Hong Kong Team Makes CAR-T Therapy for Children With Leukaemia in Just 14 Days

Researchers in Hong Kong have shown that a fully automated, locally based manufacturing system can produce CAR-T cell therapy for children with relapsed B-cell acute lymphoblastic leukaemia in just 14 days, with every one of six treated patients achieving complete remission.

The therapy in question targets CD19, a protein found on the surface of B-lineage cancer cells. Chimeric antigen receptor T-cell therapy involves collecting a patient’s own T lymphocytes, genetically engineering them with a lentiviral vector so they recognise CD19, expanding them in culture, and infusing them back into the patient. The first approved product of this kind, tisagenlecleucel, was authorised by the United States Food and Drug Administration in July 2017 for children and young adults with relapsed or refractory B-cell ALL, based on a multi-centre phase II study reporting an 83% complete remission rate and an overall survival rate of 89% at six months in a heavily pretreated population.

Despite those results, commercial CAR-T remains out of reach for many centres and families. The approved product Kymriah is currently priced at USD 475,000, and the vein-to-vein time between apheresis at a local treatment centre and infusion of cryopreserved cells shipped from overseas manufacturing facilities typically takes four to eight weeks. Some patients die of disease progression while waiting. The Hong Kong team set out to test whether a locally manufactured, fresh, non-cryopreserved product could shorten that timeline dramatically while maintaining quality and safety.

The manufacturing process relied on the CliniMACS Prodigy system, a closed, fully automated, single-use platform that handles T-cell selection, activation, lentiviral transduction, expansion, and harvest in a Grade C clean room environment compliant with Pharmaceutical Inspection Co-operation Scheme good manufacturing practice guidelines. Apheresis products were transported to the GMP centre at 15 to 28 degrees Celsius with a target yield of 2 billion CD3-positive cells. The CAR construct combined the FMC63 single-chain variable fragment targeting CD19 with a CD8 hinge, a transmembrane region derived from TNFRSF19, a 4-1BB co-stimulatory domain, and a CD3-zeta activation domain, all of human origin except the murine-derived scFv.

Quality control was rigorous. Release testing covered Gram stain, sterility, mycoplasma, endotoxin, replication-competent lentivirus, vector copy number, cell viability, CD3 and CAR expression, viable CAR-T cell concentration, and residual B cells. Potency was assessed with a standardised in vitro cytotoxicity assay requiring at least 30% specific lysis of CD19-expressing target cells at a 5:1 effector-to-target ratio. Chain of identity and chain of custody procedures included dual verification of patient identifiers, barcode-labelled containers, and electronic tracking from collection to infusion. No replication-competent lentivirus was detected in any batch.

Six patients were enrolled, three boys and three girls, with a median age at infusion of 9 years (range 4 to 14). Five were treated at first relapse and one at second relapse; two had central nervous system involvement at relapse. Before infusion, patients received lymphodepleting chemotherapy with fludarabine and cyclophosphamide, and the target cell dose was 0.5 to 3.0 million CAR-positive viable T cells per kilogram. The median cell dose actually infused was 2.64 million per kilogram, with a median transduction efficiency of 42.86% (range 33.89% to 62.1%) and median CD3-positive cell viability of 96.22%. Production efficiency was 100%, with no manufacturing failures, and the vein-to-vein time was uniformly 14 days for all six patients.

The median filled volume per bag was 28.96 mL, with a median of 76 million viable CAR-T cells (range 47 to 174 million) and a median concentration of 2.71 million viable CAR-T cells per mL. No aggregates or endotoxins were detected in any product, and tests for bacterial growth and mycoplasma by quantitative PCR were all negative. The authors noted, however, that transduction efficiency and viable cell numbers varied considerably between batches, and the factors driving this variability remain uncertain and warrant further investigation.

On the clinical side, all six patients achieved complete remission with negative minimal residual disease one month after infusion, and all remained in continuous remission at a median follow-up of 9 months (range 2 to 22 months). The median progression-free survival was not reached. Quantitative PCR and flow cytometry showed robust early expansion of CAR-T cells, with peaks of up to 120,000 copies per microgram of DNA at days 7 to 14, persistence up to 12 months in several patients, and sustained B-cell aplasia in the majority of cases. All patients received intravenous immunoglobulin replacement every four weeks after infusion, and no patients in the cohort had clinically significant or laboratory-proven infections.

Toxicity was manageable but not absent. Five patients experienced grade 1 cytokine release syndrome, while one patient with a very high disease burden of 93% bone marrow blasts developed grade 3 CRS along with grade 4 immune effector cell-associated haemophagocytic lymphohistiocytosis-like syndrome, requiring intensive care admission and treatment with dexamethasone, anakinra, emapalumab, and ruxolitinib. CRS occurred at a median of 1.5 days after infusion and resolved by a median of 6.5 days with a median of three doses of tocilizumab. All patients recovered and were discharged at a median of 14.5 days after infusion (range 10 to 87 days). All patients experienced grade 4 neutropenia around day 7, with gradual recovery over two to three months, and one patient required prolonged granulocyte colony-stimulating factor support.

The authors are careful to frame the findings as preliminary. Key limitations include the very small sample size of six patients, the single-centre design, the short follow-up period, the absence of a comparator group, and selection bias. They recommend extended follow-up of at least two years after infusion, publication of longitudinal data, and multicentre validation trials to confirm scalability, long-term persistence, and cost savings. Even so, the study demonstrates that a GMP-compliant point-of-care CAR-T platform can deliver fresh, quality-controlled, effective therapy to children with relapsed leukaemia in two weeks rather than two months, potentially offering a financially feasible alternative for centres and families priced out of commercial products worldwide.

Subject of Research: Locally manufactured point-of-care CD19 CAR-T cell therapy for paediatric relapsed/refractory B-cell acute lymphoblastic leukaemia

Article Title: Locally manufactured point-of-care CD19 chimeric antigen receptor T-cell therapy for paediatric relapsed/refractory B-cell acute lymphoblastic leukaemia: a single-centre feasibility experience

Article References: Chan, W. Y. K., Leung, A. W. K., Zhang, H., Cheuk, D. K. L., Lee, P. P. W., Jiang, G. Y. J., Wong, C. K., Chow, M. Z. Y., Chu, Y., Ng, M. H. L., Hanssens, L., Gao, L., Leung, K. T., & Li, C. K. (2026). Locally manufactured point-of-care CD19 chimeric antigen receptor T-cell therapy for paediatric relapsed/refractory B-cell acute lymphoblastic leukaemia: a single-centre feasibility experience. Frontiers in Oncology, 16, Article 1900851. https://doi.org/10.3389/fonc.2026.1900851

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1900851

Keywords: CAR-T cell therapy, CD19, B-cell acute lymphoblastic leukaemia, point-of-care manufacturing, CliniMACS Prodigy, paediatric oncology, Hong Kong, cytokine release syndrome, lentiviral vector, GMP manufacturing, relapsed leukaemia, immunotherapy