Clinical Trials Evaluating CAR T-Cell Therapies in Second-Line Treatment of Patients With DLBCL

Grzegorz S. Nowakowski, MD: We bridged the patient, maybe the patient didn’t require therapy, and now we have 3 studies with CAR [chimeric antigen receptor] T-cell therapies in the second line, with 2 consistent and 1 with a slightly conflicting result. Fred, do you want to comment?

Frederick Locke, MD: Yes, I’ll start off with the ZUMA-7 trial [NCT03391466]. This is a randomized clinical trial in the second-line setting comparing axicabtagene ciloleucel [axi-cel] to what has been our standard of care in the second-line setting for 30 years based on that PARMA study in the 1990s: if you give high-dose chemotherapy and the patient responds, then you can consolidate that with autologous hematopoietic stem cell transplant.

This study was designed to look at event-free survival [EFS] as an end point. That’s because if a patient’s randomized to standard of care and they get a dealer’s choice of protocol-defined chemotherapy regimens and they don’t respond, meaning they have stable disease, they wouldn’t benefit from a transplant. That necessitates a switch to a different therapy, and that triggers the event. It isn’t a progression event. That’s why it was designed as an EFS endpoint. Of course, of the patients on ZUMA-7 who were randomized to axi-cel, over 90%—94% if I remember correctly—got axi-cel CAR T-cell therapy. From the time of randomization, they had their leukapheresis for manufacture within 5 days and the turnaround time was very fast because axi-cel can be made very quickly. Of the patients who were randomized to standard of care, about 33% or 35% made it to autologous transplant.

It’s important to note that we aren’t comparing transplant vs CAR T. We’re comparing that treatment paradigm of chemotherapy and response. When we look at the event-free survival in ZUMA-7, CAR T-cell therapy clearly did better. Over 300 patients were randomized, about 180 in each arm, with a median follow-up of about 2 years. The event-free survival is clearly better and the stat that illustrates that best is if you look at event-free survival, the Kaplan-Meier estimate at 24 months of ongoing EFS in the axi-cel arm is much higher. About 41% of patients randomized to axi-cel got CAR T-cell therapy, remained in remission, are alive, and didn’t need other therapy. In the standard-of-care arm, that’s about 16%. That’s pretty remarkable.

Overall survival wasn’t different, and that was an interim analysis on overall survival. We’ll take a look at that at the 5-year time point and whether it’s numerically better for CAR T-cell therapy and the curves seem to separate but it doesn’t meet the statistical significance. Importantly, about 55% of all those randomized to standard of care eventually got CAR T-cell therapy off-trial.

Of the other 2 studies, one of them showed similar results. You mentioned the TRANSFORM study [NCT0357535]. It had a pre-planned interim analysis and the same study design with event-free survival as the end point. But the interim analysis revealed the wildly different differences in EFS curves, and it has much less follow-up, 6 or 7 months median follow-up. But it looks almost superimposable to the ZUMA-7 curve. One of the differences in that trial is that bridging therapy was allowed, which wasn’t allowed on ZUMA-7. But ZUMA-7 led to FDA approval for use in the second-line setting. That was patients who had relapsed within 12 months of initial therapy.

Finally, the BELINDA study [NCT03570892] is the third randomized trial. TRANSFORM was lisocabtagene maraleucel [liso-cel], ZUMA-7 was axicabtagene ciloleucel, and BELINDA is tisagenlecleucel [tisa-cel]. Unfortunately, that was also designed as a 1:1 randomization event-free survival, but it didn’t meet its primary endpoint and the EFS curves looked very similar. Certainly, the study design could have played a role. Patients were allowed to get multiple rounds of chemotherapy—not just multiple cycles—but on the standard-of-care arm they could get R-ICE [rituximab, ifosfamide, carboplatin, etoposide] for example, and then if they progressed, they could get something different, such as a gemcitabine-based regimen. They’re stretching them out, trying to get them on autologous transplant.

Similarly, on the CAR T arm, they were giving bridging therapy and the time for manufacturing was very long. I guess we have to ask the question: does tisagenlecleucel maybe have a little less activity than the other 2 agents? One final thought is that the amount of lymphodepleting chemotherapy that we give is different with tisa-cel. It’s actually a lower dose of FLU [fludarabine]–CY [cyclophosphamide]. Could that be contributing? All these factors come together but the results are pretty clear to me that with axi-cel and liso-cel, for a patient who has progressed within 12 months of initial therapy and hasn’t gotten second-line salvage chemotherapy regimen yet, we should try to get them to CAR T-cell therapy.

Grzegorz S. Nowakowski, MD: I was a little surprised that the event-free survival and progression-free survival curves in those second-line studies weren’t much better than what we had seen in the third line. I found that quite surprising. How do you explain it?

Caron Jacobson, MD: Seventy-five percent of the patients who went on these studies were primary refractory. We’ve already talked about the fact that those patients do less well with CAR T-cell therapy, so taking patients who are primary refractory and progressing is picking the worst of the worst. If you were to take people in the second line who relapse at 18 months or 2 years, we’ll see a better EFS curve.

One of the interesting points that Fred raised was looking at those overall survival curves. We’re moving into a realm where a lot of new therapies for large-cell lymphoma have activity for 12 to 24 months and maybe aren’t curing patients but they’re keeping them alive and in remission for 12 to 24 months at a time. All of a sudden, overall survival from the time of being chemotherapy refractory is much more prolonged and, just like in follicular lymphoma, is it the endpoint we should be looking at to make decisions when the readout is going to take such a long time?

What I also thought was pretty interesting is that despite the fact that patients on the TRANSFORM study were allowed to cross over to get lisocabtagene maraleucel if they either didn’t respond to salvage chemotherapy or relapsed after autologous transplant, or patients on the ZUMA-7 study were eligible for commercial CAR after that and were able to get commercial CAR, not everybody got the CAR T-cell therapy. That tells us that you lose some people if you wait too long If you over-treat them with chemotherapy.

Matthew Lunning, DO, FACP: Forty-five percent of people on ZUMA-7 who were in the standard-of-care arm didn’t make it to a commercial CAR T cell.

Frederick Locke, MD: It’s a little less than that.

Caron Jacobson, MD: Because some of them responded.

Frederick Locke, MD: Of all, some of them are still in remission.

Transcript edited for clarity.