AC: Hi, welcome to the ecancer educational myeloma discussion. Today it’s my pleasure to have Roman, hi, join me in a discussion about treating patients with multiple myeloma in the real world. My name is Ajai Chari, I’m a Professor of Medicine at the Icahn School of Medicine at Mount Sinai in New York where I’m also the Director of Clinical Research. Roman.
RH: Hi everybody, nice to be here with Ajai. My name is Professor Roman Hájek, I’m working in Ostrava in the Czech Republic and I’m Chairman of the Czech Myeloma Group.
AC: Great. So, Roman, we’ve both worked on a lot of real-world data, maybe it will be good to start with what are your thoughts on what is the role of real world or why is it important? How do you interpret real-world data for myeloma?
RH: It’s a big discussion during the last five years because we know that we need clinical trials, no doubt, they are very informative. But, on the other side, they are not reflecting the real situation. So real-world evidence is important and I always say clinical trials means perfect data on an imperfect population and, opposite, real-world evidence means imperfect data on a perfect population. The reason is clear – we have some inclusion/exclusion criteria which basically exclude from clinical trials a quite frail part of the population, so then the clinical trial doesn’t reflect the real situation. That probably translates to not such good results in the real world than they’re supposed to be, based on our results from clinical trials.
AC: Yes, I completely agree. I would say what you’ve brought up is there’s strong, what we call in research, internal validity but limited external generalisability because of the cherry-picked nature of the patient population. I think some of the other issues that distinguish clinical trials versus real world are clinical trials are often enriched for biochemical progression, there is rigorous testing. So, on the one hand, biochemical progressors may have a better time on therapy but, on the other hand, in clinical trials you’re doing repeated myeloma laboratories regularly and imaging which may not be done. I think the other features to think about are academic versus community, the ability to keep people on therapy. Perhaps what we’re seeing more and more with a drug like carfilzomib which is giving great outcomes in many clinical trials but we’re also seeing that there are some unique toxicities that may differ based on age. Depending on which clinical trial and patient population you look at, the risk-benefit ratio changes. So we’re seeing more and more the importance of this real-world data. Both in Europe and the US there is interest in looking at real-world and can you match what’s happening in clinical trials in the real world and, if not, how do we narrow that gap. So a lot of exciting work across all malignancies.
We have four abstracts to discuss today and it will be interesting to see our different takes on all of these but let me start with the first one. My colleagues and I presented this at EHA which is basically the recent treatment patterns, healthcare utilisation and costs in heavily pre-treated relapsed or refractory myeloma in the US. So the unmet need in myeloma keeps changing. We’ve had all these drugs approved recently so there is incremental benefit but the current unmet medical need really is this triple class exposure and refractory. So we’re talking about proteasome inhibitors, immunomodulatory drugs and CD38. So when patients become triple class refractory what are the outcomes of these patients in the real world and some groups have already looked and published on that but this was also to look at the healthcare utilisation and costs which has had much less data published.
In this study looking at about 14,000 real world patients we narrowed down those that had had all three classes of drugs which ended up being 154 patients and the median follow-up was seven months. What we found was that the median duration of therapy was relatively short – 4.2 months and the time to next therapy was 6.2 months. So relatively modest numbers but, again, recognising this is not unexpected because if you’ve been triple class exposed in the real world and then superimposing, as we said, real world patient comorbidities and access to treatments and ability to stay, then we looked at also the mean per member cost. So first we looked at hospitalisations, which was 0.18 per month. Then also the number of days was 1.44 and when it came to costs it was about 37,000 per month per member. When you divide that into inpatient and outpatient, inpatient was approximately 8,000 and outpatient about 20,000. So the reason also this is important is we have a lot of exciting new therapies that are being studied in this population, including bispecific antibodies as well as CAR T. So the question is what might those costs be and if you’re getting preliminary data for CAR T of a PFS of around a year we’re looking at about $480,000 for this based on what we’re currently spending. So it puts the costs that are being charged into context. So, Roman, I’m curious – what are your thoughts on this? Does this seem reasonable in terms of what the costs were obtained and the time on therapy? What do you think?
RH: I’m generally disappointed with how outcomes if we started to treat with any triplet, any expensive therapy, whatever we have, the patients in the setting like triple refractory, penta-refractory. The outcome up to now, I think, is very dismal. There are limited intervals of progression free survival from one to four or five months. Okay, we can see some very promising results with some CAR T-cells, BCMA especially, but it’s still time to optimise this therapy. So, to me, to be too focussed to treat patients penta-refractory with very expensive triplets so far is not effective. So that’s the reason why we have to be very, very aggressive at the beginning to try to cure our patients. Because the cost-effectiveness in that setting is really low.
AC: Yes, I think your point is really well taken. Perhaps one of the cleanest papers on this topic is from the UK, which presented that if you look at in the UK the median duration of therapy from initial therapy at each successive relapse it’s diminishing returns. By the time you get to your third, fourth, fifth, you’re talking about just several months. So your point is really important that our best outcomes seem to come with newly diagnosed first relapse and then we lose the steam as it goes. Also there’s the issue of attrition, that not every patient is even getting to the next line of therapy because the overall survival curves are not flat so we’re losing patients along the way. The purpose of this paper was to look at this really heavily treated population and what are the costs that we’re currently spending and then that will help set the framework for new drug therapies because very few economies in the world will be able to tolerate the costs of CAR T at $500,000 per patient. The question is how do we use these new therapies effectively, both medically but also from a cost perspective. So I completely agree and clearly this highlights the need for not just improving outcomes but in a pragmatic, cost-effective fashion.
RH: Yes, having said that, I just want to emphasise that we have to stay optimistic because we couldn’t imagine five years ago that in the first relapse and second relapse we will be able to achieve 30-40% of complete remission, even MRD negativity, with PFS in standard risk patients beyond 50 months. So, okay, this is the research, this is development and we have to be optimistic but also realistic.
AC: Absolutely. I think, for example, if these CAR Ts move to early relapse and we’re able to match this PFS with a single intervention, that’s really great for patients and the economy. So that will be great for everyone. Maybe we can now move to the second abstract which I’ll let you lead the discussion.
RH: Actually I will introduce an abstract presented by Heinz Ludwig which is focussed on an analysis of ixazomib treatment combined with lenalidomide and dexamethasone, basically, based on the treatment identical to the original phase III trial TOURMALINE-MM1. This retrospective analysis of ixazomib was made in Europe and it was quite a robust analysis, I think the biggest analysis so far with this rapid therapeutic treatment in the real world. It was done on 359 patients, as I can see, and I have two comments, basically. We already made some analysis in consortium with our Greek colleagues and UK colleagues before and then our own analysis on just check data from our registry. We basically achieved similar outcomes as in the original TOURMALINE-MM1 trial, so our progression free survival in the first analysis was 24 months and in our analysis then 22 months, so basically identical. That analysis was not so good but still progression free survival achieved was 16 months which is relevant to a real-world, not selected, population. Moreover, if we compared some variables to the original MM1 trial, this population analysed in this work were definitely older, also the ECOG was similar and ISS stage 3 was higher – 65%. So I think the outcome of this analysis just confirmed that ixazomib lenalidomide dexamethasone is a very suitable and effective treatment for even an older population. Toxicity is really acceptable, common, not surprising. Just this analysis confirmed what we saw in our analysis, so that’s perfect. It’s working in the real world, which is important.
AC: I completely agree. We talked about, at the beginning, some of the limitations of real world research which is that, in general, obviously the purpose of clinical research is to look at the outcomes – best response, PFS and OS – but we all know that those endpoints are dependent on patient factors, disease factors and treatment factors. So all it takes is to over- accrue a clinical trial with older patients with cardiac comorbidities, with high ISS stage and high risk disease and your outcomes are going to be very different than fit patients, young with favourable disease. So that’s why we’re always taught to avoid cross-study comparison and a perfect example of this is when you look at IRd, the Rd control arm of TOURMALINE, the PFS of that is about 14 months where in some of the other Rd backbone studies it can be as high as 17 months. So there’s three months’ difference with just the control arm of Rd and speaking to the patient disease and treatment factors that can affect outcomes. The purpose of this is not necessarily to do the direct cross-study comparisons but just to give a ballpark estimate of what the outcomes might be. As you alluded to, what these studies are showing is we talk about this efficacy versus effectiveness. Efficacy is determined from clinical trials, effectiveness is real world. What we’re looking at in a lot of these studies is what is the gap, because we’ve said that effectiveness is always going to be better than efficacy. Sorry, efficacy in the clinical trials is always better than real-world efficacy and so then the challenge is how do we narrow that gap. Part of it is to actually give regimens that are convenient, well-tolerated and administrable to progression. Because study after study is showing that when we give regimens the gap between effectiveness and efficacy is much higher. So the gap seems to be relatively narrow, really highlighting the need for real world data as well as effective, tolerable regimens.
RH: I absolutely agree. I have to say that we were really surprised to see that we are able to reproduce data basically on the same level in terms of progression free survival as in clinical trials. In our database in this trial it was 16 months, it’s still good. Before that we analysed seven or eight drug combinations in the real world from our database and we were never seeing such close outcomes between the clinical trials and real-world data which is just a reflection that this regimen is really suitable.
AC: You’ve done a lot of great work on this real world evidence and I forgot to mention we actually recently published in Expert Review of Haematology. We compared VRd, KRd and IRd in the real world. The challenge with doing that analysis, of course, is that in the real world there are clearly going to be biases and we saw this, that typically KRd patients were younger, they tended to have more symptomatic disease, they had more high risk disease and so you can’t just compare the raw numbers. But when you do Cox proportional-hazards modelling and adjust for those baseline differences what was striking is that the duration of therapy was actually relatively identical in the three arms. It highlights how you might think that one particular regimen is a real standout, so it goes back to that efficacy/effectiveness gap and how that applies to each regimen. So really cool work being done in these spaces. Do you want to maybe take us to the next abstract as well?
RH: Okay, I can. This abstract basically analyses HUMANS, originally it analysed the Nordic countries’ treatment pattern and healthcare resources utilisation. Basically they merged data from Denmark, Sweden and Finland. To me, it’s a nicely done analysis but at the end of the story they just confirmed what we know in Europe. We usually have to follow the guidelines. So this is no surprise that for newly diagnosed patients at the time it was bortezomib based therapy, the most frequent treatment; the same story for lenalidomide treatment in the relapsed setting. But to me it was interesting that lenalidomide-based treatment was used more frequently in Finland, for instance. I can’t say more, it’s a nice analysis saying, okay, we are treating basically the patients according to guidelines, national ones but generally European ones, based on the drugs which are reimbursed and approved in Europe. So nicely done, a robust analysis – more than 7,000 patients – so it’s a really representative analysis.
AC: I agree it validates what we think, that in Europe the regulatory bodies really determine not only what’s approved but what actually happens in the real world. Because if you don’t have access to drugs you can’t use them. So, to me, what’s interesting is really coming from the US perspective. If you look at myeloma regimens as patients go from first to second to third line of therapy it’s almost like artwork because it’s abstract art. There’s nothing standardisable, it’s just an abstract work of art. It’s like an impressionist painting and you can zoom in and maybe see a line of particular logic and then you zoom out and you lose that. It’s really one of the global variations and that’s why we’re both involved in INSIGHT, for example, this non-interventional global study. It’s the same thing, the minute you throw the US in there it further adds to the chaos. But even if you take the US out, globally there is tremendous variability in who gets what frontline. In light of the first abstract with the cost discussion that’s going to be important because if we want to make sure that patients all over the world are benefitting from novel therapies we have to make it practical. To me this raises the question, one of the striking questions is in the US it’s almost impossible to be first relapse without becoming lenalidomide refractory because we use induction, whether you’re transplant eligible or not. If you’re transplant ineligible you’re going to get an R-based regimen and be progressing and if you’re transplant eligible you go to transplant and get lenalidomide maintenance. So it’s very hard in the US to not be lenalidomide refractory. So when we look at all these Rd backbone studies, again talking about extrapolating from studies to real world, at our centre we pretty much had to close all the phase III studies with an Rd backbone because we don’t have equipoise. It would be unethical to randomise somebody who is lenalidomide refractory to get more lenalidomide. So these kinds of global practice patterns really affect the study and enrolment and outcomes. We’re seeing more that the impact of becoming lenalidomide refractory at first relapse does really impact the outcomes. A striking example is DBd in CASTOR, the PFS is almost 16 months for all patients but when you look at lenalidomide refractory it’s nine months. It’s the same study, same regimen but this is the impact. So this really highlights the need to make sure that when we do these global studies you’re considering these patient factors when we look at the outcomes.
RH: I absolutely agree.
AC: So then the last abstract that we can discuss is from Spain. Dr Escalante unfortunately couldn’t join us today but he presented on behalf of the Spanish group the myeloma frailty population in Spain. This is a preliminary analysis of the INSIGHT registry. As we said, the INSIGHT is a non-interventional global study. Here what they were looking at specifically in the 235 Spanish patients approximately, half newly diagnosed, half relapsed, they looked at various frailty status that was reported by the physician, the official myeloma frailty index and Charlson comorbidities. What you can see in this abstract is that, from a physician reporting perspective, there was only about 8% of the whole cohort when you just combine newly diagnosed and relapsed/refractory, only about 8% frail whereas when they did it by the frailty index it was almost 30%. So we see a big difference. Then the comorbidity index, as you might expect, is going to be distributed from approximately 78-80% zero to one, about 17% for two to three and then very few four and beyond. Certainly one of the limitations of this analysis is that the official myeloma frailty index was missing in almost half the patients when it was done but even in those that were done you can see already there were approximately 24 patients that were frail by the frailty index whereas there were only about 17 patients by physician.
So my take on this abstract is, first of all, that disparity is quite interesting. It will be interesting to see when we just have the descriptive analyses but I’d be curious to look at the outcomes. So how does the physician reported frailty compare to the calculated myeloma frailty index? Are they concordant, discordant, reasonably concordant? If not that’s the kind of data that we need to change practice patterns. Because we keep coming up with more and more prognostic indices and more things but at the end of the day as physicians, as we are always taught in medical school, if you’re going to do a test you should be prepared to answer how is it going to change your management. So get another frailty index, okay, that’s nice but why should I know about this? It’s nice to know the prognostic but is there a predictive component to that? So, should I be using different regimens, different doses for frail? That’s what we really need to start getting to. I would just conclude by a myeloma clinic visit is already complicated because you have look at the patient’s symptoms, their medications, all the myeloma laboratories, the imaging, the bone marrow and by the time you integrate all this data do you have time to do any complex extra testing?
What I’m really looking forward to seeing more, one of our Fellows has written an investigator initiated study where she’s incorporated gait speed. Which is great, because as you ask the patient to come from the waiting room to the office it’s just a quick walk speed from here to here. If that can replace all of this that would be great, right? Because it’s another data point, it will almost be like the new vital sign. My question would be if somebody has a slow gait speed and you’re using an IMiD, for example, might you need to give more anticoagulation than aspirin. So I want something simple like that that would help me not only take better care of patients but know that this management will be affected by this frailty type of calculation. Roman, what’s your experience with frailty and how does it incorporate in your practice?
RH: Actually, if I return to this abstract, I recall what you said, I think the most interesting is to see the different sensitivity of different indices to indicate frailty of patients which is nicely done and analysed in this abstract. On the other hand, if you have an auditorium and if I give a question who is really using real life in outpatient clinics, any index or any scores from this, if it’s myeloma frailty or Charlson comorbidity, there is such a limited number of physicians who are really using it. Because it’s sometimes time-consuming so everybody is busy, a lot of patients, in some countries limited staff who will be able to do it because this is not typically work for physicians, of course. So this is something which I will be more than happy to be overcome. Because to be focussed on trials dedicated to this frailty status, this is important. I think we usually really exclude these patients from clinical trials which is not fair and that’s one of the reasons why in the real world then the data are not so good. But probably really maybe just designing trials for typically such patients without any inclusion/exclusion or other criteria, just frailty patients, that would be nice to see some optimisation of the drug. I think it’s coming – the French guys are really doing nice work in that setting, focussed on elderly frail patients. It’s coming, many other teams, Italy, and also they are focussed on something. Sometimes it’s really just simply fine-tuning is important for us, like the reduction of dexamethasone dose. It’s really important and it has completely changed the scenario about infections, tolerability and so on in such types of patients. So it’s important to have, of course, strong therapy for younger patients in good condition without any comorbidities but then we need really, and I think it’s even more difficult to find, really very good therapies for such types of patients as our elderly generation above 80 with a lot of comorbidities. It’s possible but it’s more difficult. It’s not so simplified, like using one therapy for everybody. That’s not working.
AC: Yes, you reminded me, if you think about how are we going to improve myeloma patient outcomes the converse of that is who is lowering the overall survival curve for myeloma? It’s frail, elderly, high risk, renal failure, extramedullary disease, functional relapse. But given that this is a disease of the elderly, particularly the frail elderly, we need to do a better job, as you said beautifully. This is the population that is under-represented in clinical trials, that is affecting real world outcomes, that’s not able to get to a clinical trial. I would like to see to the point where maybe in every phase I, II and III study there’s at least a planned subgroup of elderly, frail patients so that we learn, even in phase I studies, what if older patients, we know that they have impaired hepatic and renal function and are more prone to thromboses and cardiac issues, shouldn’t we have a dedicated phase I component to these older patients? It would be nice to get to that.
The last point I was going to make is you reminded me also this year at the ASCO meeting there was actually the data from prospective risk adapted treatment in oncology patients based on frailty index. They showed, actually, by modulating the chemo they were able to reduce the toxicity and then they had better outcomes. And that’s the future is really risk-adapted therapy and I hope we can get there but with something that’s quick and easy. It can’t be a long 20 minute, 30 minute survey because that’s your entire clinic visit right there.
So any concluding comments on what we’ve talked about today, real-world data? What would your recommendation be to community doctors and cooperative groups, industry?
RH: I think the one message for the audience basically is that it’s nothing against clinical trial phase III, we need these trials. We even more need some meta-analyses confirming just the outcomes from one trial. But then, and it’s desperately needed and that’s the reason why we spent some time with this discussion, is we desperately need also confirmation from the real world. It is important.
AC: Yes, I completely agree. The themes that we’ve talked about today, for the community oncologists particularly who may not be just seeing myeloma, these themes are across all of oncology. The initiatives being done to narrow this efficacy real-world gap so efficacy from the clinical trials which are intended for registration but the effectiveness in the real world, that gap really needs to be narrowed across the spectrum and it starts with making studies more representative, the eligibility criteria permissive and doing treatments that are actually implementable in the real world. So we’ll look forward to narrowing that gap with time.
So, thank you Roman. It’s been a pleasure and we hope the audience has found this helpful. Thank you for your attention.