In May 2026 the haematology department of the Second Affiliated Hospital of Zhejiang University School of Medicine published an account of a 68-year-old patient with relapsed multiple myeloma who, after nearly a year of successive chemotherapy regimens, received a CD19/BCMA dual-target CAR-T infusion. Twenty-eight days later no abnormal plasma cells were detected in his bone marrow. By day 40 his symptoms had resolved and his blood counts, immunoglobulins and kidney function had improved. He was discharged for long-term follow-up.
It is a striking recovery, and the hospital's account is worth reading. It is also one patient, reported by the treating team, with no control group and no peer-reviewed dataset attached. This article uses the case to explain a genuine design problem in CAR-T, then separates what the case demonstrates from what it cannot. Readers considering treatment in China deserve both halves.
What multiple myeloma does
Bone marrow is the factory that produces red cells, white cells and platelets. Plasma cells are the specialised cells that make antibodies. In multiple myeloma a clone of plasma cells becomes malignant: it multiplies without control, crowds out normal blood formation, and releases large amounts of a useless monoclonal protein that can damage the kidneys.
The consequences follow directly. Too few red cells cause fatigue and dizziness. Crowded-out normal white cells leave the patient vulnerable to infection. Low platelets cause bleeding. The monoclonal protein can block and injure the kidneys. In the reported case, the patient presented with dizziness, blurred and double vision and difficulty walking, and was found to have severe anaemia, thrombocytopenia, badly impaired kidney function and more than half of his marrow occupied by malignant cells.
The diagnosis IgG-kappa multiple myeloma, Durie-Salmon stage III and ISS stage III, with plasma cell leukaemia, describes a very advanced, high-risk disease. Plasma cell leukaemia is a rare and aggressive leukaemic phase of myeloma. It is an important detail when interpreting the result: this was a patient with an extremely poor prognosis and few remaining options, not someone with a routine relapse.
Ten months of treatment that did not hold
The reported treatment history is a familiar sequence in relapsed myeloma. First-line VRD combined bortezomib, lenalidomide and dexamethasone. When resistance appeared, the team moved to DVD and KRD regimens containing newer targeted agents. Finally they used DR-PACE, a highly intensive salvage combination with substantial toxicity. Across roughly ten months the disease was briefly suppressed and then returned in full. This is what relapsed or refractory means in practice: not that nothing was tried, but that each line of therapy stopped working.
That history is the context in which a cell therapy is considered. It also explains why the case cannot be generalised. A patient who has exhausted four lines of intensive treatment has a different disease biology and a different risk tolerance from someone earlier in the course of myeloma.
The failure mode dual targeting is built for
A CAR-T cell is engineered to recognise one surface marker. In B-cell malignancies that marker is usually CD19; in myeloma it is usually BCMA. The design has a known weakness. A tumour is not a uniform mass of identical cells, and under the pressure of a targeted therapy, cells that no longer display the target can survive and repopulate the disease. This is called antigen escape or antigen loss. It is one of the main reasons a patient can respond and then relapse.
Dual-target CAR-T addresses that geometry rather than simply adding a second drug. The hospital's account describes the two targets as complementary: BCMA is described as highly specific to myeloma cells and the main attack surface, while CD19 is described as reaching a more primitive cell population with greater potential to drive resistance. Giving one cell two independent recognition systems means a tumour cell would need to lose both markers at once to hide. The reported aim is a carpet-sweep clearance that lowers the chance of escape and later relapse.
The short version
The problem: a single-target CAR-T can be evaded when tumour cells stop displaying the marker it recognises.
The design: engineer the cell to recognise two markers at once, here CD19 and BCMA, so escape requires losing both.
The case: one 68-year-old patient with relapsed myeloma and plasma cell leukaemia, with no abnormal plasma cells detected at day 28.
The limit: one patient, reported by the treating hospital, is not a response rate and not a comparison with standard care.
The biological claim about CD19 reaching a primitive resistant compartment is the team's stated rationale, and it remains debated in the myeloma literature. CD19 expression on myeloma cells is not universal. Dual targeting is a plausible and actively studied strategy, not a settled mechanism. We describe it as the hospital presents it so readers can evaluate the argument, not to endorse it as proven.
The four steps, and the fever that followed
The reported procedure follows the standard autologous CAR-T pathway. T cells are collected from the patient's blood by apheresis. In a GMP-grade laboratory a gene is introduced so the cells display the chimeric antigen receptor. The cells are expanded to a therapeutic dose and tested. Before infusion, lymphodepleting chemotherapy clears space and resources for the incoming cells. Finally the product is infused intravenously.
After the infusion the patient developed grade 1 cytokine release syndrome, with recurrent fever. The team managed it with close monitoring, corticosteroids and anti-infective support. Fever in this setting is a sign that the engineered cells are activating and expanding; it is expected and manageable, but it requires a unit that recognises escalation early and has the drugs and ICU backup to respond.
| Time point | Reported finding | What it means |
|---|---|---|
| Before infusion | Four successive regimens over about ten months, all followed by relapse; severe anaemia, thrombocytopenia and kidney injury | Relapsed or refractory, very high-risk disease |
| After infusion | Grade 1 cytokine release syndrome with recurrent fever | Expected immune activation, managed on the ward |
| Day 28 | No abnormal plasma cells detected in bone marrow | A deep response at the first formal assessment |
| Day 40 | Headache, double vision and leg weakness resolving; immunoglobulins and serum kappa light chain falling and stabilising; liver and kidney function normalised; blood counts improved | Clinical and laboratory recovery consistent with the marrow finding |
| Follow-up | Discharged for long-term monitoring | Durability unknown at the time of publication |
What one patient cannot tell you
The most important sentence in this article is that a case report is not a response rate. There is no denominator. We do not know how many other patients were treated in the same programme, what happened to them, whether the treatment was part of a registered trial, or how long this patient's remission lasted. We do not know the manufacturing success rate, the toxicity profile across a cohort, or how the result compares with other salvage options. Publishing a successful case is standard practice for a hospital; it is also a selection of one outcome from a larger, unseen set.
None of that makes the recovery unreal. It means the correct use of the information is to understand what dual targeting is trying to achieve and what questions to ask, not to estimate the odds of a similar result. Any percentage quoted to a patient should be traceable to a named trial with a stated sample size and follow-up, not to a case narrative.
"No abnormal plasma cells detected at day 28" describes one patient at one time point. It says nothing about how often this happens, or for how long.
There is also a distinction between an approved product and a centre's own cell-engineering programme. China has approved nine CAR-T products, including BCMA-directed therapies for myeloma. A dual-target construct made and used at a single hospital is a different regulatory category from a commercially licensed product. Patients should ask which applies, because the answer determines what evidence, what price and what protections are attached.
A note on this hospital
The Second Affiliated Hospital of Zhejiang University School of Medicine is a major academic centre in Hangzhou, and the case was reported by its own haematology team. We are describing a published account, not making a referral. This hospital is not listed among the partner centres on this site, and nothing here should be read as an endorsement, an availability guarantee or an invitation to seek treatment there.
The hospitals page explains the categories of centre we work with and how a shortlist is built.
What to ask about any dual-target proposal
- Is the construct a licensed product or a hospital-made cell therapy, and under what regulatory pathway is it given?
- Is the treatment part of a registered clinical trial, and what is the registration number?
- What published cohort data exist for this exact construct, with what sample size, follow-up and response rate?
- How were antigen expression and target positivity confirmed on my own tumour cells before treatment?
- What is the centre's published rate of severe cytokine release syndrome and neurological toxicity for this product?
- What does the written quotation cover, and what is the plan if manufacturing fails or the disease progresses during the waiting period?
- Has a specialist independent of the treating centre reviewed the proposal and the alternatives?
What this means if you are considering treatment in China
Dual-target CAR-T is one of the more credible answers to a real limitation of first-generation cell therapy, and Chinese groups are among those testing it. A dramatic single case is a reasonable starting point for curiosity. It is not a reasonable basis for a treatment decision, a budget or a flight booking.
We can help you establish which product is being proposed, whether it is licensed or investigational, what published evidence supports it, and what an itemised quotation includes. We do not recommend specific therapies or centres, and we do not predict outcomes.
Our analysis of why CAR-T design differs across cancers explains the wider design landscape.
This article is educational and is not medical advice. Only a treating specialist with the full medical record can assess whether any therapy is appropriate.