It was the battle of Boston’s best pastry shops, competing for the top prize in a CancerNetwork^® Face-Off event. Experts in lymphoma from Dana-Farber Cancer Institute in Boston, Massachusetts, divided into 2 teams–Modern Pastry Shop and Mike’s Pastries–gathered to debate treatment options in the lymphoma setting. The teams competed in 3 key areas: real-world updates, use of autologous stem cell transplant (ASCT) or chimeric antigen receptor (CAR) T-cell therapy, and patient cases.

What is Face-Off? It is an educational program designed as a competition for teams to present to and against each other.

How does it work? There are 3 rounds: data presentations, topics, and patient cases. During each round, both teams can present and defend their ideas and challenge the other team. The judge determines who is worthy of the top prize.

Round 1: Data Presentation

Modern Pastry Shop on the TRANSFORM Study

Presented by Jennifer Crombie, MD

The phase 3 TRANSFORM trial (NCT03575351) assessed lisocabtagene maraleucel (liso-cel; Breyanzi) vs standard of care (SOC) as second-line treatment for patients with relapsed/refractory large B-cell lymphoma (LBCL).¹ A total of 89 patients were enrolled in the liso-cel arm vs 91 in the SOC arm, with 61 patients crossing over and 57 receiving liso-cel. The median follow-up was 33.9 months.

Between the liso-cel and SOC arms, 48% and 66% were men, the median age was 60 and 58 years, and the most common LBCL subtypes were diffuse LBCL (DLBCL) and not otherwise specified in 58% and 54%, respectively. The best response to the first line of therapy showed a complete response (CR) in 33% vs 30%, a partial response (PR) in 39% vs 50%, stable disease in 8% vs 5%, and progressive disease in 21% vs 14%.

Event-free survival (EFS) was 29.5 months (95% CI, 9.5-not reached [NR]) in the liso-cel arm vs 2.4 months (95% CI, 2.2-4.9) in the SOC arm (HR, 0.375; 95% CI, 0.259-0.542). At 36 months, the EFS rates were 46% (95% CI, 35.2%-56.5%) and 19% (95% CI, 11.0%-27.3%).

The progression-free survival (PFS) was NR (95% CI, 12.6-NR) and 6.2 months (95% CI, 4.3-8.6) between arms, respectively (HR, 0.422; 95% CI, 0.279-0.639). The 36-month PFS rates were 51% (95% CI, 39.9%-62.0%) and 26.5% (95% CI, 15.9%-37.1%).

In the liso-cel arm, the overall response rate was 87% (95% CI, 78.3%-93.1%) vs 49% (95% CI, 38.3%-59.6%) in the SOC arm. The duration of response was NR (95% CI, 16.9-NR) in the liso-cel arm and 9.1 months (95% CI, 5.1-NR) in the SOC arm (HR, 0.603; 95% CI, 0.364-1.000), with 24-month rates of 60.5% (95% CI, 49.7%-71.4%) and 43.5% (95% CI, 28.8%-58.1%). The overall survival (OS) was NR (95% CI, 42.8-NR) in the liso-cel arm vs NR (95% CI, 18.2-NR) in the SOC arm (HR, 0.757; 95% CI, 0.481-1.191), with 36-month rates of 63% (95% CI, 52.7%-72.9%) and 52% (95% CI, 41.2%-62.4%).

Safety was consistent with previously reported results in the treatment-emergent (TE) period. In the post-TE period, 16% of patients in the liso-cel arm and 14% in the SOC arm had any-grade adverse effects (AEs), with 9% and 5% experiencing grade 3/4.

Mike’s Pastries on Real-World Implementation of TRANSFORM

Jacobson / These are impressive outcomes. It shows an incredibly effective product in a clinical trial setting with very good efficacy for a CAR T-cell product. I will say that a CR rate of 75% and [the] PFS that we’re seeing here don’t seem to match up with what we’re seeing in the real world. It doesn’t even match up with other clinical trials of liso-cel. This does seem to be an anomalous population for efficacy. The safety probably seems similar, although in my practice, more than 50% of patients have CRS [cytokine release syndrome]. It’s much closer to 75% or 80% of patients having at least grade 1 CRS. Just like other randomized trials of CAR T-cell therapy vs SOC, [liso-cel] is clearly a winner. I don’t think the numbers here necessarily match up in terms of efficacy in the real world.

LaCasce / The other issue currently is the reliability of manufacturing. The longer time to manufacturing and more off-spec products when we use this in the real world have been mitigated by the change in the label now. There are fewer patients who are off spec, but it does take longer to manufacture these cells. That is one of the main things that we will probably argue is a consideration when you have a patient who has aggressive, explosive disease.

Ryan / In a sense, you self-select the patients who are going to liso-cel because you think that they can wait long enough for bridging. I would just agree in terms of the safety that CRS is probably higher than 50%. I have seen some inpatient AEs like neurotoxicity. The rates were very low and probably still low but perhaps just taking potentially sicker patients through, and maybe that is resulting in higher rates of some of the toxicity that were reported in the study.

Fisher / The other team was talking about the limitations of the trial, so I’ll focus on some of the pluses. It is certainly less toxic than some of the other CAR T products available. It is more efficacious than the standard of high-dose chemotherapy [with] ASCT. It also expanded in the real world the population of patients who can undergo CAR T because of the decreased toxicity. People who would not have gone on this trial, for example, are able to get through this therapy.

Merryman / Another strength of this trial is that it allowed bridging chemotherapy rather than just bridging steroids, which reflects the real-world experience for many patients. A key patient group missed by all the randomized trials is patients who have secondary CNS [central nervous system] involvement. We don’t have great data to guide therapy for those patients in the setting.

Crombie / For real-world data, we’ve been fortunate to contribute patients to the [Center for International Blood and Marrow Transplant Research] registry and have participated in a few real-world liso-cel studies. Last year, ASH [American Society of Hematology] showed outcomes of patients in the real-world setting treated with liso-cel, and they found that in that cohort, outcomes were pretty similar to what was seen in the trial. There are limitations to these types of studies, but more than half may not have even been eligible to participate in this study and still had similarly favorable outcomes.

Mike’s Pastries Presents ZUMA-7

Presented by Caron A. Jacobson, MD, MMSc

The phase 3 ZUMA-7 trial (NCT03391466) looked at axicabtagene ciloleucel (axi-cel; Yescarta) vs SOC chemoimmunotherapy in patients with relapsed/refractory LBCL.² The primary end point was EFS. Crossover was not allowed for this protocol.

The median PFS in the axi-cel arm was 14.7 months (95% CI, 5.4-43.5) vs 3.7 months (95% CI, 2.9-5.3) in the SOC arm (HR, 0.51; 95% CI, 0.38-0.67). The 1-year PFS estimates were 52% vs 28%, 2-year were 46% vs 27%, 3-year were 44% vs 26%, and 4-year were 42% vs 24%.

The median EFS was 10.8 months (95% CI, 5.0-25.5) in the axi-cel arm and 2.3 months (95% CI, 1.7-3.1) in the SOC arm (HR, 0.42; 95% CI, 0.33-0.55). The EFS rates at 1 year were 49% and 20%, at 2 years 44% and 19%, at 3 years 41% and 19%, and at 4 years 29% and 17%.

The median OS was NR (95% CI, 28.6-not estimable [NE]) vs 31.1 months (95% CI, 17.1-NE) in the SOC arm (HR, 0.73; 95% CI, 0.54-0.98; P = .03). The OS rates at 1 year were 76% and 63%, at 2 years 60% and 51%, at 3 years 56% and 48%, and at 4 years 55% and 46%.

In the axi-cel arm, the most common grade 3 or higher AEs included neutropenia (69%), anemia (30%), leukopenia (29%), and hypophosphatemia (18%). In the SOC arm, the most common grade 3 or higher AEs were thrombocytopenia (57%), neutropenia (41%), febrile neutropenia (27%), and leukopenia (22%). Serious AEs of grade 3 or higher in the axi-cel arm included encephalopathy (12%), hypotension (11%), and pyrexia (9%). In the SOC arm, the only grade 3 or higher serious AE was delirium (1%).

Modern Pastry Shop Compares Efficacy and Safety

Crombie / We saw the trials were similar. Some of the key differences were the ability to use bridging therapy, which, as Reid mentioned, mirrors what is often done in clinical practice. The TRANSFORM study allowed for crossover. We saw an OS advantage with axi-cel and showed the advantage of using CAR T-cell therapy earlier. Those are some of the key differences.

Fisher / It’s another example where we put on the boxing gloves, [where] CAR T is better than a stem cell transplant. I don’t know if I’ve seen a breakdown of patients going to [ASCT] who had a PR vs CR in their outcomes. Because now, if we were considering ASCT, let’s say someone who’s more than 12 months out, if they had a PR, we’d probably say they [had a] failed [therapy] and go to CAR T-cell therapy. Those with a PR wouldn’t go to ASCT now. I don’t know how that would influence the numbers. I’d be curious as to that. It doesn’t answer the question as to which CAR T product is better, though.

Merryman / Unrelated, maybe tangential to this conversation, but it still is a little unclear: If you give someone bridging [therapy] and they happen to be in CR, are they better with an ASCT or CAR T-cell therapy? We don’t have data that directly answer that question. Those patients can still do well with ASCT, but clearly, as a strategy, second-line CAR T-cell therapy is winning.

Jacobson / To piggyback on Reid’s point, a lot of times we have to make decisions about a population of patients and not about individual patients in front of us. It looks like, based on the survival curves and also the outcome of CAR in the third line for these patients, that if you wait until that point to see if they’ve been in a CR, you’ve missed the boat for being able to cure. That’s a very small population of the whole pie. If you don’t take them to CAR and wait to see how they respond, fewer patients are going to survive their lymphoma. Those patients may do quite well, but if you want to treat the population as a whole and maximize the number of patients who survive, you have to take the CAR approach.

LaCasce / The other thing I’d add is that the high-dose chemotherapy is very myelotoxic. Something that we are seeing now, as we’re following these patients for a long period of time, is that people start to develop clonal hematopoiesis or risk for myelodysplastic syndrome [MDS] and acute myeloid leukemia [AML]. If you’ve had ASCT and then you require CAR T-cell therapy, that risk goes up substantially. Because these patients tend to remain in long-term remission from their disease, this becomes a big issue.

Davids / As a clinician, it’s nice doing less ASCT because those patients struggle. Christine, maybe I can put this one to you because this point was brought up earlier around efficacy outcomes and manufacturing reliability. Is that influencing how you think about the optimal sequencing of CAR T in these patients?

Ryan / The manufacturing reliability did play a lot into the potential selection of product, but more recently, [it is] probably affecting that decision less. In terms of sequencing, though, that’s more of a question about which CAR T-cell therapy product. I wouldn’t say that necessarily affects the decision about whether I’m seeing a patient deciding about going to CAR or not.

Round 2: Hot Topics

Mike’s Pastries on Using ASCT Up Front

Presented by Ann S. LaCasce, MD, MMSc

This has been our standard of care for many years. People who have transformed follicular lymphoma and who had prior therapy, not just de novo transformed disease where we wouldn’t consider doing consolidation, [may be eligible for ASCT]. We generally, if they are in CR, take them to ASCT. I certainly have many patients like this who have done very well with durable remissions for both their aggressive lymphoma and their follicular lymphoma. These are patients who are chemotherapy sensitive, so you could argue that these are patients who benefit from a high-dose chemotherapy approach. Yes, there are toxicities associated with high-dose chemotherapy, but many of these patients do well for many years and may not need subsequent therapy. When they relapse, then we would think about CAR T-cell therapy. For patients who do not achieve a CR with standard chemotherapy for their transformed disease, we would take those patients to CAR T-cell therapy. It’s about demonstrating that these patients [are] chemotherapy sensitive, and if they are, there are plenty of data to suggest that these patients in CR who go to ASCT do very well. We can save the CAR T-cell therapy for later if they need it.

Modern Pastry Shop on Pro CAR T-Cell Therapy at Relapse

Fisher / The argument that we’ve had with that is the long-term risk of cytopenias and mild dysplasia, eventual leukemia in patients who are exposed to multiple calculators from pre–high salvage therapy, high-dose chemotherapy, and even fludarabine-based therapy with CAR T-cell therapy. That’s what you have to weigh out.

Davids / Reid, who’s the right patient for this strategy of pro CAR T relapse?

Merryman / The level of evidence for consolidative ASCT in this population is not high. It’s retrospective studies that showed an EFS benefit, not an OS benefit, and with all the caveats of retrospective data. The other argument that you could make is that there are 2 large data sets now, one from the European Consortium and one from the US, that patients with transformed follicular lymphoma probably do better with CAR T-cell therapy than patients with de novo DLBCL. These patients are going to do better with CAR T-cell therapy later, and the evidence for using ASCT consolidation is weak. I would lean toward waiting and using CAR T-cell therapy later.

Crombie / We’ve seen in these other studies that earlier CAR T-cell therapy or doing CAR T-cell therapy after ASCT doesn’t necessarily result in improved outcomes. Having a less heavily pretreated patient go to CAR T-cell therapy could be an advantage.

Jacobson / We’re doing what Matt warned us not to do earlier, which was not to compare different trial populations and outcome populations against each other. Until we have a randomized study to answer this question, we should be doing the things that maximize the opportunities for a cure for our patients. If they’re chemotherapy sensitive and you take them to [ASCT], a proportion of those patients will be cured, and the ones who aren’t can go on to get CAR T-cell therapy. I would say without a head-to-head study, it would be hard to argue the other way right now.

Crombie / Becoming the standard was done many years ago, and we didn’t have CAR T. We have even more options now for DLBCL. Even if a patient relapses after CAR, there are other new strategies available. To commit someone to an ASCT so early may put them through additional toxicity that is not warranted.

Davids / What about the sequencing argument in terms of someone who goes to ASCT and then later relapses? Can they get CAR T-cell therapy? If you’re going the route of CAR T-cell therapy at relapse, would you later take them to an ASCT? Probably not. You might lose a line of therapy.

Fisher / After ASCT, after CAR T-cell therapy is unlikely to be effective; you can certainly consider allogeneic stem cell transplant after CAR T-cell therapy. It gets to be a more difficult problem.

Crombie / We do have newer therapies, and patients may be in worse shape after going through ASCT. Logistically, it also requires a patient to be in the hospital for many weeks. A lot of patients would probably prefer not to do that if they had a reasonable backup option down the line, if needed.

Merryman / If you look at the EFS benefit from the retrospective studies, it’s somewhere between 10% and 15%. I quote a 5% to 10% risk of MDS and AML from an ASCT. If there is a benefit, it seems like a small one.

Davids / Does [treatment selection] differ with, let’s say, a 40-year-old vs a 70-year-old [patient]?

LaCasce / For sure.

Jacobson / Yes, for the 40-year-old [patient], you want to keep open all the possibilities to get them to cure. At 70, you weigh the risks and benefits a little bit more and hope that CAR T cells will.

Merryman / I don’t think any of us would take a 70-year-old [patient] to ASCT in this scenario.

Davids / Where’s the cutoff then? What age?

Fisher / Sixty-four.

Round 3: Patient Cases

Modern Pastry Shop on Relapsed DLBCL With CNS Involvement

Presented by Jennifer Crombie, MD

A 53-year-old woman with stage IV non–germinal center B-cell DLBCL involving the kidney and bone marrow achieved complete remission with polatuzumab vedotin (Polivy), rituximab, cyclophosphamide, doxorubicin, and prednisone but experienced relapse at 6 months with both systemic and CNS disease.

Should this patient receive salvage chemotherapy (rituximab plus high-dose cytarabine, dexamethasone, and cisplatin alternating with high-dose methotrexate) followed by thiotepa-based ASCT if achieving CR or proceed directly to CAR T-cell therapy with bridging?

Crombie / This is a challenging clinical situation. It’s rare that we see this, but when we do, it’s always difficult to manage. Patients have systemic disease that’s chemotherapy refractory with an early relapse, and you’re dealing with the management of the CNS involvement. We do have data that show patients who are chemotherapy sensitive and can achieve a thiotepa-based ASCT can have favorable long-term outcomes. Based on available data and the long-term responses that we’ve seen in those patients, that’s a worthy goal. The problem is that sometimes it’s hard to get a patient like this to get into remission in both the systemic and the CNS. I would try to give chemotherapy and a thiotepa-based ASCT. If the disease were progressing, then I would go to CAR T-cell therapy because you don’t have another option.

Jacobson / I agree, but I would add that I think what we’ve learned about CAR T-cell therapy for CNS disease is that it can cause dramatic responses in a subset of patients, and for some patients, that can be a couple years. Unlike in systemic DLBCL, we see relapses 3 and 4 years after. I am skeptical that CAR T-cell therapy in the setting of active disease, maybe as a consolidation therapy, can cure the disease, but in the setting of active disease, [it] can cure CNS lymphoma. We know that thiotepa ASCT can cure a subset of patients. That compels me to give this a try. We all know this patient who has essentially early relapsing systemic disease; their CNS disease may respond, but their systemic disease may be a problem. They may have to go to CAR T-cell therapy just because you can’t get both sets of diseases into remission. It’s worth a shot because the 5-year data for ASCT look better than the 5-year data for CAR T-cell therapy. I’m glad we have CAR for patients who are not chemotherapy sensitive, but I don’t think it replaces it for patients with active disease.

Fisher / A lot of [patients] who get CAR T-cell therapy in this setting, though, are not chemotherapy sensitive and not felt to be ASCT candidates. I’m not sure if we’re apples to apples there.

LaCasce / The CNS disease is what I worry about; that’s what is rapidly fatal. CAR T-cell therapy can be tough sometimes with CNS disease. You have to look at the individual patient; they’ve left them with meningeal disease. We have to involve our neuro-oncology colleagues because the toxicity for these patients can be very high. If they do respond and have a complete remission, I agree: I want to get them to a thiotepa ASCT. It’s a very tough situation.

Mike’s Pastries on Young Patient With High-Risk Progressive Disease

Presented by Caron A. Jacobson, MD, MMSc

A 32-year-old patient with primary refractory DLBCL demonstrated progression during the regimen of rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone after cycle 4, with rapidly enlarging lymphadenopathy. The patient is otherwise healthy (ECOG performance status of 1) with high tumor burden and markedly elevated lactate dehydrogenase (LDH) levels.

Given the aggressive clinical course and patient eligibility for cellular therapy and considering the urgency, toxicity risks, and efficacy profiles, what factors would guide your choice between available CAR T-cell therapy products?

Jacobson / This is exactly the patient who would’ve been favored for axi-cel in our propensity-weighted analysis. These are the factors we thought of. We thought of pretreatment LDH, we thought of a pretreatment tumor burden, we thought of high pretreatment inflammatory markers, all things that predict poor outcome with cell therapies, and axi-cel outperformed liso-cel in this patient population. [I’m] going back to our earlier discussion about not wanting to risk an out-of-spec product or potentially a product that wasn’t manufactured at all and wanting to get CAR T-cell therapy into this patient rapidly, because to be honest, this patient’s probably not going to respond very well to bridging therapy. All those things make me want to pick the product that’s most likely to get to the patient in the fastest time frame, and that would be axi-cel.

Crombie / You can make an argument based on the data that the toxicity profile is better. At our institution, we have a lot of comfort, as was said before, with axi-cel because we’ve been using it for a long time and it’s very reliable. You can make the argument based on the data that are available. The safety profile is better, and liso-cel could be used for everyone. We all do the same thing in our group.

Ryan / At a young age, at this point, we have a lot of options for bridging still. You have polatuzumab [vedotin], you have bispecifics, you have liso-cel. Yes, [the patient has] primary refractory [disease], but if you wanted to make the argument for liso-cel, you probably could get this patient to that with the tools that we have available.

References

1. Kamdar M, Solomon SR, Arnason J, et al. Lisocabtagene maraleucel versus standard of care for second-line relapsed/refractory large B-cell lymphoma: three-year follow-up from the randomized, phase III TRANSFORM study. J Clin Oncol. Published online July 7, 2025. doi:10.1200/JCO-25-00399

2. Westin JR, Oluwole OO, Kersten MJ, et al; ZUMA-7 Investigators; Kite Members. Survival with axicabtagene ciloleucel in large B-cell lymphoma. N Engl J Med. 2023;389(2):148-157. doi:10.1056/NEJMoa2301665