At this year’s meeting we reported the end of phase I results for the KTE-X19 CAR T-cell trial in adults with relapsed and refractory acute lymphoblastic leukaemia. Acute lymphoblastic leukaemia remains an unmet need, the majority of adult patients will relapse and when they do relapse the outcomes are poor.

This phase I was launched specifically for patients with primary refractory disease or those who had relapsed after two or more lines of therapy. The KTE-X19 is very similar to the CAR T-cell therapy that’s been approved in diffuse large B-cell lymphoma. It utilises the same murine anti-CD19 as CFV and the same CD28 co-stimulatory domain. Where it is different is we isolate T-cells prior to viral transduction and manufacture so that we can avoid transduction of the blasts.

We overall apheresed 54 patients and ultimately treated 45. Infection was the most common reason why patients didn’t proceed from the apheresis to the treatment with KTE-X19. For the 45 that we did enrol, as perhaps expected, they were a heavily refractory population, 68% had three or more lines of therapy including 42% who had prior blinatumomab. We reported at the previous meeting that blinatumomab doesn’t appear to be associated with inferior response to KTE-X19. About 30% had had prior transplant and around 37% were primary refractory. These patients had very high blast bone marrow burdens when they came onto the study, the median was around 60%. In fact, the study did not allow for the treatment of MRD, patients had to have greater than 5% marrow blasts just to be able to enrol.

Of the 45 patients that we treated we reported on safety data at this year’s meeting and then we have efficacy data for 41 of the 45, simply because of our data cut-off time point. The safety, we have no new safety signals. We saw two patients with a grade 5 attributable mortality, one with multi-organ failure in the setting of CRS. This patient was treated at the highest cell dose so I should rewind and say that we explored three different doses for this trial – 2x106, 1x106 and 0.5x106 cells/kg. We ultimately settled on the 1x106 cells/kg dose and I’ll come back to that in a moment.  We saw two deaths on the study, one was at the 2x106 cell dose in the setting of CRS and multi-organ failure and the other one was at the 0.5x106 cells/kg dose where the patient passed away of a stroke, again in the setting of CRS and neurotoxicity. Both of these have been previously reported so we’re delighted to say that the overall mortality amongst the 45 patients that we treated remains quite low, it’s 4% for the trial and is in line with what we’ve seen in diffuse large B-cell lymphoma and the other studies of CAR T-cell therapy that have been conducted to date.

In terms of safety we still do see high rates of CRS and neurotoxicity. If you look across the entire trial it was around 30% for grade 3 and above CRS and it was around 38% for grade 3 and above neurotoxicity. We wanted to figure out whether we could do better. We were already seeing very encouraging activity and an improvement in the overall CRS rate as we went from 2x106 to 1x106 cells/kg. We wanted to see whether we could do better in the group treated at 1x106 cells/kg so we developed a modified toxicity algorithm. Specifically we approached these patients with steroids at the onset of grade 2 neurotoxicity rather than waiting for grade 3 neurotoxicity to emerge and we also restricted tocilizumab for the treatment of CRS. So we’re no longer recommending to investigators if they have isolated neurotoxicity try tocilizumab first, we said, no, just got straight to dexamethasone, again if they’re meeting that grade 2 cut-off.

What we saw was dramatic. If we restrict our analysis just to those patients treated at the 1x106 cell dose previously we saw around 57% grade 3 toxicity and I think it was 8% grade 4 neurotoxicity. With the modified treatment approach we saw no grade 4 toxicity and grade 3 toxicity in 11% which equates to one patient out of that group. Not only do we see a decrease in toxicity we also saw a decrease in the duration of that toxicity so the median went from 20.5 days to 11 days in those patients who experienced any neurotoxicity. So we were excited because we feel like we’ve got a dose now that brings down our CRS rates and a toxicity algorithm that brings down our neurotoxicity rates, what about efficacy? Again, although it’s a phase I trial and this is a secondary endpoint, it still deserves some attention because we just want to understand this is going to be feasible as we move forward into the phase II clinical trial.

So, looking across all three doses we saw very high rates of CR/CRI, all of which were MRD negative so in total 68% achieved a CR/CRI, again, all of which were MRD negative. If we focus on our recommended phase II dose the 1x106 cells/kg dose, 84% achieved a CR/CRI, again all of which are MRD negative. More importantly, these responses were consistent across all of our key covariants – prior transplant didn’t affect this, if they were primary refractory it didn’t affect this endpoint – and so we’re very encouraged by this response rate.

We also did present durability of response. Again, this is preliminary because this is a phase I trial, but we’re very encouraged. For this highly refractory patient population the median duration of remission at the 1x106 cells/kg dose was 12.9 months once censored for transplant. Interestingly, when we no longer censor for transplant we don’t reach our median duration of response which begs the question is this a bridge to transplant? The answer is I don’t know. We transplanted a total of three patients at the 1x106 cells/kg dose and it’s really too few patients to make any meaningful conclusions to say, yes, this is the standard we should be taking or, no, this isn’t the standard we should be taking. For now it’s going to be an individualised decision and I really like some of the data that’s emerging around the role of MRD or in some cases very proliferative leukaemias that may be associated with high LDH or low platelet count, as the Hutch has developed. There may be surrogates that we can use to say, hey, this is a higher risk patient where we want to think about transplant in this setting rather than approaching it as a one size fits all.

So, to summarise very quickly, we treated 45 patients including 23 at the 1x106 cells/kg dose. We had very encouraging efficacy, very encouraging safety. We are already about two-thirds of the way into our phase II when we’re hopeful we’ll have some good news to share in the near future.