SOHO 2026 | Lessons from Phase I KITE-222: antigen escape, expansion, and the future of CAR-T in AML

Yeah, I mean, I think we’ve been very disheartened. We’ve had a tough go in myeloid CAR-T. So this KITE-222 targets CLEC12A or what’s also known as CLL1. This has been our third main target. 33123 have been looked at the most, but CLEC12A, definitely a lot of data and actually maybe a little bit more pure on leukemic stem cells. I think we have a couple of issues. One is who we’re treating. There’s like a logistical practical issue, which is a huge issue. One, these trials take forever. Two, we’re treating very refractory patients, some of which have proliferative disease that are very poor candidates for CAR-T, but they’re the ones that are trying to get on these kind of last-ditch trials, you know, from that perspective. And we’ve said for years, like, lower burden, you know, better effector-to-target ratio, these things matter, but yet the ones that we’re trying are fifth line, second allo, you know, from that perspective. And so I think there is a real logistical challenge. And there’s a challenge, I think, you know, of course we need to think about safety but how long it takes us to get through cohorts. So even in this trial I would argue that we actually didn’t finish the trial yes we got to dose level three but we didn’t even get to like fully build out like it was this truly enough patients even at dose level three in this trial I think was a question. We did I think, so this is essentially the same technology with a CD28 costimulatory domain as with other approved Kite products, and what we did see is that expansion, although there was, was quite a bit less. And so the intensity of lymphodepletion may matter. So we did amend the protocol to increase the dose of cyclophosphamide, but a small subset of, we did see better expansion. And we had one patient that achieved blast clearance, although that patient actually received double lymphodepletion. And actually, unfortunately, maybe died from complications of low blood counts, potential HLH from that perspective. So there’s just so many questions in all of this that it’s been difficult.

I do think there’s a couple of things. One, we have to think much more outside the box. And so I’m more, I’ve been excited, but more excited. Can we target more than one antigen going for efficacy? So for example, the Senti Bio program that Steve Strickland’s presented originally now got updated at ASH, looks pretty good. I mean, small numbers, but response rates are quite reasonable. It is with a different and high-dose lymphodepleting sort of salvage chemotherapy. So I think there’s questions on what’s CAR versus what’s chemo. But I think those sort of strategies may make sense or allo concepts. And so I’m excited by avan-cel, which is a CD123. This was a plenary at EHA presented by the German group. And they see expansion actually on par with what they’ve seen with lymphoid targeting therapies. And so because we don’t have anything, I think autologous versus allogeneic is a great question. We had Selectus where we did see some, I had one amazing response, 18-month MRD negative. So I think the point is we have to improve the technology. Think about expansion. Maybe allo could be better in some of these settings. I think two, we have to treat the right patients. And three, we got to be more out of the box. Yes, I would love to have one antigen, get it approved, then go after another one. But I think we have to think more out of the box. We did see clonal escape even on this trial. And I think this is probably the key other finding is actually the CLL1 cells went away in a decent proportion of patients. Actually, it was CLL1-negative cells that were there. And I think the problem was preclinically, people say, hey, CLL1 is on most of these leukemic stem cells. But what is most? What percent? Is it high or low expression? I think this heterogeneity is a big issue. So maybe if we treated a really more homogeneous population of true high CLL1s, would they have done different? I don’t know. Because again, the trial ended earlier than I’d like. But these sort of questions, I think we have to try to answer multiple of them in sort of future CAR-T trials. I’m still very optimistic. We will have our sort of cellular therapy or bispecific T-cell engager approval in the not distant future. But we’re not going to be the same as targeting CD19 or BCMA.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.