Manufacturing Is the Medicine: Why CMC and GMP Decide Whether a Cell or Gene Therapy Reaches Patients

For cell and gene therapies, manufacturing is not a preliminary step toward the medicine. It substantially is the medicine. Chemistry, manufacturing, and controls (CMC) and good manufacturing practice (GMP) are the factors that decide whether a living product ever reaches a patient, because for a re-engineered T cell or a viral vector, the process defines the product. Get the process wrong and there is no molecule to fall back on. Much of a program's fate is settled on the factory floor, not at the bench.

Why the process is the product

A small-molecule drug can be characterized down to the last atom, and two batches made by different routes can be shown to be the same substance. A living cell therapy cannot be reduced to a single structure. An autologous CAR-T product starts from one patient's own cells, is transduced with a vector, expanded, and returned, all within a narrow window. The final product is defined by how it was made, which starting material was used, which reagents, which incubation times, which release assays. Change the process and, in a real sense, you may have changed the drug. This is why regulators treat comparability, the demonstration that a product is essentially unchanged after a manufacturing change, as a central scientific question rather than paperwork.

That biology creates a distinctive bottleneck. Conventional biologics frameworks assume large batches, long shelf lives, and the ability to quarantine and retest. Cell therapies often have batches of one, shelf lives measured in hours to days, and limited material for repeat testing. The US Food and Drug Administration's Center for Biologics Evaluation and Research (CBER) has long recognized this, and its cellular and gene therapy guidances describe manufacturing controls, potency, and comparability as the recurring hard problems (FDA, Cellular and Gene Therapy Guidances).

What the recent FDA CMC item actually is

In January 2026 the FDA posted material describing a flexible approach to CMC oversight for cell and gene therapies, and CBER issued a related final guidance on CMC flexibilities for products heading toward a Biologics License Application. It is important to read this correctly. Independent regulatory analyses noted that the announcement largely reflects existing policy and practice and aligns with the agency's past guidance. In other words, this is a clarification and consolidation of how CBER has already been reasoning, not a new rule that lowers the bar.

The flexibilities are specific and phase-appropriate. Before a product is made for later-phase trials, sponsors are not expected to meet every element of the commercial drug-manufacturing regulations in 21 CFR part 211 in full; expectations scale with development stage. For minor manufacturing changes, CBER has signaled it will not demand overly stringent comparability data. On process validation, the agency has indicated there is no fixed requirement for three process-performance-qualification lots; the number should be scientifically justified. On commercial specifications, there is room to refine release limits as post-approval experience accumulates.

None of this removes the core obligation. Phase-appropriate GMP still means GMP. Flexibility on the number of validation lots is not flexibility on sterility, identity, or potency. Reading a clarification of existing expectations as permission to cut corners would be a serious misinterpretation of what the agency said.

The bottleneck in numbers

The manufacturing bottleneck is not theoretical. In real-world use, a meaningful share of patients approved for autologous CAR-T never receive an in-specification product on the first attempt. A survey from the European Society for Blood and Marrow Transplantation documented that manufacturing failures and out-of-specification products occur with commercial CD19 products in routine practice, and that failure rates differ meaningfully from one product to another. A report from the UK National CAR T Panel examined manufacturing failure in patients approved for treatment of large B-cell lymphoma; in that cohort a small but clinically important minority of patients experienced a failed manufacturing run, and the report traced the consequences for those who could not be treated in time. When the process fails, the patient, often already progressing, may run out of options while a remanufacture is attempted.

Economics compound the problem. Published cost-of-goods analyses for autologous CAR-T have put per-dose manufacturing costs in the six figures, with skilled labor and materials such as viral vectors as leading drivers. A 2025 analysis in Frontiers in Medicine describing a quality management system for decentralized manufacturing highlighted long contract-manufacturing lead times and constrained capacity as structural limits on how many patients can be served (Frontiers in Medicine, 2025). Capacity gates access as surely as biology does.

What this means for developers and patients

Three implications follow. First, invest in CMC early. Locking a robust, well-characterized process before pivotal trials avoids the comparability trap, where a late process change forces expensive bridging studies or a repeat trial. Second, treat potency and release assays as scientific instruments, not checkboxes; a therapy you cannot reliably measure is one you cannot reliably release. Third, read regulatory flexibility as an invitation to bring good science and clear justification, not as a loosening of quality. The direction of travel, toward automation, closed systems, and decentralized or point-of-care models, is promising, but each of these still has to earn its GMP credentials.

For patients, the honest message is that a therapy can be scientifically sound and still fail to arrive because it could not be made in time, to specification, at a cost the system can bear. Solving manufacturing is not secondary to solving the biology. It is the same problem, seen from the factory.

This article is educational and is not medical advice.