Clinical Cases in Pharmacology Clinical Cases  ·  Hematology I  ·  Hematologic Neoplastic Disorders
Hematology I · Case-HemNeoplastic-0011 — Hematologic Neoplastic Disorders

DLBCL in Early Relapse: Second-Line Standard Versus a Chemosensitive Patient's Own Response

Two trials moved the second-line standard in early-relapse DLBCL to CAR-T, and this patient sits inside their enrolled population. She also sits outside the group whose outcomes generated the result.

Abbreviations, terms, and other agents mentioned in this case DLBCL — diffuse large B-cell lymphoma  ·  R-ICE — rituximab, ifosfamide, carboplatin, etoposide — a salvage chemotherapy regimen  ·  PET — positron emission tomography  ·  CAR-T — chimeric antigen receptor T-cell therapy  ·  ZUMA-7/TRANSFORM — the phase 3 trials establishing CAR-T over transplant in early-relapse DLBCL  ·  CR — complete response  ·  CHOP — cyclophosphamide, doxorubicin, vincristine, prednisone — the R-CHOP backbone  ·  CD19 — a B-cell surface antigen targeted by CAR-T therapy
Presentation

L.G., a 55-year-old woman who manages a regional produce distribution warehouse, relapsed with diffuse large B-cell lymphoma eight months after finishing frontline R-CHOP, discovered on a routine surveillance scan before any symptoms returned. She started salvage therapy with rituximab, ifosfamide, carboplatin, and etoposide, and after two cycles her repeat PET scan came back completely clean — no residual fluorodeoxyglucose uptake anywhere her original disease had been, the kind of response that, before CAR-T existed as an option, would have sent her straight to autologous transplant with real confidence in the outcome.

ZUMA-7 and TRANSFORM changed the second-line standard for exactly her situation — early relapse inside twelve months — establishing CAR-T as superior to salvage chemotherapy plus transplant across their full enrolled populations on an intent-to-treat basis. The difficulty is that her eight-month relapse put her in those trials' population while her PET-negative response took her out of the group that drove their result. In ZUMA-7's standard-of-care arm, roughly half the patients never responded to salvage at all and so never reached a transplant, which is a large part of why the comparison came out as lopsidedly as it did; in the supplementary analysis of the minority who did respond and were transplanted, the confidence interval around CAR-T's advantage was wide enough to be consistent with no effect. That is a post-randomization comparison inside one arm, not a powered subgroup test, and it cannot overturn the headline finding. It does mean the headline finding was largely generated by patients whose disease behaved in exactly the way hers did not.

L.G. · 55 Second-line, post-salvage
History
DLBCL, relapsed 8 months after frontline R-CHOP (early relapse)
Salvage response
PET-negative complete response after 2 cycles R-ICE
Performance status
ECOG 0
Autologous transplant eligibility
Confirmed eligible, no organ dysfunction
CAR-T access
Apheresis available; manufacturing turnaround estimated 4-5 weeks

Lymphoma tumor board, second-line planning

Hematologist-Oncologist Opening

ZUMA-7 and TRANSFORM both showed CAR-T superior to salvage chemotherapy plus transplant across their full enrolled populations, on an intent-to-treat basis. Her clean response to salvage doesn't change which arm of those trials produced the better outcome overall — CAR-T is the current second-line standard for early relapse, and I'd offer it as the primary recommendation.

Transplant Physician Response

The overall result is real, and I want to be precise about what sits under it rather than overstate it. In ZUMA-7 about half the standard-of-care arm never responded to salvage and therefore never reached transplant — they registered an event immediately. That structure is a legitimate part of the trial's design and it is also why the intent-to-treat comparison looks the way it does. In the supplementary analysis restricted to the responders who were actually transplanted, the interval around CAR-T's advantage included no effect at all. TRANSFORM, as far as I'm aware, has no equivalent published analysis, so I'd stop short of saying both trials show this.

And a post-randomization comparison inside one arm is weaker still than a prespecified subgroup — it is exactly the kind of analysis that gets over-read. I'm not overturning the trials' main finding with it. I'm saying the main finding was built substantially on patients whose salvage failed, and hers didn't.

Clinical Pharmacologist Final

There's a practical dimension underneath this that isn't about which therapy is superior on paper. CAR-T manufacturing and insurance authorization here are estimated at four to five weeks — a real window during which a narrow, confirmed chemosensitive remission could begin to slip. Autologous transplant after an already-documented CR can proceed on a materially faster timeline. That's not a tiebreaker for efficacy, but it's a real cost to naming CAR-T the default without weighing it.

Regimen selected
Autologous Stem Cell Transplant
Consolidative Transplant · Proceeding promptly
Selected given her confirmed PET-negative complete response to salvage — the exact population transplant has historically performed well in — and the faster available timeline relative to CAR-T manufacturing.
CAR-T Cell Therapy — Held in Reserve
CD19-Directed CAR-T
Named explicitly as the next step if she relapses again post-transplant, at which point second-line CAR-T's own trial data would no longer be the relevant comparison — later-line CAR-T is separately well-established.
Where this was left

Agreed: proceed to autologous transplant now, given her confirmed clean remission and the faster available timeline, rather than pursuing CAR-T as the default second-line choice.

What today's decision does not settle: whether a patient with a messier salvage response — measurable disease on restaging, a partial rather than complete remission — should be read the same way. The hematologist-oncologist's position that the trials' intent-to-treat result should govern regardless of subgroup response was never actually tested against that harder case here; it simply wasn't the case in front of the team today, and stays untested by this particular decision.

Educational content only — a composite teaching case, not a real patient encounter or a substitute for clinical guidance. About These Cases →