Leo I. Gordon, MD, FACP: Well, 1 thing is the product themselves, the costimulatory domains in the products. The axi-cel is a CD28 product, and the other 2 are 4-1BB costimulatory domains. And there are some biologic differences between the 2. It looks like the CD28 results in perhaps a little bit more rapid expansion of the CARs, whereas the 4-1BB costimulatory domain may be a little bit slower but maybe with more persistence; we don’t know that. There are also some differences in the lisocabtagene maraleucel based on data from your institution, David, that suggested in laboratory with an equal amount of 1:1 ratio, CD4 and CD8 cells might be preferable, at least in animal models. And so when we’re using the Juno product, the liso-cel, it’s a 1:1 ratio, and it’s a 4-1BB costimulatory domain.

David Maloney, MD, PhD: What do we know about these patient populations in these 2 trials?

Matthew Lunning, DO: Well, I think beyond the relapsed/refractory large cell population aspect of it, it’s really what the patients are allowed to do. For instance, on the ZUMA-1 trial, no bridging therapy beyond steroids was used versus how the other constructs are allowing bridging therapy. So really you have to ask yourself, “Can I bridge this person? Do I need bridging therapy if I’m using a commercial product or if it’s on the trial?” I think that that’s really an individual discussion. The majority of the people are getting bridging therapy. Even in the real-world experience, there appears to be more bridging therapy with the use of axi-cel as more data emerge with regard to debulking and potentially outcomes.

Leo I. Gordon, MD, FACP: I think what I would also say is that the question of bridging therapy brings up another point that we haven’t talked about yet that we have to recognize as we’re trying to interpret these clinical trials. And that is this type of therapy just by its nature—by the fact that there are limited slots on clinical trials and the fact that it takes anywhere from 2 to 3 weeks to manufacture the product, the fact that bridging therapy is not allowed in some—has an automatic selection bias that we have to be aware of. And I think there’s a selection bias for clinical trials in general to include more favorable patients, but there’s especially selection bias in these studies because it takes a while, and these people are sick. And those patients who could eventually make it through to get the CAR might be a better group of patients. So we have to be aware that our data might be skewed to the more favorable. And I think as we begin looking at the real-world experience, and there are a number of abstracts at this meeting that address that, we may see a slight drop-off in our results, and we have to be aware of that.

Michael Pulsipher, MD: Let me throw out a question, though. You make it sound like some of these products are better than other products. Based on these 3 studies, is that really true?

Leo I. Gordon, MD, FACP: I think we can’t say that yet. In fact, even if we have final data in all 3, cross-trial comparisons are very difficult to do. The only way we’ll know the answer to that, and I’m not sure this will be done, is to have randomized studies.

Michael Pulsipher, MD: Exactly. A key point to emphasize is the patients aren’t exactly the same. They’re different. In each trial, the way the product was handled, the intervals between infusion, there are a lot of factors that really weigh in to that. So we know that they respond. We have some evidence of long-term response, but we really don’t know comparatively what’s best.

Matthew Lunning, DO: I think you could even drill down to the lymphodepletion chemotherapy, which was different for all 3 regimens. CY/FLU, or Cytoxan/fludarabine was an option, but in the JULIET trial, I think that you didn’t have to get necessarily any lymphodepleting chemotherapy, and bendamustine could potentially have been an option.

Michael Pulsipher, MD: And many studies have shown the right lymphodepleting chemotherapy makes a big, big difference. So that’s important to point out.

Matthew Lunning, DO: But it’s interesting, the differences in the doses, too. You’ve got to pay attention with the construct also.

Nilanjan Ghosh, MD, PhD: The other thing with ZUMA-1 is that they allowed only refractory patients or patients who had relapsed within 1 year post autotransplant. The other 2 trials used relapsed/refractory and also autotransplant. So that was a little bit of a different thing with ZUMA-1.

Matthew Lunning, DO: I think the liso-cel TRANSCEND trial was a very interesting trial to watch, as it navigated through the dosing cohorts to get to the pivotal cohort, which became more in line with the axi-cel cohort.

Leo I. Gordon, MD, FACP: One question comes up about lymphodepleting therapy, to clarify this a little bit. I think it’s clear to us, but I want to make sure that everyone understands that it’s designed as an immunologic treatment to allow expansion of the CARs. And I heard a question at the ASH [American Society of Hematology] meeting yesterday as to whether the lymphodepleting therapy might have caused the remission or might have been actually the reason the patient had remission. And I think we want to stress that these patients had had multiple treatments. They were refractory to this kind of treatment. And it’s very, very unlikely that any of the responses were related to the Cytoxan/fludarabine.

Matthew Lunning, DO: You can also argue that a lot of these patients haven’t seen fludarabine. We’re going to get to toxicity, but potentially some of the toxicities that we may see after CAR T cell may be related to some of the lymphodepleting chemotherapy.

David Maloney, MD, PhD: I think these are excellent points, and I think the 3 trials have all shown tremendous activity, with complete rates as high as above 50% and the durable responses appearing to be in that 30% to 50% range. And there are a lot of differences. But there are also differences in safety. What do we know about these 3 different products?

Leo I. Gordon, MD, FACP: Well, I think that the 2 issues we’re seeing in terms of safety are cytokine release syndrome and neurologic toxicity. And one of the problems in interpreting the data is not all the trials have used the same criteria to define those problems. I think in your institution, you’re beginning to harmonize that and to look at that. And I think we need to do that. It appears that the cytokine release syndrome seems a little bit lower with the 4-1BB product, but that might be a consequence of the way we’re measuring cytokine release syndrome. Neurologic toxicities are in the range of about 30% to 35%, with most of the neurologic toxicities being grade 1 or 2, not grade 3 or 4. I would add that even grades 1 and 2 neurologic toxicity for a patient are significant. Maybe for a clinical trialist they’re not, but for patients’ loss of word finding, confusion is a significant event.

Matthew Lunning, DO: It’s an event for their caregiver too.

Michael Pulsipher, MD: They certainly need to be aware that this is going to happen, and the big question is whether these are long-term toxicities or they contribute to long-term toxicities. And so far, we haven’t seen that in low-grade neurotoxicities, but we need to watch them.

Leo I. Gordon, MD, FACP: Correct.

David Maloney, MD, PhD: I think the lesson is that so far, it looks like most people can recover from these toxicities, but they can be severe, and patients have a fairly high chance of ending up in the ICU [intensive care unit] with almost any of the products if they have severe cytokine release syndrome or neurologic toxicity. So I think this will affect the ability to give these products in the outpatient setting. We certainly have experience with the 4-1BB constructs, using them as an outpatient setting, whereas the CD28, axicabtagene, pretty much requires hospitalization for at least 7 days. And I think the median was around 16 days in the clinical trials.

Transcript Edited for Clarity