Prof Evangelos Terpos – University of Athens, Athens, Greece
Prof Nina Shah – University of California, California, USA
ET: Nina, thank you so much that we are together today for this, I think, interesting topic. And my first question is when we read a clinical trial or a clinical paper on multiple myeloma we read about progression free survival or response, especially minimal residual disease negativity. So, for you, what do you believe is the most relevant endpoint for a clinical trial – PFS or MRD?
NS: It really depends on the trial but I would say in general I really like to see the PFS data because I think that’s the most robust data to get an idea of how the drug is doing, particularly for a disease like multiple myeloma where maybe if things go up and down by numbers, maybe the light chains go here or there, it doesn’t really mean you have to change treatment until it has reached a certain threshold. So that’s the question you want to answer to the patient is how long will I be on this therapy or how long will it be before I need to change therapy. For that I think the PFS is very useful, although I will say that MRD is becoming the faster way of trying to predict PFS, right?
Many of the trials that look at combining or comparing regimens for induction treatment or transplant or no transplant, the primary endpoint is PFS because, again, you want to be able to tell the patient, ‘Hey, you’re going to be able to live this long without changing therapy.’ Overall response rate , as you know, is one of the ways the agencies decide if these drugs are good enough to approve in later lines. Because of that, not only is it response rate by IMWG but also MRD status that is helpful to us to understand very quickly how good that therapy is.
ET: Then we have the hazard ratio that goes into this equation. What is your opinion about the hazard ratio?
NS: So, for example, if the hazard ratio is 1, of course that means there was no difference between the two arms. But, as the hazard ratio gets lower you can say roughly, and this is the way I think about it, for example a hazard ratio of 0.5 there is a 50% decrease in risk of progression. It’s not the exact statistical way to say it but that’s the way to think about it.
ET: We like to see the median PFS in the trial and we’ve seen that in one trial, for example, one regimen gives you 18 months and the other gives 15 months compared to the same, let’s say, non-experimental arm. We may have the opinion that this 18 months is much better compared to the 15 but then when you go to the hazard ratio you can see that this is totally probably the opposite because the control arm is doing totally different.
NS: We have so many good drugs in myeloma, we want to offer our patients the best that they can get. So I think the hazard ratio is one way. Understanding the patient populations is a very important one as well. This goes back to that Table 1, so who are the patients – are they as heavily pre-treated in both of the trials? Are they the same refractoriness to certain agents that are important to us, especially in the relapsed/refractory setting?
NS: I think all of us, we love MRD but, again, we don’t know what to do with it. It doesn’t mean there’s no disease, it means we just can’t find it. It’s minimal detectable disease, as some people call it. So as we just talked about, the quality of the myeloma is so much more aggressive in the later lines, the fifth, sixth, seventh line. Even if the MRD is zero by 10-6 analysis, that doesn’t mean that by 10-8 it’s zero; it means that you just can’t find it at that level of sensitivity. We know myeloma is incurable so whenever one cell, progenitor cell, any cell that’s left is going to grow quicker and more aggressively in the later lines than it does in the front line. That’s exactly why I think we’re seeing people going to MRD negativity with some of these very, very potent immunotherapies and yet at one year they progress.
I do want to say that so far what we’ve seen in the later line trials, and similar to the front line trials, if you achieve MRD negativity, whatever the measure is, if you compare the patients who get it, MRD negativity , versus who don’t always the patients who have got MRD negativity do better by PFS. So we know that. So then the question is how do we measure it. We measure it the best we can. We’re trying to use NGS, now there’s some aspect-based analyses. Whatever we can do to understand the most sensitive way of finding it, at least we’ll be able to tell our patients, ’You’ll do better if you’re in this negativity arm.’
NS: I think the best place to measure MRD for those patients is right before maintenance therapy. So after the transplant, before you start something new, because that’s a very clean breakpoint that you can say, ‘Okay, how is this patient going to do?’ That is what has been done in a lot of the trials. But I think that having at least that data, it helps me to understand where this patient is going. Maybe if I know that their MRD is positive, even though their M-protein might be very low, that I would consider very quickly acting upon them if they start to have biochemical progression.
NS: We’ve seen this recently where PFS disease control is good but then the toxicities make the overall survival not good. This is very important because what this means for us is, one, we have to report the AEs very clearly and, two, that if we choose this drug we must pay attention to those AEs and, for example, give preventative antibiotics or be ready with tocilizumab or whatever it is that we’re going to use to mitigate any of these side effects. So I think that AEs are really important because they tell you about the patient experience and we sometimes forget to look at that table.
NS: For anything with a comparator arm I find that progression free survival, for me, is the most important. But I do agree that getting an earlier answer with MRD data is usually predictive of what will happen in PFS.