Current Challenges in CAR T Manufacturing

CAR T therapy has undergone a remarkable transformation over the past decade, evolving from an experimental treatment into a clinically validated therapeutic platform with seven FDA-approved therapies and a rapidly expanding global market estimated at $6 billion.^1,2

Yet despite its clinical success, manufacturing remains a fundamental bottleneck to broader patient access.

The challenge facing CAR T is no longer clinical validation; it is scalable delivery. Manufacturing remains constrained by donor-to-donor variability, patient-specific production, and complex multi-step workflows operating within tightly controlled GMP environments.^3,4 These factors drive up cost, increase process variability, and limit manufacturing throughput, ultimately restricting patient access to potentially life-saving therapies.

In addition, existing analytical approaches were either never designed to work in real time or have limited capacity to predict manufacturing outcomes. Most inline sensors or probes rely on downstream or metabolite-based measurements that describe what a cell has done, rather than what a cell can become. This limits process visibility and reduces opportunities for informed intervention before failure occurs.⁵

As the field moves toward larger patient populations and commercial-scale manufacturing, the industry faces a critical challenge: how do we move from reactive manufacturing to predictive control? Solving this challenge is essential to improve manufacturing robustness, reduce the cost of goods, increase manufacturing capacity, and expand patient access to advanced cell therapies.

Achieving these goals will require more than automation alone. It will need technologies capable of revealing cellular states earlier in the manufacturing process, enabling informed intervention, and a transition from retrospective testing to proactive process control.⁵

To address this challenge, Cytomos has developed AuraCyt^®, a predictive cell intelligence platform designed to reveal cellular states earlier in the manufacturing process and support more informed manufacturing decisions.

From dielectric signatures to predictive cell intelligence

Cells carry rich biophysical information that is not always visible through conventional biological markers. Dielectric spectroscopy accesses this information by measuring the frequency-dependent electrical behavior that arises from a cell’s intrinsic structure.^6-9

Specifically, the cell membrane separates the conductive interior of the cell from its surrounding environment, shaping how the cell responds to an applied electric field. As the frequency increases, the influence of the cell membrane decreases, allowing the electrical response to become progressively more sensitive to intracellular structure and composition as well as membrane properties.^6,8,9 The resulting electrical response reflects the intrinsic physical state of the cell.

AuraCyt captures this response at the single-cell level using ultra-wideband dielectric spectroscopy. By measuring each cell across more than 200 frequencies, the platform generates a high-dimensional digital fingerprint of every cell that reflects the cellular state and captures subtle changes associated with emerging behavior. In the current configuration, this produces 507 parameters per cell, generating biologically grounded, AI-ready digital datasets for cell characterization and predictive analysis.¹⁰

This approach represents a paradigm shift in cell analytics: using physics to predict biology. As changes in cellular state often precede measurable biological outcomes, AuraCyt signatures can reveal early signals and emerging cellular trajectories before they are detected by conventional assays. This provides a biophysics-first route to predictive cell intelligence.^5,10

Applying predictive cell intelligence to CAR T manufacturing

To demonstrate how predictive cell intelligence can be applied in a real manufacturing workflow, AuraCyt was evaluated in a CAR T manufacturing case study focused on transduction efficiency and process risk.

Transduction efficiency is a critical determinant of the CAR T manufacturing success, yet current analytical methods typically confirm transduction only after key manufacturing decisions have been made. Cytomos decided to investigate whether predictive cell intelligence could provide earlier insights into manufacturing outcomes.

AuraCyt was evaluated as a rapid, label-free platform for monitoring transduction efficiency and characterizing transduced and non-transduced T-cell populations over an eight-day manufacturing workflow.

Donor-derived T cells were activated with IL-2 and expanded using a standard CAR T manufacturing process. Following activation, the cell population was divided into two arms: one transduced with a CD34-CAR construct and a matched non-transduced control.

AuraCyt analysis was performed on Days 0, 4, 6, and 8. Flow cytometry was performed on Day 6 to confirm CAR-CD34 expression, providing a conventional benchmark for transduction assessment against which earlier AuraCyt measurements could be compared. Cell viability and concentration were monitored throughout the process (Figure 1).

AuraCyt identified distinct dielectric signatures associated with donor T cells (Day 0), T-cell activation, and CAR-CD34 transduction. Flow cytometry confirmed CAR-CD34 expression on Day 6, providing a conventional benchmark for transduction and CAR expression. This aligned with AuraCyt analysis, which showed clear separation between non-transduced, activated, and CAR-CD34⁺ T cell populations using label-free single-cell measurements (Figure 2).

Across the eight-day workflow, the AuraCyt Signatures of CAR T cells and non-transduced controls followed distinct trajectories (Figure 3).

Using Day 0 donor T cells as a reference population, AuraCyt tracked the evolution of both transduced and non-transduced cells over time. Comparison of the resulting dielectric signatures using effect size analysis demonstrated divergence between the two populations by Day 4, prior to conventional confirmation of CAR expression by flow cytometry on Day 6 (Figure 3).

These findings suggest that AuraCyt can detect transduction-associated changes in cellular states and identify emerging CAR T manufacturing trajectories before conventional transduction assessment. This supports its potential as a rapid, label-free analytical approach for improving process visibility and enabling earlier manufacturing insight.¹⁰

Implications for CAR T manufacturing and beyond

These findings demonstrate how early AuraCyt signatures can reveal CAR T manufacturing trajectories before conventional assessment. By providing earlier insight into cellular states, AuraCyt has the potential to move manufacturing from retrospective quality control towards predictive process control.

Earlier visibility of process trajectory could improve manufacturing consistency, reduce variability, and support more informed decision-making during critical manufacturing windows. In turn, this may help lower the cost and complexity of delivering advanced cell therapies at scale.

Beyond CAR T, the same biophysics-first approach could be applied across cell therapy, biologics, and other advanced manufacturing workflows, providing a foundation for predictive biomanufacturing and AI-enabled process optimization.

 

Lindsay Fraser, PhD, is CSO at Cytomos.

 

References

1. U.S. Food and Drug Administration (FDA). Approved Cellular and Gene Therapy Products. Center for Biologics Evaluation and Research, FDA.

2. MarketsandMarkets. CAR T-Cell Therapy Market Report 2026–2031.

3. Ayala Ceja M, Khericha M, Harris CM, Puig-Saus C, Chen YY. CAR T Cell Manufacturing: Major Process Parameters and Next-Generation Strategies. Journal of Experimental Medicine. 2024.

4. Baguet C, Larghero J, Mebarki M. Early Predictive Factors of Failure in Autologous CAR T-Cell Manufacturing and/or Efficacy in Hematologic Malignancies. Blood Advances. 2024.

5. Mendoza R, Carter K, Peterson S. Early-stage Analytical Development Strategies for Cell Therapy. Cell & Gene Therapy Insights. 2024;10(11):1493–1503.

6. Gawad S, Schild L, Renaud P. Micromachined Impedance Spectroscopy Flow Cytometer for Cell Analysis and Particle Sizing. Lab on a Chip. 2001.

7. Cheung K, Gawad S, Renaud P. Impedance Spectroscopy Flow Cytometry: On-Chip Label-Free Cell Differentiation. Cytometry Part A. 2005.

8. Liang W, Zhao Y, Liu L, et al. Determination of Cell Membrane Capacitance and Conductance via Optically Induced Electrokinetics. Biophysical Journal. 2017.

9. Tehrani FD, O’Toole MD, Collins DJ. Tutorial on Impedance and Dielectric Spectroscopy for Single-Cell Characterisation on Microfluidic Platforms: Theory, Practice, and Recent Advances. Lab on a Chip. 2025.

10. Fraser L, Wright D, Hunter B, Giakoumelou S, Bellot F, et al. Biology Empowering AI: Rethinking Prediction in Biomanufacturing. Poster presented at the International Society of Cell & Gene Therapy (ISCT) Annual Meeting, Dublin, Ireland; 6–9 May 2026. Supported by the attached ISCT poster.

The post Current Challenges in CAR T Manufacturing appeared first on GEN - Genetic Engineering and Biotechnology News.