The case of a 51-year-old with multiple myeloma, standard risk, stage II was the topic of discussion during a recent Targeted Oncology Case-Based Roundtable moderated by Ajai Chari, MD.
Ajai Chari, MD, associate professor Medicine, Hematology, and Medical Oncology, The Mount Sinai Hospital, discussed the case of a 51-year-old patient with multiple myeloma during a Targeted OncologyTM Case-Based Roundtable.
Targeted OncologyTM: What are the data for using autologous stem cell transplant (ASCT) in patients such as this?
CHARI: Some brief data from the IFM/DFCI2009 study [NCT01191060] that had VRd [lenalidomide (Revlimid), bortezomib (Velcade), dexamethasone] for 3 cycles, after which they got ASCT in the consolidation phase or got more VRd, then lenalidomide maintenance for a year. These are the French data. In the United States [DFCI (Dana-Farber Cancer Institute)] version, the lenalidomide is given until progression of disease [PD].1,2
[After 44 months of follow-up], the control arm without ASCT had a 36-month median PFS [progression-free survival] vs 50 months for the ASCT arm, so an extra 14 months. After 90 months of follow-up, the PFS didn’t change much, and the OS [overall survival] rate is not different [either], at 60.2% without ASCT vs 62.2% with ASCT.1,2 This is important because people make their decisions based on OS.
I would say that, although we always look for OS end points, our outcomes in myeloma are getting so good. If you’re going to wait for an OS end point to make treatment decisions, you’ll be waiting a long time because this was with a long follow-up of 90 months.
In 90 months, there wasn’t an OS difference, despite a significant PFS difference, which speaks to the impact of salvaged therapies. Although we can risk stratify multiple myeloma, there are high-risk patients who relapse and might be able to demonstrate OS outcomes, but not for the typical patients with multiple myeloma. The FDA routinely accepts PFS as a surrogate end point, particularly for less heavily treated patients and those with newly diagnosed multiple myeloma [NDMM], because of the prohibitive duration of follow-up. Sometimes people talk about molecular data and whether somebody should get a particular treatment. In the IFM/DFCI2009 study, relative to the hazard ratio, ASCT was better in all patients. Although [there was] a wide confidence [interval in those with high-risk cytogenetics], which is not surprising because they had [only approximately] 40 patients.
I don’t think ASCT harms patients. But [to know] whether it overcomes patients, you need well-designed prospective studies with an adequate number of patients.
There’s a lot of debate about the use of MRD [minimal residual disease] status, especially coming from chronic myeloid leukemia, where you have these deep molecular remissions, and you may think this is really practice-changing. The ASCT arm always does better than the arm without ASCT.
Whether you’re looking at standard-risk or high-risk [cytogenetics] based on transplant, standard-risk seems to do better.2 That’s an important update.
What do you consider the standard-of-care regimen for induction?
In speaking to the potential utility of MRD, the FORTE study [NCT02203643] from the Italian group is a very important study. They used a 1:1:1 randomization. There were several questions being asked, [such as] what the best induction regimen was. It was 4 cycles of KCd [carfilzomib (Kyprolis), cyclophosphamide, dexamethasone] for 1 arm vs 4 cycles of KRd [carfilzomib, lenalidomide, dexamethasone] for 2 arms.3
All 3 arms got mobilized, 2 arms got ASCT, and 1 KRd arm did not get ASCT, but went on with 8 more KRd cycles. The dosing of carfilzomib is 36 mg/m2 twice weekly. Keep that in mind because many of you [probably] do weekly chemotherapy, but here it is twice weekly for an entire year. That’s a pretty dose-intensive regimen, and we need to think how many patients, physicians, and nurses would adhere to a year of chemotherapy twice weekly. The other 2 arms received ASCT followed by KCd or KRd consolidation, and the second 1:1 randomization was to either KR or R alone maintenance.
Francesca Gay, MD, PhD, presented some updated [data at the American Society of Hematology 2020 meeting and later published in Blood].4 Interestingly, after a 1-year follow-up, there was no big difference between the ASCT and the non–ASCT arms. This is one of the limitations of MRD because it is a surrogate, but OS and PFS can also be difficult, and MRD doesn’t preclude you from needing to follow up patients with these survival-based end points that are being used as surrogates.
Based on median PFS, the winner—by far—is the KRd plus ASCT, then the arm with 12 cycles of KRd [KRd12], then KCd plus ASCT. KRd12 without ASCT did better than KCd.4 This is an important take-home message because it confirms the older IFM2013-04 study [NCT01971658] that VTd [bortezomib, thalidomide (Thalomid), dexamethasone] was superior to VCd [bortezomib, cyclophosphamide, dexamethasone] by efficacy. It seems a better partner for a PI [proteasome inhibitor] is [an immunomodulatory drug (IMiD)], whether in Europe you use thalidomide [or] in the United States you use lenalidomide.
It seems like the IMiD plus PI combination is better. I still think there may be slight exceptions for somebody who is admitted in the hospital with renal failure or cord compression, and you need to start treatment immediately. You may have some issues with getting [an IMiD]. In the real world, just because you start somebody on a [cyclophosphamide]-based regimen, doesn’t mean with the next cycle you couldn’t switch, but in the European policy, they’re pretty much locked into it.
[The] first message is [that] the KRd is superior. The second message is [that], although MRD [initially] suggested relatively comparable results—with the notable exception that even early on there would seem to be a benefit for high-risk patients going to transplant— with a longer 45-month follow-up, ASCT with KRd did the best and was statistically significantly better than the non–ASCT group.
Pretty much everybody benefited from ASCT, so KRd with ASCT is favored, and the HRs are well below 1 [across the board], with a few crossing the 1 threshold because, again, smaller numbers. The message here is [that] if KRd 36 mg/m2 twice weekly for an entire year did not beat ASCT, then that gives us pause as to the utility of MRD as a surrogate for decision-making and the importance of how you get to a particular remission.4
Now we have both the French data for VRd with ASCT with a 90-month follow-up and KRd with a 45-month follow-up, and both studies show that, even in 2021, ASCT has not gone away. If you want to displace ASCT, I argue that you need to show that novel regimens are doing better than it. Not only have we not seen that they’re not doing better but they are also not keeping up with it, so we still have work to do. We’ll see what the newer regimens are, but we should still be considering ASCT, [especially] for patients below age 70.
One of the tricky parts is, for patients over age 70 or 75, you need to reduce the melphalan [Alkeran] dose typically to condition for the ASCT. Then you are in a [murkier] situation.
The second randomization was the KR vs R, and the KR [seemingly] did better across all subgroups, showing the utility of doublet maintenance therapy.
The carfilzomib dosing for the maintenance is 36 mg/m2 on day 1, 2, 15, and 16. So it’s still 4 visits per month, which is a big ask, but it is an option for patients.
Can you discuss trials looking at quadruplet therapy?
In the CASSIOPEIA study [NCT02541383], everybody was transplant eligible. All patients received 4 cycles of VTd or VTd plus daratumumab [Darzalex; D-VTd], stem cell collection, ASCT, then 2 cycles of consolidation with VTd or D-VTd.5
There was an important difference in the United States version of this study. In the CASSIOPEIA study, there’s a possibility a patient could get daratumumab up front in induction and consolidation but not get it in maintenance. In the United States version, it is daratumumab throughout, and there is no double randomization. Another difference is, in this study, the daratumumab maintenance started off every 8 weeks. That’s important because we now have pharmacokinetic data, and we all know daratumumab is a monoclonal antibody, but the trough level is quite below where we think efficacy is maintained. In the United States study, daratumumab also started off at 8 weeks, then it was modified to go to monthly for every-4-week dosing of maintenance. Those are the 2 major differences other than [lenalidomide use in the United States one].
The CASSIOPEIA study was a large study with 1000 patients that had 2-part end points. One randomization before induction, and [the second randomization before maintenance].
What was the efficacy in the CASSIOPEIA trial?
The [first] primary end point was stringent complete response [sCR], and that was favorable. It was 29% for the quadruplet arm vs 20% for the triplet arm, which was statistically significant [P ≤ .001].
MRD negativity status was 64% in the quadruplet arm vs 44% in the triplet arm, which was also statistically significant [P < .0001]. The take-home message is [that] we currently have deeper responses with the addition of daratumumab. Everybody seemed to benefit in [the subgroup analysis], except the high-risk patients.
In the high-risk population, there was a difference between improving, overcoming, and hurting. The addition of daratumumab didn’t overcome high-risk, but it also did not hurt those patients. The high-risk group remains an unmet need, and we still need to do better. With continuing therapy, both arms have a deepening of response, but the quadruplet arm always seems to do better.
Whether you look at MRD using next-generation flow cytometry or next-generation sequencing, D-VTd did better than the VTd.
Perhaps the most important thing from this study is the Kaplan-Meier curve after a 2-year follow- up. There was already a median PFS separation favoring the quadruplet over the triplet therapy [HR, 0.47; 95% CI, 0.33-0.67; P < .0001]. That’s an important message because it is not just sCR and MRD that are the primary end points, but it translates into PFS benefit, which, barring OS, is the most important end point.
The 2021 American Society of Clinical Oncology meeting had an update about the second randomization in all-comers.6 As you might suspect, many people would criticize that the control arm was observation, and if it is already established that lenalidomide is a better maintenance than placebo, then why are we doing this? Keep in mind that, when studies are designed, there’s always a lag in what we learn. In this study, the observation arm was the control. [At a median follow-up of 35.4 months, median PFS was not reached with daratumumab and was 46.7 months with observation [HR, 0.53; 95% CI, 0.42-0.68; P < .0001], so almost a 50% reduction in the risk of PD or death with the addition of daratumumab on maintenance.6
This was consistent across most subgroups. However, one notable exception is that this only appeared in the patients who got VTd in induction and consolidation. If you got daratumumab up front, there didn’t seem to be any benefit of getting it as maintenance again [HR, 1.02; 95% CI, 0.71-1.47], but if you got only VTd, there was a benefit [HR, 0.32; 95% CI, 0.23-0.46].6
The message here is, if you didn’t get daratumumab up front, it’s a good maintenance, but it doesn’t answer the question of how it compares with lenalidomide maintenance. The second [unanswered] question is: If you had been giving daratumumab as intended, which is every 4 weeks, how it might have worked for those who had it up front?
What was the version of the quadruplet regimen investigated in the United States?
The United States randomized phase 2 trial, the GRIFFIN study [NCT02874742], had transplant-eligible patients with NDMM. [Randomization was 1:1 with 4 cycles of] VRd, ASCT, VRd consolidation, then R maintenance in the control arm and [4 cycles of D-VRd, ASCT, D-VRd consolidation, and DR maintenance] in the experimental arm. There’s no double randomization. [Initially], the daratumumab [maintenance] was every 8 weeks, but then it moved to every 4 weeks.7
Another consideration that was different from the European version was that they used [cyclophosphamide] for mobilization, whereas they used granulocyte colony-stimulating factor [in the GRIFFIN study]. If that failed to get adequate cells, they could go back and use cyclophosphamide, but it was not permitted as a mobilization strategy from the get-go. The median age in the GRIFFIN study was 60, ISS III [was approximately] 13%, and high-risk was 15%, which was comparable with the CASSIOPEIA study. [The issue in the United States] randomized studies is adherence. In the D-VRd arm, 90% of those patients got ASCT, whereas [only 76% did] in the VRd arm.
If you did an intention-to-treat analysis vs actual treatment, you might get different results. The primary end point of sCR after consolidation was statistically significant at 63.8% in D-VRd vs 47.4% in VRd [P = .0253] and CR or better was [approximately] 82% vs 61% [P = .0014]. This is after the induction, ASCT, and consolidation. There is a deepening of responses in both but, again, the depth is maintained with the addition of daratumumab.8
[In the subgroup analyses] for sCR, most are in favor of D-VRd, with the exceptions of ISS III and high risk. But again, the sample sizes for them were exceedingly small. Remember this is a randomized phase 2 study with much fewer patients than CASSIOPEIA. ISS III only had [approximately] 13 patients in each arm, and the same for high risk, so you have a wide confidence interval.
For MRD negativity status, everything is favoring D-VRd, too. In the ISS III [subgroup], you have more MRD negativity [favoring D-VRd], and in the high-risk [subgroup, too], but it also still has a wider confidence interval. Overall, MRD negativity was 62.5% for the quadruplet therapy vs 27.2% [for the triplet therapy].
OS is not surprising. The VRd arm is a great regimen. Many of you are using that and have been using that for the past decade for good reason. If you were an early adopter of VRd, you were doing that based on a single-arm, phase 2 study with [approximately] 60 patients. The randomized studies did not report until maybe a few years ago.
We don’t have long enough data, and I’m not surprised we’re not seeing OS difference with a great triplet regimen to begin with. For PFS, there seems to be a potential separation, but it’s premature to say anything. There were very few patients in that follow-up. There are [approximately] 7 patients who are still relapse free vs the VRd. We clearly need more follow-up. We saw from the FORTE [study], and all those other studies, that you need more follow-up to see if these translate. If you believe in MRD, then there [was favoring of the quadruplet therapy] pretty much across all subgroups.
[For me,] the one use of MRD is as a surrogate when PFS is prohibitively long. My crystal ball would say that if the MRD negativity is deeper with the addition of daratumumab [across the board], this should pan out through PFS benefit with more time.
How did adverse events (AEs) affect the patients on the GRIFFIN trial?
For safety, we know daratumumab and isatuximab [Sarclisa] targets CD38, which is expressed on myeloid precursor cells. So we do see neutropenia of all grades at 58% for D-VRd vs 35% for VRd, and grade 3 and higher [neutropenia] is 41% vs 21%. Thrombocytopenia is also higher in D-VRd.7,8 A lot of the [numbers on] AE tables— which I find challenging—are always absolute and never adjusted for time on therapy. Although it’s less of an issue here because the follow-up is relatively short.
[Of course,] the question is [whether] the cytopenia translated into any significant clinical problems. From an infection point of view, upper respiratory tract infection [URTI] of grades 3 and 4 are not high grade, so those were not different. All grades of URTI were 63% for D-VRd vs 44% for VRd.8
Of course, subcutaneous daratumumab has very little infusion-related reactions. It started off with IV [intravenous], and later in the study during maintenance, we were allowed to switch. Initially, all grades of infusion reactions were 42% because it was IV. Typically, with subcutaneous daratumumab, we’re getting around 10% for all grades and very few for grade 3 and 4.
What are your thoughts on the trials you have mentioned overall?
The IFM/DFCI2009 study and the CASSIOPEIA were phase 3, and the others were phase 2. Median follow-ups were remarkably different for all, and for the CASSIOPEIA and GRIFFIN [studies], they were relatively short.
One of the most important things we don’t talk about enough is how much chemotherapy was given. For the IFM/DFCI2009 study, it was about 4 cycles of induction to post consolidation. The control arm VRd had a total of 6 cycles lasting [approximately] 6 months. The FORTE study had 8 and 12, CASSIOPEIA [had] 6, and GRIFFIN [had] 4.5.
All studies, apart from the GRIFFIN study, use cyclophosphamide mobilization. Because the end points are very different, if you look at post consolidation very good partial response, usually the experimental arm does better. It was 78% vs 69% [in the IFM/DFCI2009 study], but the FORTE [study was] quite comparable at 89% vs 87%. In CASSIOPEIA, it was 83% vs 78%, and in the GRIFFIN study, it was highest at 91% vs 73%.
For the PFS differences, the data are not mature enough. But all 3 studies, apart from the GRIFFIN study, show an improvement. VRd plus ASCT was superior by PFS. KRd plus ASCT was superior. D-VTd was superior as well.