ASH 2022: Lymphomas and CLL

CME

Key Studies in Lymphomas and CLL: Independent Conference Coverage of ASH 2022

Physicians: Maximum of 1.00 AMA PRA Category 1 Credit

Released: April 17, 2023

Expiration: April 16, 2024

John M. Burke
John M. Burke, MD
Peter Martin
Peter Martin, MD

Activity

Progress
1
Course Completed

MATRix/IELSG43: Phase III Trial of High-Dose Chemotherapy + ASCT vs Nonmyeloablative CIT as Consolidation for Primary CNS Lymphoma

John M. Burke, MD:
In the randomized IELSG32 trial, the combination of methotrexate/cytarabine/rituximab/thiotepa (MATRix) was shown to improve outcomes compared with methotrexate and cytarabine alone in primary central nervous system lymphoma (PCNSL), a highly aggressive form of NHL.30,31 High-dose chemotherapy followed by ASCT (HD-ASCT) also has been shown to be effective and feasible in younger patients (younger than 65 years of age) with PCNSL.32

At ASH 2022, Illerhaus and colleagues33 presented results from the multicenter, open-label phase III IELSG43 study (primarily conducted in Europe) evaluating HD-ASCT after induction therapy with the MATRix regimen in younger patients with PCNSL. The MATRix induction regimen was administered for 4 cycles, after which patients were randomized to receive 1 of 2 consolidation regimens: either HD-ASCT with thiotepa and carmustine or busulfan (in blue) or the rituximab/dexamethasone/etoposide/ifosfamide/carboplatin (R DeVIC) regimen (in orange) for 2 cycles. Patients were recruited between 2014 and 2019.  

MATRix/IELSG43: Baseline Characteristics (ITT)

John M. Burke, MD:
Of the 346 patients enrolled, 116 did not make it to randomization for various reasons, including declining to be randomized, AEs, and others. The baseline characteristics of the remaining 229 randomized patients were fairly well balanced, with a slight increase in poor Eastern Cooperative Oncology Group performance status in patients who received HD-ASCT.

MATRix/IELSG43: Efficacy Outcomes

John M. Burke, MD:
As would be expected, response rates were the same in both groups after the 4 cycles of MATRix induction and before consolidation, with 40% of patients achieving CR and 60% achieving PR. Even after the 2 treatments during consolidation, there was no difference in response rates between R DeVIC and HD-ASCT, with approximately two thirds of patients achieving CR.  

Despite the lack of a difference in response rates between the 2 arms, PFS and OS were improved in patients who received HD-ASCT vs R-DeVIC. At a median follow-up of 45.3 months, the 3 year PFS rates with HD-ASCT and R-DeVIC were 79% vs 53% (HR: 0.405; 95% CI: 0.252-0.650; P = .0002), respectively, and the 3 year OS rates were 86% vs 71% (HR: 0.456; 95% CI: 0.256-0.812; P = .0077), respectively. Although not shown here, all examined subgroups in the study benefited with HD-ASCT vs R-DeVIC.

MATRix/IELSG43: Safety


John M. Burke, MD:
There were high rates of hematologic toxicities in both groups, with slightly higher rates with HD-ASCT. Rates of infections and mucositis also were higher with HD-ASCT. There was no difference in neurotoxicity between the groups.

Of note, treatment related mortality was higher with HD-ASCT at 3%, whereas with R-DeVIC it was 0%.

MATRix/IELSG43: Clinical Implications

John M. Burke, MD:
The IELSG43 trial is, to my knowledge, the largest phase III trial of HD-ASCT in patients with PCNSL and the first to prove the benefit of HD-ASCT in first remission for patients with PCNSL. 

In my practice, I do not routinely use the full MATRix induction regimen, but rather use an alternative induction regimen. Furthermore, I already have been sending patients with PCNSL who achieve remission after induction for ASCT. So for me, this trial just confirms what I already have been doing. 

I do find it interesting that the 3-year PFS and OS rates were better with HD-ASCT vs R-DeVIC, despite similar remission and response rates immediately after consolidation. 

Peter Martin, MD:
I agree. In general with PCNSL, the more intense the treatment a patient can tolerate, the better their outcomes. That was the case with the MATRix regimen, clearly built on just high dose methotrexate, and it seems to be the case with high dose chemotherapy, as well. The recent CALGB 51101 (Alliance) trial suggested a benefit for ASCT consolidation compared with nontransplant consolidation, but that trial was even more plagued by the challenge of getting patients to consolidation.34 Even though it is clear that we should be trying to increase the intensity of therapy for these patients to improve outcomes, there will be a limit where eventually toxicity outweighs gains in efficacy. At some point, other innovative ways will be needed to deal with PCNSL. 

John M. Burke, MD:
The OptiMATe trial (NCT04931368), which currently is recruiting, will be investigating a de-escalated induction regimen with the aim of reducing toxicity, and we look forward to these results.

Pooled Analysis of PhaseED-Seq MRD-Negativity Assessment as Surrogate Endpoint After Frontline Induction Therapy in Patients With DLBCL

John M. Burke, MD:
The current approach to monitoring for relapse in patients with DLBCL is to use PET, with a PET CR being associated with improved PFS and OS and a favorable prognosis.35 The next study we will discuss is aiming to introduce end-of-treatment MRD testing into the DLBCL space for a similar purpose. 

Detecting MRD through measuring cell-free circulating tumor DNA (ctDNA) with DNA sequencing has been proven useful for predicting the risk of treatment failure in DLBCL.36 However, current methods of detecting ctDNA, such as clonoSEQ and CAPP-Seq, have some limitations and imperfections. PhaseED Seq is a technique that tracks combinations of variants on the same DNA strand—phased variants—rather than single nucleotide variants (SNVs) in the DNA, as is the case for current methods of detecting ctDNA.37 The tracking of phased variants leads to lower rates of “background noise” as compared with tracking of SNVs, thereby increasing the sensitivity of PhaseED-Seq, despite phased variants being much less common than SNVs. 

In a study presented by Roschewksi and colleagues38 at ASH 2022, PhaseED-Seq was applied to ctDNA samples from patients with aggressive B cell lymphoma enrolled on 5 prospective clinical trials, as listed on the left side of the slide, to determine the prognostic value of MRD status at various time points in the treatment of DLBCL. Pretreatment samples were used to identify phase variants, which then were tracked as MRD in samples during treatment on the first day of cycles 2, 3, and 4 and at end of therapy (EoT). A positive MRD was defined as a level >1 x 10-6. Finally, PhaseED-Seq ctDNA MRD results were compared with the performance of PET/CT scans in defining remission and in predicting relapse.

PhaseED-Seq for MRD-Negativity Assessment in DLBCL: Baseline Characteristics

John M. Burke, MD:
In total, 290 samples from 112 patients were included in this study. The median age of patients was 61 years, and there were slightly more male patients (61%) than female patients (39%). Most patients (79%) had DLBCL and advanced-stage disease (72%).

PhaseED-Seq for MRD-Negativity Assessment in DLBCL: Sensitivity of MRD Detection and PFS Prognosis

John M. Burke, MD:
I would like to highlight 5 important findings from this study. First, of 17 patients who ultimately progressed, 16 (94%) had detectable MRD at EoT. In addition, of those with detectable MRD at EoT, 94% relapsed with a lead time of up to 30 months. So, the test is good at predicting who is going to relapse.

Second, of patients who did not progress in follow-up, 91% had uMRD at EoT. In addition, of 70 patients with uMRD at EoT, 99% did not relapse in a median follow up of 17 months. So, the test is good at predicting who is not going to relapse.

Third, among eventual non-progressors, the rate of uMRD increased with additional cycles of therapy: 24% at Cycle 2, Day 1; 64% at Cycle 3, Day 1 and Cycle 4, Day 1; and 91% at EoT. So, the most accurate time to predict likelihood of cure is EoT.

Fourth, at EoT, MRD by PhasED-Seq was better than PET/CT at predicting relapse, and was highly predictive of relapse even in patients who were PET/CT-negative. Of the 62 patients who were PET/CT-negative, 11 had detectable MRD, two thirds of whom relapsed, whereas 51 had uMRD, only 1 of whom relapsed.

Finally, MRD testing by PhasEd-Seq appears superior to clonoSEQ and CAPP-Seq in a cross-trial comparison.

PhaseED-Seq for MRD-Negativity Assessment in DLBCL: Clinical Implications

John M. Burke, MD:
PhasED-Seq is not currently commercially available or ready for routine use in the clinic. In addition, larger, definitive, prospective studies still are needed to confirm these results. However, the implications of this study are far-reaching and have the potential to change the way we manage lymphoma in numerous ways.

First, patients who have uMRD after treatment may have such a low risk for relapse that perhaps they may not need to follow up with an oncologist or undergo surveillance scans. That statement needs to be proven, but one certainly can generate that hypothesis right now.

Second, patients who have detectable MRD after treatment may have such a high risk of relapse that they need either more frequent monitoring than is currently recommended or to have immediate or consolidation treatment to prevent near-certain relapse. Again, this needs to be tested, but it is a potential way this technology could be used.

Finally, patients who do not achieve early uMRD after just a few cycles of treatment may be selected for intensification strategies to increase their likelihood of long term cure.

Peter Martin, MD:
I fully agree. If the very reasonable hypotheses you posited pan out after testing, this ultrasensitive method to detect ctDNA MRD would allow us to identify patients who currently are either undergoing unnecessary surveillance or not undergoing sufficient surveillance, as well as identify those potentially missing out on early treatment that could improve outcomes.

ELM-2: Phase II Trial of Odronextamab in R/R DLBCL

John M. Burke, MD:
As a reminder, the phase I ELM-1 trial evaluated the safety and efficacy of the CD20xCD3 bispecific antibody odronextamab in patients with FL or DLBCL who received ≥2 lines of treatment, including an anti-CD20 antibody. In ELM-1, odronextamab demonstrated favorable outcomes in patients with DLBCL, with an ORR of 53% and a CR rate of 53%.17

Now, we will continue the discussion of the multicohort phase II ELM-2 trial evaluating odronextamab in patients with R/R B-cell NHL, which was introduced by Dr Martin earlier in the section on FL,18 but now with a focus on patients with R/R DLBCL who had received ≥2 lines of therapy including an anti-CD20 antibody and an alkylator.39 Again, to mitigate risk of CRS, odronextamab was given in an optimized step-up manner. For cycle 1, the study was initiated with 0.5 mg on Day 1 and 2, 10 mg on Day 8 and 9, and 160 mg on Day 15. This was modified to 0.2 mg on Day 1, 0.5 mg on Day 2, 2 mg on Day 8 and 9, and 10 mg on Day 15 and 16. In cycles 2-4, patients received 160 mg odronextamab on Days 1, 8, and 15, and then 320 mg every 2 weeks in cycle 5 onward. Treatment was continued until disease progression.

ELM-2: Baseline Characteristics of DLBCL Cohort

John M. Burke, MD:
The DLBCL cohort of ELM-2 enrolled 140 patients with a median age of 66 years who had received a median of 2 prior lines of therapy (range 2-8). Similar to other trials evaluating CD20xCD3 bispecific antibodies, this patient population was heavily pretreated: 57% had primary refractory disease, 86% were refractory to their last therapy, 79% were refractory to anti CD20 antibodies, 16% had received a prior ASCT transplant, and 22% had received a prior CAR T-cell therapy.

ELM-2: Efficacy in DLBCL Cohort

John M. Burke, MD:
With a median follow-up of 21 months, odronextamab achieved an ORR of 49%, with a CR rate of 31%. Similar response rates were achieved even in patients with prior CAR T cell therapy, with an ORR of 48% and a CR rate of 32%. This indicates that prior CAR T-cell therapy did not appear to affect response to odronextamab.  

The median DoR with odronextamab was 10 months, but unfortunately the median PFS was only 4 months.

ELM-2: Safety in DLBCL Cohort

John M. Burke, MD:
Notable toxicities with odronextamab included CRS in 53% of patients with the step up dosing and grade 3 CRS in only 1 patient. Immune effector cell‒associated neurotoxicity syndrome occurred in fewer than 5% of patients. Similar to what Dr Martin showed us earlier in FL, the step up dosing schedule does appear to reduce the rate of high grade CRS.

ELM-2: Clinical Implications in DLBCL Cohort

John M. Burke, MD:
Similar to what we saw earlier for FL, odronextamab efficacy appears to be similar to that reported for other bispecific antibodies in patients with R/R DLBCL. With glofitamab in this population, the ORR was 52%, the CR rate was 39%, the median DoR was 18 months, and the median PFS was 5 months.40 With epcoritamab, the ORR was 63%, the CR rate was 39%, the median DoR was 12 months, and the median PFS was 4 months.22 Both glofitamab and epcoritamab were accepted for Priority Review by the FDA for R/R LBCL based on these findings,21,41 and I anticipate that the findings of the ELM-1 and ELM-2 studies will be submitted for FDA approval this year.

Although clearly some patients benefit from bispecific antibody therapy, with a median PFS of <6 months across all bispecific antibodies, none is a “home run” in the treatment of R/R DLBCL. Regardless, I suspect bispecific antibodies will become relatively commonly used for patients who either relapse after CAR T-cell therapy or do not receive it at all. 

Peter Martin, MD:
I agree. These agents will be particularly important in a high risk population, and it will be great to have them available as soon as possible. Right now, they are available only through clinical trials, and there simply are not enough clinical trials to go around for all of these patients.