Nilanjan Ghosh, MD, PhD: It appears that using a CD28 [cluster of differentiation] costimulatory domain can make the cells grow faster and rapidly expand; this may have practical implications. For the axicabtagene ciloleucel [axi-cel] product, the median turnaround time is 17 days. For patients who are awaiting to get this therapy urgently, this may be an important consideration. At the same time, it should be noted that, at least from the data we have, the rapid expansion could lead to higher cytokine release syndrome [CRS] and neurotoxicity. It would be hard to compare the different studies because they were not randomized. The other thing that would be interesting to see in retrospect would be the costimulatory-domain choice to figure out about persistence of CAR-T cells in the long-run and whether it has an effect on T-cell exhaustion.

Matthew Lunning, DO: I personally do not think you can compare the toxicity standpoints directly. Maybe they will be in the same setting in the future—that’s really the only way to get a definitive notion about whether one construct is more effective than another.

If you want to take a step back and look at the patient population in which axi-cel was studied, using their CRS guidelines, you know there was all-grade CRS in the 90% range and all-grade neurotoxicity in the 50% range. They had characteristic days of onset for the CRS and neurotoxicities, with broad ranges with regards to toxicity. But the toxicity can be mitigated or reduced with the use of tocilizumab [Actemra], which blocks the IL [interleukin receptor]-6 receptor. If the neurotoxicity does occur to a significant grade, then dexamethasone [Ozurdex] can be used to help lessen the severity.

Speaking towards the toxicity profile with other constructs, it’s difficult because, specifically in the tisagenlecleucel [tisa-cel], they used a different grade of scale for CRS. That’s where it becomes difficult to jump between grade ½ or grade ¾ toxicities between constructs. You have to acknowledge different lympho-depleting chemotherapies, patient populations, dosing options, and disease burdens. Some patients were unable to receive bridging chemotherapy, whereas a majority of patients with other constructs were. How does that play into the differences between constructs, toxicity and safety profiles? I don’t think we can say right now. We can only report the numbers and hope we can increase the safety and decrease the toxicity, by finding strategies or algorithms that lead to lessening the one without compromising the other.

Nilanjan Ghosh, MD, PhD: The follow-up for these studies have not been long; however, the JULIET and ZUMA-1 studies had a bit longer follow-up than the TRANSCEND study. The median follow-up for the JULIET was 14 months, and the median follow-up for ZUMA-1 was 15 months, whereas the median follow-up for TRANSCEND was less than a year; it was actually about a year for survival endpoints and 8 months for durability of response. The longer follow-up is certainly needed. It does appear that a 3-month time point is important for looking at responses and predicting long-term durability.

Transcript Edited for Clarity