About 30% of CAR T-cell therapies manufactured each year go unused because the intended recipient's condition deteriorates during the 10-to-21-day production window, according to analysis published in 2024.

That waste rate frames a rapid-manufacturing proposal from researchers at Portugal's Instituto de Biologia Experimental e Tecnológica (iBET), who argue that cutting ex vivo cell expansion to three days or less would preserve more therapeutically potent populations and enable lower per-patient doses.

"Rapid manufacturing changes the manufacturing goal as, instead of maximizing the number of cells produced outside the body, it aims to make fewer but fitter cells, by limiting expansion to three days or less," Marta Costa, PhD, told GEN. "This preserves naïve and stem cell-like memory T cells, which proliferate and persist much better once infused, and this can, ultimately, result in lower CAR T cell doses."

What does a three-day window actually change?

Current protocols extend expansion until cultures reach roughly 10⁸ cells per batch, a density target that typically leaves harvested populations fully differentiated and exhausted. Compressing the timeline to roughly one-seventh of the upper production range elevates the share of naïve and stem cell-like memory T cells in the final product.

Costa frames the trade-off directly: "The evidence increasingly shows that cell quality matters more than cell quantity. Shorter processes produce fitter cells, cost less, increase capacity, and help patients who cannot wait weeks."

That combination—lower per-dose cell counts, shorter facility occupation, and faster delivery—could reshape dose-finding budgets in early-phase trials. Costa and colleagues predict equivalent efficacy at substantially lower dose levels, though the supporting data set remains limited to early-phase and preclinical work.

Why does automation become the gating factor?

Reproducibility pressures rise sharply when expansion compresses from two weeks to roughly three days. Costa identifies closed automated systems as the enabling layer. "Automation is what makes speed safe and allows for reproducibility," she said. "It also makes decentralized and point-of-care manufacturing realistic, allowing CAR T cell manufacturing to run at many hospitals supported by a closed, standardized and sensor-equipped system that is essential to ensure product quality."

She points to the CliniMACS Prodigy and Lonza's Cocoon as platforms integrating leukapheresis through final formulation in a single closed configuration. Both sit under evaluation for rapid protocols, she noted.

A separate bottleneck remains. Several release assays for sterility, viability, and potency still require a week or more to complete—longer than the proposed expansion itself. Faster analytics, not faster cell growth, may become the rate-limiting step for vein-to-vein time.

Could AI replace fixed manufacturing recipes?

Costa argues the current paradigm applies a single process recipe regardless of how an individual patient's apheresis material responds. Predictive modeling of growth kinetics, paired with continuous online monitoring of cell state, could allow operators to adjust parameters dynamically and select harvest timepoints based on cellular phenotype rather than calendar duration, she said.

Such adaptive manufacturing would require validated digital twins or surrogate markers—neither of which has reached regulatory acceptance for autologous cell therapy. Costa's framing remains forward-looking: these insights could enable more adaptive manufacturing strategies, including selecting the optimal harvest timepoint based on the evolving state of the cells rather than a predefined culture duration.

What does this mean for R&D budgets?

The wider context is structural. In vivo CAR T strategies, which bypass ex vivo manufacturing altogether through vector delivery inside the patient, are gaining momentum. Costa ties that trend to the rapid-manufacturing case and stops short of prediction: "We will see what the future holds."

For R&D managers weighing capital allocation, the calculation now spans three competing paths: shorten ex vivo expansion with existing automation, invest in AI-driven process-control overlays, or pivot toward in vivo platforms that remove the manufacturing facility from the timeline entirely.

What remains unverified

The iBET proposal rests on a small evidence base; randomized comparative trials of three-day versus standard expansion have not yet reported clinical readouts. The 30% annual wastage figure derives from one 2024 analysis cited by Costa, and methodology details—disease indication, patient cohort, and definition of "unused"—were not disclosed in available materials. Cost-per-dose savings and capacity-utilization gains remain projected rather than measured. Vendor benchmarks for the CliniMACS Prodigy and Cocoon under rapid protocols sit at an early stage.

Whether three-day expansion translates into measurable clinical benefit at scale is the question that principal investigators and manufacturing directors will need answered before retooling facilities around closed, sensor-equipped platforms and AI-driven release decisions.