Arthur He: Hey, good afternoon, Andrew and team. This is Arthur on for [indiscernible]. I just had a quick one regarding the GRANITE. Just curious, when you see the ctDNA reduction opposed to the chemo, did you see, is there any correlation for the reduction with the baseline character of the tumor? I just curious, is there any way you guys could minimize these noisy background if you go into use ctDNA to further as a biomarker for the evaluation in the future?
Andrew Allen: We haven’t performed those analyses yet. Obviously, it’s a phenomenon that relates to the tumor response to chemotherapy, specifically to FOLFOX/FOLFIRI plus bevacizumab. And that regimen is only used in colorectal cancer. Derivatives of it, I guess, used in pancreatic cancer, but it wouldn’t have great utility outside of this colorectal cancer setting. And so determining who does well on chemotherapy is obviously not particularly critical to our development program going forward.
Arthur He: All right, thanks. And my second question for the update, further readout in the third quarter, besides the PFS, mature PFS and more ctDNA data, what other data set could we expect can give us more information regarding the pivotal study design?
Andrew Allen: Yes, so the overall survival data will still be very immature at that point. Median overall survival is around two years in this disease. So again, following the same logic, if PFS is basically around one year or just shy, you need to give it another year to get to very mature OS data. Now we are doing some additional analyses. Karin, our Head of R&D is on the phone here. So, Karin, perhaps you want to give a little flavor as to what additional correlates we might have for the Q3 update.
Karin Jooss: Yes, we are looking in a subset of the patient population. We perform TCR-Seq, we perform ALICE bot [ph] analysis. So we do look at translational data, which we actually have published significant data in our nature medicine papers. So we do some of that work, but we don’t anticipate the data to look any different to our goal for manuscript. But TCR-Seq an ALICE bot you can anticipate that we are also potentially, if we have enough cells, perform ICS, looking for poly functionality or even killing. So we have a great toolbox, and depending on the number of cells we have from these blood trials, we will, in a subset only of these patients looking for T cell response, TCR-Seq, as well as functionality of the T cells.
Arthur He: Great. Thanks for the addition of color. Thanks for taking that question.
Andrew Allen: Thank you.
Operator: The next question comes from Corinne Johnson with Goldman Sachs. Please proceed.
Unidentified Analyst: Hi, this is Omari on for Corinne. I have a couple questions. What gives you confidence that the PFS will strengthen as the data matures? And on the hazard ratio is the piecewise Cox pH model of prespecified analysis.
Andrew Allen: So the confidence comes from two primary sources directly and one indirect source. First of all, we saw the same phenomenon of PFS and OS extension in the Phase 1/2 study in third line colorectal cancer patients, not a randomized study, but those who had reductions in their bone biomarkers, including ctDNA, had this apparently extended PFS and OS. So we’re repeating that observation in this study. Secondly, the high risk analysis gives you a way of peering into the future, and it’s obviously not a perfect analysis. No one can foretell the future. If they could, it would be easy. But it’s a reasonable way of asking the question, if I expect the low risk group to behave similar to the high risk group, then what would I see in the future?
And obviously, the high risk analysis is quite mature and looked very encouraging. So that’s the second direct bit of evidence to encourage optimism for the Q3 mature data. The indirect support comes from the notion that the data we’re waiting for are in the low risk patients. And data from other players, most notably Moderna, suggests that those low volume disease patients are the ones that do best on vaccine based immunotherapy. Therefore, if you believe that that applies to this study, we should see signals at least as strong, if not stronger, in the lower risk population as the dataset rounds out and matures. Your second question was about the GPW statistical test, and I couldn’t quite catch it, was the question, are we using that in the summer in Q3?
Unidentified Analyst: Piecewise Cox pH model. Was that a pre specified analysis for this study?
Andrew Allen: Yes. So we’ve actually been in a lengthy dialogue with FDA around the way to statistically analyze the study because it is clear that patients are randomized at the beginning of their induction chemotherapy, and for the first, on average, five months, the treatments are identical across the two arms of the study, and then the treatments diverge, and therefore, the Kaplan-Meier plots of progression free survival will not meet the proportional hazards assumption. And that is a requirement if you’re going to use the standard logrank test. So we knew that the way the study was designed, logrank was not the appropriate statistical test and appropriate way to analyze these curves that, in fact, what you wanted to do was a time weighted system.
So we entered this discussion with the agency and we settled with them on GPW, the global. Sorry, generalized piecewise analysis. And we’ve done a very simple model for this study. Any progression event prior to six months is given a weighting of zero, and any progression event after six months is given a weighting of one. That’s the way we’ve analyzed these data and that was pre-specified.
Unidentified Analyst: Thank you.
Andrew Allen: Thank you.
Operator: Thank you. At this time, there are no further questions in queue. I’d like to turn the call back to Andrew Allen for closing comments.
Andrew Allen: Thank you very much. That represents the end of our call, so we have nothing further to add. Just like to thank you for your time and attention. Obviously, we’re very excited to see these data in Q3, and we hope to be able to really move the ball forward for these patients who’ve been waiting a long time, for some reasons, for optimism and with that, thank you very much.
Operator: Thank you. This does conclude today’s teleconference. You may disconnect your lines at this time. Thank you for your participation, and have a great day.
