reflects. If every batch produced throughout the clinical program fell between N and Q, that is the specification regulators will require at approval, even if the original range ran from A to Z. A commercial process designed around the broader range suddenly has to consistently hit a narrower target it was never built to meet, and the yield problems that follow are expensive and slow to resolve.

To avoid this, developers need to be intentional during the clinical program about using product from across the full manufacturing range, not just from the optimal center. If a wider specification is going to hold up at submission, the clinical data needs to show that patients were treated with product from those outer limits. It is counterintuitive, but building that evidence into the clinical program early is the only way to preserve the manufacturing flexibility the commercial stage will require.

A pattern worth addressing earlier than most do

The three issues described here are not the result of technical failures or unpredictable regulatory behavior. They reflect a consistent gap between where developers are focused during the pressure of clinical execution and where the consequences of manufacturing decisions eventually land. Each is addressable before pivotal enrollment begins. Almost none are addressable afterward without the kind of delay, cost, and regulatory back-and-forth that now characterizes a significant share of submissions in this space.

Cell and gene therapy has produced genuinely transformative clinical results. Delivering on that science consistently, and at scale, depends on the manufacturing discipline required to carry it the rest of the way.