Matthew Lunning, DO: There are differences between axi-cel, tisa-cel, and liso-cel, and you really have to go down to the CAR-T cell construct itself while they are all targeting against CD19 [cluster of differentiation 19], to see the receptor that lives on the outside of the lymphoma cell or the diffused large B-cell lymphoma [DLBL] cell. What then happens after engagement may be a little different; this is represented in the costimulatory molecule. In the axi-cel construct that is CD28, which provides signal 2, whereas in tisa-cel and liso-cel, 4-1BB is the costimulatory molecule. I think the other differences between the constructs is the dosing. The dosing with axi-cel is given on a weight-based formulation up to a total cell dose of 2 x 10⁸ CAR-T cells. Whereas, tisa-cel is actually a variable dosing: it can go anywhere from .6 x 10⁸ all the way up to 6 x 10⁸ CAR T-cells. Lastly, with liso-cel, that CAR-T construct is actually split between a flat dose of CD4 cells and a flat dose of CD8 cells. But currently I believe in the core dataset the total cell dose is 1 x 10⁸CAR T-cell doses.

Leo Gordon, MD: It’s hard to compare the 3 studies, partly because they weren’t done as a phase III randomized study. As I said, there were some differences in patient populations. There are major differences in the product themselves, which is probably the most important thing. We know this is sort of a second generation of CAR-T cells—it became apparent that one needed to have a so-called co-stimulatory molecule to get expansion to these cells. Without the costimulatory molecule, expansion of these cells is difficult. There are 2 costimulatory molecules that are currently in play: the CD-28 molecule, which is what’s used in the ZUMA studies, and the 41BV costimulatory molecule, which was used in the TRANSCEND study from Juno Therapeutics and Juliet study from Novartis.

The other difference is, in the TRANSCEND study, there are preclinical data from Stanley Riddell, MD, at the Fred Hutchinson Cancer Research Center, that suggested that an equal amount of CD4 and CD8 cells infused in an animal model might benefit in terms of efficacy. In the TRANSCEND study there’s a fixed ratio of CD4 to CD8 cells that are given to the patient, whereas that ratio is not preestablished in the ZUMA-1 or JULIET studies.

Whether that makes a difference or not is undetermined. Certainly the preclinical model suggests that it will, but we have no evidence yet that is going to make a difference. In fact, when the KITE group looked at their ration retrospectively, they were close to one-to-one comparisons. Although it wasn’t necessarily planned, the ration turned out to be one-to-one. Whether that’s going to make a difference, we do not know. There was also a difference in dosing in the studies.

There is a sense that CD28, as a costimulatory molecule, leads to more rapid expansion of the CARs, whereas the 41BB costimulatory molecule leads to a less rapid expansion. All this is going to have to be sorted out. Whether there will be a randomized comparison ever, I do not know.

Transcript Edited for Clarity