At ASCO 2020 we presented an international phase II randomised study of the PD-L1 inhibitor durvalumab combined with tremelimumab, a CTLA-4 inhibitor for patients with advanced hepatocellular carcinoma. The study had several different parts – first there was part one which was really examining the safety of the combination of these two drugs and we reported those results earlier. For ASCO this year we reported the randomised phase II portion with expansion cohorts looking at four different treatment arms. There was the durvalumab or PD-L1 inhibitor monotherapy arm, the tremelimumab CTLA-4 inhibitor monotherapy arm and two different combination arms. One combination arm was entitled T75+D which was the combination of tremelimumab at a flat dose of 75mg for four sequential doses combined with the PD-L1 inhibitor durvalumab as a backbone continuously every month. Then the next combination arm was a novel regimen of a higher dose of tremelimumab, 300mg flat dose, given only once as a priming dose along with the durvalumab monthly as a backbone.
This novel fourth arm of the T300mg dose was based upon translational and preclinical studies showing that a single higher dose of tremelimumab as a CTLA-4 inhibitor can elicit a burst of proliferative activated T-cells in the peripheral blood and, furthermore, that this burst does not recur with repeated dosing. So that led to the hypothesis that perhaps a single dose of tremelimumab would be enough to prime the immune response in this combination and perhaps could also spare the toxicity that we see with sequential prolonged CTLA-4 inhibitor dosing. So that was the rationale for the study – four different arms, again, two different monotherapy arms and two different combination arms.
It’s noteworthy that patients in this phase II trial were predominantly a second line population with about 77% having received prior sorafenib in the T300+D novel priming dose arm. Though eligibility did allow patients to enrol who had refused prior sorafenib, so a subset of patients in each arm were enrolled in the first line context without prior systemic therapy. Overall, however, the predominant population was a second line patient population.
The primary endpoint of the study was safety and all four arms demonstrated acceptable safety. In fact, the rates of corticosteroid requirements, one of the potential toxicities that we see with CTLA-4 inhibition, were favourable across all four arms, particularly for the novel combination arm of T300+D, with only about 24% of patients requiring steroids for immune mediated adverse events.
When we look to the efficacy or the activity of the four different arms we saw that the combination arm with the priming dose of T300+D had the highest rate of objective response, around 24% of patients having a meaningful and measurable tumour shrinkage. Then the duration of these responses was quite long and, in fact, the median was not met at the time of reporting. So the outcomes for the combination arm with the priming dose tremelimumab were best for response rate by comparison with the other three arms.
When we looked at the important endpoint of overall survival the novel priming arm of the T300, or 300mg tremelimumab priming dose plus durvalumab backbone, also had the highest median at 18.73 months by comparison with the other three arms, lending support to the strength of its objective response rate outcomes and validating the translational hypothesis that led to the development of that arm.
We also looked at landmark proportions of patients with survival after 12 months and 18 months and saw that 60% of patients in the T300+D arm were alive at 12 months and 52% at 18 months, another important landmark event for a CTLA-4 plus PD-L1 combination like this.
So taking the data in sum we found that the T300+D arm had the best ratio of benefit to risk and thus has been the arm that has been taken forward into the phase III trial, HIMALAYA, looking at this combination regimen with the single priming dose of tremelimumab plus durvalumab backbone versus durvalumab monotherapy versus the standard control arm of sorafenib as monotherapy.
A particularly interesting outcome of this study were the biomarker analyses where we looked at peripheral blood using flow cytometry for lymphocyte populations on day 15 and showed that there was a T-cell population comprised predominantly of activated or proliferative T-cells that were CD3 positive, CD8 positive and Ki67 positive that discriminated between the four different treatment arms and, in fact, were highest in the T300+D priming dose arm, supporting again the translational rationale that led to the development of that priming dose arm. Further, if we looked at the proportion of those activated and proliferative T-cells across all four arms we saw that the patients who had complete response or partial response to treatment also had the highest proportions of this proliferative T-cell signature population on day 15. This really reinforced the mechanism of response and highlights that the combination with the single priming dose of tremelimumab is activating a peripheral T-cell population to mediate its effects.
What does this mean looking to the future?
As we look forward to the results for HIMALAYA it will be quite exciting to see how this novel priming dose of tremelimumab plus durvalumab backbone performs both for efficacy as well as the important secondary endpoint of toxicity in hepatocellular cancer first line population. I also think that the mechanistic biomarker results from the study 22, the phase II trial, really underscore the potential for this priming dose approach to be explored in other tumour types potentially as a means to achieve the T-cell activation and immune response without the consequences of toxicity that can arise with prolonged exposure.