Transcript:Krishna V. Komanduri, MD: I think the next thing that’s really exciting in non-Hodgkin’s lymphoma right now, and certainly is a topic of a lot of focus at this meeting, is autologous cellular immunotherapy and CAR-T cell therapies. And I think we have some very interesting data that we’ve presented at this meeting. We have the Abramson abstract, and Leo, you’re part of that study looking at relapsed/refractory DLBCL. I thought that perhaps you could share with us your experience.

Leo I. Gordon, MD: This is a study looking at JCAR017, which is the latest iteration of the Juno CAR-T. So far, there have been 39 patients, 28 have been phoresed, we’re reporting on 14 treated, and 10 of those patients have had prior transplant. So, they’re a fairly heavily pretreated group of patients. The overall response rate, and when you do PET scans on day 30 and then at day 90, is about 80%, with about 70% complete remissions. Now, the problem with that is that the follow-up so far has been pretty short.

One of the major concerns about CAR-T—and I don’t think anybody is sure yet about what the reason for this is—is the cytokine release, the cytokine storm that occurs, and the neurologic toxicity that occurs. And in the ALL patients, for some reason, that has been a much higher risk. We haven’t seen that very much. It’s less than 20% for the cytokine storm, and there have been no cases of neurologic toxicity, at least so far in this trial.

We are using Cytoxan/fludarabine as lympho-depleting chemotherapy just before the cells are infused. My experience has been that people have tolerated that fairly well. They do actually have neutropenia prior to coming in for the CAR-T. And we’re watching them fairly closely in the hospital, and many other places. I know Cam Turtle and the group in Seattle have been doing it long enough that they’re doing most of these as outpatients actually. And I think Steve Schuster at Penn also has done most of these as outpatients in their iteration of CAR-T therapy. So, I think the concept is a very interesting one. To me, it’s similar to allogeneic transplant, but without the risk of graft-versus-host disease. I think what’s interesting, to me anyway, is that if you look at trying to collect the cells and trying to maintain them, the idea of perhaps using checkpoint inhibitors to increase the number of central memory T cells that you want to collect for this, and then to maintain them afterwards, might be something that will be added to some of these protocols. It may increase the efficacy of this approach.

Krishna V. Komanduri, MD: I think a few things that are worth pointing out about the JCAR017 program is that there are measured equal proportions of CD4 and CDA T cells, which is a little different than the JCAR015 program. The JCAR015 program has been recently the subject of a lot of attention in ALL because there have been clinical holds related to cerebral edema and neurotoxicity. We don’t know if that’s because of the composition of the product. We also know though that B-ALL (B-cell acute lymphoblastic leukemia) does appear to be a different disease ironically—even though we don’t know that there are CD19 expressions in the brain—that it may have something to do with the biology of B-ALL. The rate of neurotoxicity is higher. Certainly, I think it seems to be true of cross programs, both the Novartis trials and the Juno studies have seen higher rates of neurotoxicity in B-ALL.

We have, as I said, been part of the ZUMA-1 trial that’s being reported as a late breaker on Tuesday in refractory aggressive non-Hodgkin’s lymphoma. Again, like the JCAR017 data that you were talking about, it’s a CD19 CAR. We certainly saw a significant fraction of patients with cytokine release syndromes, as well as neurotoxicities. But the rate of deaths in the study was 3%; 2 out of the 3 deaths were thought to be associated with the product and related to either cytokine release syndromes or, especially, neurotoxicity. But overall, there was an over 75% response rate to CAR T cells in a very highly refractory group of patients historically—50% of those patients had CRs. And at 3 months, it is notable that there are roughly 39% to 40% of those patients that are still in remission.

Now, there are mixed things. I think that if you looked at Jim Cokendolpher’s studies at the NCI, that I think led to the IP that were formed at Kite, the initial response rates are actually a little bit lower, but the 3-month response rates were more similar to what they saw initially. The interesting thing is in the initial NCI experience: patients who had a CR at 3 months typically maintained that CR, and those CRs were durable, but that is not something that we’re used to seeing even in the allogeneic transplant setting. We’re used to seeing a fall off, and we certainly don’t see plateaus of relapsed curves at 3 months.

So, I think that although these data are very, very exciting, the CR rates and the overall response rates might be 5 to 6 times what you might expect with other historical therapies in this setting. I think that the durability of these therapies still needs to be observed. We really have a 3-month follow-up; we’ll see 6-month follow-up soon. But I do think that from a regulatory standpoint, we’re likely to see approval on the basis of what we’ve seen, in 2017. But there’s a lot to learn.

Transcript Edited for Clarity