Summary of the Second Annual BMT CTN Myeloma Intergroup Workshop on Minimal Residual Disease and Immune Profiling

Abstract

The second annual Blood and Marrow Transplant Clinical Trials Network (BMT CTN) Myeloma Intergroup Workshop on Minimal Residual Disease and Immune Profiling was convened on December 7, 2017 at the American Society of Hematology (ASH) meeting. During this workshop, investigators from around the world presented their latest research involving assessment of minimal residual disease (MRD) and immune profiling (IP) in myeloma. This document summarizes the workshop presentations as well as relevant ASH abstracts and focuses on the regulatory issues involved in the integration of MRD and IP assessment in clinical trial design and practice.

Introduction:

The incorporation of therapies such as immunomodulatory drugs (IMiDs), proteasome inhibitors and monoclonal antibodies into the newly diagnosed and relapsed/refractory setting have led to higher response rates and improved survival for patients with multiple myeloma. In particular, achievement of minimal residual disease (MRD) negativity has been associated with improved survival outcomes.¹ Previously reported studies evaluating MRD have been heterogeneous with respect to the sensitivity of the MRD assay and the technique used, leading to international efforts to standardize MRD assessment.^2–4 While MRD status is now incorporated into the International Myeloma Working Group (IMWG) response criteria⁵ and it is evident that MRD negativity can serve as a prognostic biomarker, it is not clear if MRD can be used as a surrogate endpoint or to guide treatment decisions. In addition, although several reports have highlighted varying immune profiles that appear to correlate with survival outcomes,^6–8 the clinical utility of immune profiling (IP) has yet to be determined.

In 2016, the first annual BMT CTN Myeloma Intergroup Workshop on Minimal Residual Disease and Immune Profiling was convened to discuss emerging data and technologies for MRD and IP assessment and to develop strategies to incorporate MRD/IP into clinical trial design. ⁹ The second annual workshop was convened at the ASH meeting on December 7, 2017. Here we summarize the presentations from the workshop, review the relevant myeloma MRD/IP abstracts from ASH 2017, and discuss future considerations for the clinical and/or regulatory uses of MRD and IP.

Defining MRD:

In 2016, the IMWG published their consensus criteria for MRD assessment and incorporation of MRD status into the response criteria.⁵ They clarified that a first-pull bone marrow aspirate of 2–5 mL should be used for MRD analysis. The criteria distinguish between flow MRD-negativity and sequencing MRD-negativity. If multiparametic flow cytometry (MFC) analysis (aka next generation flow (NGF)) is to be used, then it is specified that a validated eight-color, two-tube method should be utilized, as per the established Euro-Flow procedure.¹⁰ Five million cells are to be assessed and the MFC method should have a sensitivity of at least 1 in 10⁵ plasma cells. If next generation sequencing (NGS) is to be used, then the IMWG specifies that a validated assay such as Lympho-SIGHT (Sequenta) with a minimum sensitivity of 1 × 10⁻⁵ be utilized. In September 2018, the United States Food and Drug Administration (FDA) granted De Novo designation for the clonoSEQ assay (Adaptive Biotechnologies) for the detection and monitoring of MRD in patients with myeloma or B-cell acute lymphoblastic leukemia (https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm622004.htm). The clonoSEQ assay has a sensitivity of 1× 10^-6. The IMWG also defined “Imaging plus MRD-negative” status as MRD negativity by NGF or NGS in addition to PET/CT-negativity. Finally, the concept of “sustained MRD-negativity” was introduced as achieving imaging plus MRD-negativity in assessments that are a minimum of one year apart. Although these definitions will be incorporated into ongoing and future studies, there continues to be significant heterogeneity in the sensitivity thresholds which are being utilized in recently presented abstracts, as detailed below.

MRD and IP analysis from completed clinical trials:

Myeloma MRD from UK studies:

Roger Owen provided an overview of the MRD testing performed in the Myeloma XI trial as well as several other recent UK studies. The Myeloma XI trial enrolled 1971 newly diagnosed myeloma patients, including 1248 transplant-eligible and 723 transplant-ineligible patients. Following completion of a variety of induction regimens, consolidation based on response to induction, and then autologous stem cell transplant (ASCT) for patients in the transplant-eligible group, all patients underwent randomization to lenalidomide maintenance, no maintenance or lenalidomide plus vorinostat maintenance. Data regarding the lenalidomide plus vorinostat subgroup were not presented. MRD assessment was performed via MFC at 10⁻⁴ using International Clinical Cytometry Society/European Society for Clinical Cell Analysis (ICCS/ESCCA) consensus performed at a single central laboratory. DNA was also stored for NGS testing in the future. An analysis has been performed on a subset of patients (n=389) who had evaluable samples at six and twelve months post-maintenance randomization.¹¹ MRD negativity at six months was observed in 56% (206/389) of patients. MRD status at 6 months post-initiation of maintenance had a significant impact on progression free survival (PFS): the median PFS was not reached for the MRD-negative subgroup while it was 24 months for those who were MRD-positive (HR 0.22, 95% CI 0.14–0.34, p<0.0001). In addition, superior overall survival (OS) was observed for MRD negative patients (HR 0.42, CI 0.21–0.87, p=0.0153). Notably, maintenance therapy improved PFS regardless of MRD status and 30% of patients receiving maintenance converted from MRD-positive to MRD-negative compared to 4% in the observation group (p=0.0045).

The Myeloma X study treated patients with bortezomib/doxorubicin/dexamethasone as salvage therapy following relapse after ASCT. Patients were subsequently randomized to salvage ASCT or cyclophosphamide (400 mg/m² PO weekly x 12 weeks).^{12, 13} In this study, 174 patients were randomized and day 100 MRD testing was available in 95 patients. Both PFS (HR 0.39, CI-0.24–0.61, p<0.0001) and OS (HR 0.52, CI 0.27–0.99, p=0.0434) were prolonged in patients who were MRD negative as compared to MRD positive. The PADIMAC trial (Phase II study of bortezomib, Adriamycin and dexamethasone (PAD) therapy for previously untreated patients with multiple myeloma: Impact of minimal residual disease (MRD) in patients with deferred ASCT) treated patients with PAD induction followed by stem cell harvest.¹⁴ Patients in a partial response (PR) underwent ASCT while those in very good partial response (VGPR)/complete response (CR) were observed and then offered ASCT at time of relapse. MRD assessment was performed following stem cell harvest and at day 100 post-ASCT (n=77). MRD negativity at time of harvest was associated with superior PFS. In addition, patients who achieved a PR, underwent ASCT and were MRD negative at day 100 had similar PFS as those who were MRD negative and in VGPR/CR after induction. Finally, the MUK five trial randomized patients in first relapse or with primary refractory disease to cyclophosphamide/bortezomib/dexamethasone (8 cycles of 21 days) vs carfilzomib/cyclophosphamide/dexamethasone (6 cycles of 28 days). Of the 292 evaluable patients, MRD samples were available in 62% (n=182). MRD was assessed at the end of induction and 18% of patients were MRD negative (~10% MRD-negative in the intention-to-treat population).

BMT CTN PRIMeR:

The PRIMeR (Prognostic immunophenotyping for multiple myeloma response) study is an ancillary MRD study associated with the BMT CTN 0702 STaMINA (Stem cell transplantation for multiple myeloma incorporating novel agents) trial. The STaMINA study involved 750 patients randomized to three arms: 1) single ASCT followed by lenalidomide maintenance, 2) single ASCT followed by consolidation with four cycles of VRD (bortezomib, lenalidomide, dexamethasone) and then lenalidomide maintenance, and 3) tandem ASCT followed by lenalidomide maintenance.¹⁵ To date, no differences in PFS or OS have been observed amongst the three arms. Bone marrow and peripheral blood samples were collected at randomization, prior to initiation of maintenance and at one year post-randomization. Marcelo Pasquini provided information regarding the design and results thus far from the PRIMeR study. The primary endpoint was to evaluate MRD status across treatment arms at the one-year time point. The number of bone marrow samples available for MRD were 302 at baseline, 314 prior to maintenance, and 294 at year 1. MRD was assessed centrally using 4- and 6-color MFC with 10⁻⁵ sensitivity. MRD negativity rates were 43% prior to transplant, 78% prior to maintenance and 84% at one year. MRD status is being analyzed to determine whether this is more prognostic for PFS than traditional disease response.

EMN 02/HO95:

The RV-MM-COOP-0556 (EMN02/HO95) study enrolled 1499 newly diagnosed patients.^{16, 17} Patients received VCD (bortezomib, cyclophosphamide, dexamethasone) induction followed by stem cell collection and randomization to ASCT (single or double) vs 4 cycles of VMP (bortezomib, melphalan, prednisone). Patients then underwent a second randomization (R2) to consolidation with 2 cycles of VRD vs nothing and then all patients received lenalidomide maintenance. Stefania Oliva discussed the MRD testing that was performed as part of this trial.¹⁸ MRD was assessed in patients suspected in being in CR pre-randomization (R2), prior to maintenance and then every six months during maintenance therapy until clinical relapse. MRD assessment was performed using the EuroFlow protocol³ with a maximal sensitivity of 10⁻⁵ centralized in three European laboratories. The cut-off for MRD positivity was defined as ≥ 20 clonal plasma cells out of at least 1 × 10⁴ acquired plasma cells or at least two million leukocytes. Quality checks were done amongst the three labs to compare sensitivity and demonstrate correlation between protocols. Prior to maintenance 76% of patients were MRD negative. Of the 24% who were MRD positive prior to maintenance and had subsequent MRD analysis performed after at least one year of maintenance, 44% and 48% became MRD negative after one and two years of maintenance, respectively. Of the 316 patients assessed for MRD, the median PFS was not reached for those who achieved MRD-negativity while it was 38 months for those who were MRD-positive (HR 0.33, CI 0.2–0.53, p<0.001). A landmark analysis at one year of maintenance therapy showed a statistically significant difference for the two-year PFS rate: 92% vs 65% (p<0.001) for MRD-negative vs –positive. Subgroup analysis revealed that high risk cytogenetics and ISS stage III patients were at highest risk for MRD-positivity. Despite this, those patients with high risk cytogenetics or ISS III who did achieve MRD-negativity had improved PFS vs those with MRD-positivity.

Incorporating MRD and IP assessment into current and future clinical trials:

GMMG-CONCEPT:

Katja Weisel presented the GMMG-CONCEPT study (A Clinical Phase II, multicenter, open-label study evaluating induction, consolidation and maintenance treatment with isatuximab (SAR650984), carfilzomib, lenalidomide and dexamethasone (I-KRd) in primary diagnosed high-risk multiple myeloma patients). This study will involve 117 transplant-eligible patients and 36 transplant-ineligible patients, all with high risk disease as defined by del(17p), t(4;14) or gain(1q21) and ISS II/III. In the transplant-eligible arm, patients will receive six cycles of I-KRd induction followed by single or double ASCT, consolidation with 4 cycles of I-KRd and then I-KR maintenance until progression. For the transplant-ineligible group, patients receive a total of 12 cycles of I-KRd followed by I-KR maintenance until PD. The primary objective is MRD-negativity after consolidation using MFC at 10⁻⁵ sensitivity with experimental MRD assessment being evaluated with allele-specific oligonucleotide-PCR, NGS and diffusion weighted magnetic resonance imaging (DW-MRI). All patients in VGPR/CR will undergo MRD assessment and all MRD-negative patients undergo MRD assessment every six months. The secondary objective of the study is PFS while amongst the tertiary objectives are overall response rate, duration of MRD-negativity and OS. Exploratory objectives include determining the best method for defining the MRD negative state, evaluating genetic polymorphisms such as HLA, KIR and FcGR variations at baseline and evaluating immunologic reconstitution during maintenance. The latter will be evaluated in both the bone marrow and peripheral blood using immune profiling via MFC and will assess CD4 T cells, CD8 T cells, Treg’s, NK cells, iNKT cells, monocytes, neutrophils, myeloid derived suppressor cells and dendritic cells. At the time of the workshop, seven patients had been enrolled.

SWOG S1803 (DRAMMATIC Study):

Amrita Krishnan presented the concept for the DRAMMATIC Study: Phase III study of daratumumab + lenalidomide (LD) or lenalidomide (L) as post-autologous stem cell transplant maintenance therapy in patients with multiple myeloma (MM) using minimal residual disease to direct therapy duration. Lenalidomide maintenance following ASCT has been shown to significantly prolong both PFS and OS.¹⁹ Furthermore, as noted above, the data from the Myeloma XI trial reveal that patients who achieve MRD negativity post-ASCT have improved PFS. ¹¹ In the relapsed/refractory setting, the addition of daratumumab to lenalidomide/dexamethasone significantly increased MRD negativity rates.²⁰ The DRAMMATIC study plans to address whether lenalidomide + daratumumab maintenance is superior to single agent lenalidomide maintenance and whether MRD status can be used to guide maintenance duration. Patients will be randomized to single agent lenalidomide vs lenalidomide plus daratumumab maintenance post-ASCT. Following two years of maintenance, MRD will be assessed by NGS (sensitivity 10⁻⁶). Patients who are MRD positive will continue maintenance therapy. Patients who are MRD negative will be randomized to either continuing maintenance therapy or stopping therapy. The primary objective is to compare the OS between the two maintenance arms. Secondary objectives for the first randomization include PFS, ORR, MRD-negativity rates and toxicity while for the second randomization the objectives include comparing PFS between MRD negative patients randomized to indefinite vs discontinued maintenance therapy. A total of 950 patients enrolled over 6 years are required to detect an increase in the median OS from 10 to 16.7 years in the combination arm (HR 0.6). At the time of the workshop, the plan was to assess MRD status via NGS over multiple time points.

GEM2014MAIN:

Noemi Puig discussed the Spanish Myeloma Group’s (GEM) past and current efforts to assess MRD and IP. As has previously been reported, second-generation 8-color MFC was used to evaluate MRD in 162 transplant-ineligible patients enrolled in the PETHEMA/GEM2010MAS65 study.⁸ Thirty-four percent of patients achieved MRD negativity and this was associated with improved time to progression (TTP) and OS. In addition, IP was assessed by evaluating 15 bone marrow cell subsets (erythroid and myeloid hematopoietic progenitors, erythroblasts, mast cells, eosinophils, basophils, monocytes, neutrophils, B lymphocytes and their respective precursor, naïve and memory subsets, natural killer T cells plus natural killer cells and remaining T lymphocytes) using a single 8-color combination (CD45, CD138, CD38, CD56, CD27, CD19, CD117, CD81).⁸ Based on the IP phenotype, patients were grouped into three clusters with different TTP and OS outcomes. The authors identified a subset of MRD-positive patients with a favorable IP. Currently this group is performing immune monitoring in the GEM2014MAIN study which is evaluating lenalidomide/dexamethasone vs lenalidomide/ixazomib/dexamethasone maintenance therapy post-ASCT. Following completion of two years of maintenance therapy, patients who are MRD-negative will be observed while patients who are MRD-positive will continue lenalidomide/dexamethasone maintenance for an additional three years. A single 8-color (9 marker) flow strategy is being used to identify 18 different B cell subsets, including precursors, naïve, and memory with heavy and light chain distribution. In addition, two more 8-color tubes are being used: one to identify more than 10 T-cell subsets and the second for the characterization of NK subpopulations. An example was shown demonstrating the IP of 4 patients in the GEM2014MAIN study after one year of maintenance therapy with different patterns observed when evaluating NK, CD4 and CD8 T-cells, B and plasma cells. The overall goal of this work is to integrate patient factors with the tumor landscape, MRD assessment and immune profiling to identify predictors of outcome.

Ongoing developments in MRD and IP:

Peripheral blood IP and MRD assessment:

Manisha Bhutani presented work done by the Levine Cancer Institute investigating whether the peripheral IP is distinct in MRD-positive vs -negative patients following ASCT.²¹ Thirty-six patients who underwent ASCT had bone marrow and blood specimens obtained between days 60–90 post-ASCT. The bone marrow was analyzed for MRD using next generation flow cytometry while the blood was analyzed for activation, polarization and functionality using flow cytometry. The MRD assay was based on EuroFlow with the threshold for MRD positivity defined as >15 abnormal plasma cells in 1 million nucleated cells (1.5 ×10⁻⁵). Using this assay, 6/36 (17%) patients were MRD-negative. Mature natural killer (NK), NKT-like and T cells were identified based on distribution of CD3 and CD56 and were then subsequently evaluated for major histocompatibility complex (MHC)-I cytotoxicity, CD1d cytotoxicity, activation and anergy via surface expression of killer ‘inhibitory’ Ig-like receptors (KiR2DS4, KiR3DL1), natural killer group 2 proteins (NKG2A, NKG2D), natural killer p46 protein (NKp46), programmed death receptor 1 (PD1) and T-cell inhibitory receptor (Tim3). Differences in the peripheral IP of MRD positive and negative patients were found including lower NK/γδ T cell numbers and higher NK/NK-T activation phenotype in MRD-positive patients. In addition, they monitored the IP of patients undergoing IMiD-based maintenance therapy and found that the differences in NK/γδ T cell numbers between MRD-positive and negative patients normalized in response to maintenance therapy. Future directions for this work include blood and marrow IP using NGF (62 variables) and NGS (αβ/γδ T cell receptor sequencing) in a variety of contexts including MGUS, smoldering and active myeloma patients,²² longitudinal monitoring during lenalidomide maintenance,²¹ and as correlative endpoints in ongoing and planned clinical trials.

Evaluation of clonal heterogeneity and MRD:

Niels Weinhold presented data from studies in which bone marrow aspirates and focal lesions were molecularly evaluated and compared. In these studies, patients underwent both a standard iliac crest aspirate as well as a CT-guided aspirate of focal lesion/s. CD138 cells were subsequently selected from the specimens and whole exome sequencing and copy number arrays performed. As recently reported by Rasche et al.,²³ these studies revealed the spatial heterogeneity of myeloma. In particular, evidence was provided for the regional evolution of myeloma as well as the limited exchange between sites, likely a consequence of spatial constraints within the bone marrow. In addition, focal lesions appeared to contain advanced clones. Differential responses of these clones to therapy were shown in an example of a patient who had multiple sites biopsied across multiple time points. At baseline, three different biopsy sites yielded the same dominant clone, however two months later after chemotherapy, there was one residual focal lesion which now had a different genomic profile. Seven months after that the patient had progressed and biopsy of two different sites revealed that the second clone was now dominant in both sites. In another example, a patient had four different clones identified at baseline from four different biopsy sites. Following four months of treatment, including a first ASCT, the patient was noted to be in a stringent CR with MRD positivity in the bone marrow (0.004%). Three months later following the second ASCT, the marrow was now MRD negative (<0.001%) however focal lesions were still observed in the marrow on imaging. The patient then received consolidative chemotherapy but then relapsed four months after the second ASCT. At that time, biopsy of two sites showed the presence of the baseline clone which had now acquired an additional 144 mutations as well as a second dominant clone. Finally, a third case was presented in which a relapsed/refractory patient with known disease progression was assessed as MRD negative in bilateral iliac crest aspirates. In aggregate, these studies reveal the extraordinary complexity of myeloma with clonal evolution and spatial heterogeneity and underscore the importance of interpreting marrow MRD testing in the context of imaging.

Radiographic assessment of MRD:

Jens Hillengass presented an overview of the different bone marrow infiltration patterns observed using MRI, including minimal, diffuse, focal and mixed which are estimated to be present in 20%, 30%, 30% and 20% of patients. Focal lesions can be detected using PET/CT, MRI T1, MRI T2 or DW-MRI. The heterogeneous infiltration of the bone marrow by myeloma can result in disparate MRD results depending on where the aspirate is obtained. While several studies have reported the prognostic significance of residual lesions on PET/CT at time of or after ASCT, the prognostic significance of MRI has been less evident.^24–27 Advances in DWI, which utilizes differences in the motion of water molecules between tissues, has led to the development of whole-body DWI which could be used to evaluate burden of disease and changes following therapy.²⁸ Finally, there are emerging imaging modalities which could be used to evaluate MRD and IP. The incorporation of zirconium-89-labeled monoclonal antibodies into PET imaging (immuno-PET) has been evaluated in a variety of malignancies.²⁹ This technique has the potential to not only provide visualization of the tumor, but also to quantify uptake of the radiolabeled monoclonal antibodies.³⁰ A recent report described the development of zirconium-89-labeled reconstituted high-density lipoproteins that have high uptake by tumor-associated macrophages (TAMs), which could have the potential for providing a way to noninvasively monitor TAM immunology.³⁰

MSKCC-NCI MRD Consortium:

Ola Landgren provided updates on the Memorial Sloan Kettering Cancer Center (MSKCC)-National Cancer Institute (NCI) MRD consortium which is part of the MSKCC Biomarker Development Initiative. This initiative is focused on establishing MRD as a novel surrogate endpoint to accelerate novel drug development and encompasses evaluation of myeloma biology, MRD assay development and development of assays that are independent of bone marrow sampling. As noted in other presentations, the genetic landscape of myeloma is quite complex. This institution is developing a targeted exome sequence platform “myTYPE” which utilizes FISH-seq to evaluate 120 myeloma-specific genes. This panel includes genes frequently mutated in myeloma, genes in the NFκB pathway, treatment targets such as proteasome subunit genes, immunotargets (e.g., BCMA, PD1, CD38), and candidate genes that may be associated with the development of myeloma. As a recent study revealed that at time of diagnosis patients have multiple sub-clones³¹ and that these subclones may have differential responses to therapy, the incorporation of an assay such as myTYPE at diagnosis may lead to improved decision making regarding choice of therapy. With respect to MRD assessment, Roshal et al., recently reported the development of a single tube 10-color panel MFC panel which has the advantage of decreased antibody costs and equivalent sensitivity as the EuroFlow two tube 8-color method.³² With respect to an assay that is bone marrow-independent, MSKCC has been investigating the use of matrix-assisted laser desorption/ionization (MALDI)-time of flight (TOF)-mass spectrometry (MS) analysis of monoclonal immunoglobulins. In this assay, either 20 μL of serum or 1 mL of urine is purified in five separate reactions using nanobody immunoenrichment to identify the different immunoglobulin isotypes.³³ As has been previously reported, this assay is more analytically sensitive than immunofixation electrophoresis and can simultaneously measure free light chain ratios for IgG, IgA and IgM.³³ Finally, preliminary studies in mice have demonstrated the feasibility of utilizing ⁸⁹Zr-labeled daratumumab as CD38 immuno-PET to visualize myeloma involvement with initial human studies planned in 2018.

Summary of ASH2017 abstracts on MRD and IP in myeloma:

The number of studies incorporating MRD and IP as exploratory endpoints is rapidly increasing. As shown in Tables 1 (MRD) and 2 (IP), there were multiple studies presented at the 2017 ASH annual meeting that reported included analysis of MRD and IP in myeloma patients. Despite established guidelines for MRD assessment,⁵ there continues to be heterogeneity with respect to how MRD assessed: while the majority of studies utilized MFC or NGS analysis, this was performed with varying sensitivity thresholds. Although historically, assessment of MRD focused on response to treatment in the newly diagnosed population, an increasing number of studies are now incorporating this analysis in the relapsed/refractory setting. In aggregate, these studies add to the body of existing literature demonstrating improved survival outcomes associated with achievement of MRD negativity, regardless of the methodology used to assess MRD status. However, none of these studies directly address the question as to whether MRD status can be used to tailor therapy. It will be important, once these abstracts are published in full, to undertake a formal analysis of the data so that evidence-based recommendations can be generated. Finally, as shown in Table 2, it is difficult to draw any conclusions regarding the studies assessing IP as these studies evaluated different patient populations with different assays at different time points, highlighting the need for consensus guidelines in this area.

Table 1.

Summary of ASH 2017 abstracts evaluating minimal residual disease (MRD) in plasma cell disorders

Study	MRD methodology (sensitivity)	Patient Population
ALCYONE. Mateos et al.37	NGS (10−5)	TNE ND randomized to D-VMP vs VMP
IFM2009; Avet-Loiseau et al.38	NGS (10−6)	TE ND receiving RVD induction randomized to consolidation with ASCT/RVD vs RVD followed by lenalidomide maintenance
Myeloma XI; de Tute et al.39	MFC (4 ×10−5)	TE and TNE ND
GEM2000 MM, GEM2010, GEM2012; Medina et al.40	NGS compared with MFC (≥10−5)	TE and TNE ND
EMN02/HO95. Cavo et al.16	MFC (10−5)	TE ND
Chu et al. 41	qPCR for patient-specific variable region sequence (10−5)	TE ND
Huang and Li42	MFC	TE ND
Pethema/GEM2012. Paiva et al.43	MFC (3 × 10−6)	TE ND
Terpos et al.44	MFC (2 × 10−6)	Sustained CR for ≥ 2 years after first line therapy
Rasche et al.45	MFC (10−4 to 10−5) Functional imaging: PET/CT, DWIBS	TE ND or following salvage therapy
PADIMAC. Popat et al.14	MFC 10−4	TE ND
MM5. Huhn et al.46	PCR of paired BM and PB samples	TE ND
Fernandez et al.47	MFC and PET/CT	TE and TNE ND
Pawarode et al.48	MFC (3×10−5 to 5×10−6)	TE ND, not achieving CR pre-ASCT
Solovev et al.49	MFC (10−5)	TE ND
MMY1001. Facon et al.50	NGS (10−4, 10−5, 10−6)	RR receiving DPd
Gay et al.51	MFC (10−5)	TE ND receiving KRD vs KCD induction
CASTOR. Spencer et al.52	NGS (10−4, 10−5, 10−6)	RR receiving Vd vs DVd
POLLUX. Dimopoulos et al.53	NGS (10−4, 10−5, 10−6)	RR receiving Rd vs DRd
Rosinol et al.54	MFC (3×10−6)	TE ND receiving RVD
Brudno et al.55	MFC	RR receiving BCMA CAR T-cells
Korde et al.56	MFC	ND receiving KRD induction
GEM-CESAR. Mateos et al.57	MFC	High-risk smoldering MM treated with KRd induction, ASCT, KRd consolidation and maintenance
Jimenez-Zepeda et al.58	MFC	Newly diagnosed and relapsed AL patients
Foureau et al.21	MFC (1.5 × 10−5)	Post-ASCT receiving IMiD maintenance

Abbreviations: AL, light chain amyloid; ASCT, autologous stem cell transplant; BM, bone marrow; CR, complete response; CyBorD, cyclophosphamide/bortezomib/dexamethasone; DPd, daratumumab/pomalidomide/dexamethasone; DRd, daratumumab/lenalidomide/dexamethasone; DVd, daratumumab/bortezomib/dexamethasone; D-VMP, daratumumab/bortezomib/melphalan/dexamethasone; DWIBS, diffusion-weighted magnetic resonance imaging with background suppression; IMiD, immunomodulatory drug; KCD, carfilzomib/cyclophosphamide/dexamethasone; KRD, carfilzomib/lenalidomide/dexamethasone; MFC, multiparametric flow cytometry; MM, multiple myeloma; ND, newly diagnosed; NGS next generation sequencing; PB, peripheral blood; Rd, lenalidomide/dexamethasone; RVD, lenalidomide/bortezomib/dexamethasone; RR, relapsed/refractory; TE, transplant eligible; TNE transplant not eligible; Vd, bortezomib/dexamethasone; VMP, bortezomib/melphalan/prednisone.

Table 2.

Summary of ASH 2017 abstracts evaluating immune profiling (IP) in plasma cell disorders

Study	IP methodology	Patient Population
Terpos et al.44	MFC BM	Sustained CR for ≥ 2 yrs after first line therapy
Bhutani et al.22	MFC of PB	MGUS, smoldering MM and MM
Foureau et al.21	MFC PB and BM	Post-ASCT receiving IMiD maintenance
Manasanch et al.59	MFC, gene expression profiling and exome sequencing of PB and BM	High-risk smoldering MM treated with pembrolizumab
Ocio et al.60	MFC PB and BM	Patients receiving pembrolizumab for persistent residual disease after 1-2 prior lines of therapy
POLLUX. Van de Donk et al.61	CyTOF PB	RR receiving Rd vs DRd
Neri et al.62	scRNA-seq BM	RR receiving daratumumab + IMiD
Danziger et al.63	RNA microarray from paired whole BM core biopsies and CD138-enriched BM aspirates	ND
MM-014. Qian et al.64	MFC PB	RR receiving DPd

Abbreviations: ASCT, autologous stem cell transplant; BM, bone marrow; CR, complete response; CyTOF, cytometry by time of flight; DPd, daratumumab/pomalidomide/dexamethasone; DRd, daratumumab/lenalidomide/dexamethasone; IMiD, immunomodulatory drug; MFC, multiparametric flow cytometry; MGUS, monoclonal gammopathy of undetermined significance; MM, multiple myeloma; ND, newly diagnosed; PB, peripheral blood; RR, relapsed/refractory; scRNA-seq, single cell RNA-seq.

Establishing MRD as a surrogate endpoint:

The i²TEAMM Initiative:

Nikhil Munshi presented the ongoing efforts of the i²TEAMM (International Independent Team for Endpoint Approval of Myeloma MRD) Initiative. As has been noted above and in a recent meta-analysis,¹ achievement of MRD-negativity in newly diagnosed patients is associated with superior PFS and OS outcomes. The i²TEAMM initiative is composed of myeloma research groups from the US and Europe (GEM, IFM/DFCI, MRC, EMN/HOVON, GMMG, BMT-CTN), pharmaceutical industry and an independent statistical/analytical group. The primary objective of this initiative is to evaluate and validate MRD as a surrogate endpoint for PFS in myeloma clinical trials through prospectively planned meta-analytic surrogacy analysis based on patient-level data. Fourteen clinical trials will be evaluated, including 13 randomized trials, 2 trials with multiple cohorts, 2 trials with more than two treatment arms and 5 trials with more than one randomization. Among the 11 randomized trials involving newly diagnosed patients, 7 trials used 1^st generation flow cytometry (10⁻⁴), 2 used 2^nd generation (10⁻⁵) and 2 used next generation flow (10⁻⁶). Of these 11 trials, all assessed MRD prior to maintenance and four assessed MRD after or during maintenance. The primary assessment of surrogacy will be trial-level to measure how precisely the treatment effect on the true endpoint may be predicted based on the observed treatment effect on the surrogate endpoint. It was noted that individual patient-level correlation is not sufficient to establish surrogacy of an endpoint as it does not directly follow that a treatment that alters an intermediate event will alter the long-term event in a predictable manner. The inclusion and exclusion criteria for study selection are under evaluation, however currently proposed inclusion criteria included randomized multicenter studies involving at least 200 patients per trial. Surrogacy will need to be determined for each of the different myeloma patient populations (e.g., newly diagnosed vs relapsed/refractory) separately. The data elements that will be collected include the MRD status, cycle by cycle conventional response, PFS and OS data.

Regulatory perspectives of MRD and IP testing:

Nicole Gormley from the United States Food and Drug Administration (FDA) discussed some of the regulatory issues that must be considered when proposing to use MRD or IP. She noted that MRD has distinct potential uses as 1) a clinical tool to monitor for relapse and guide therapeutic decisions and 2) a regulatory tool to determine patient stratification, patient selection, risk-based treatment assignment or as a surrogate endpoint. Similarly, the potential uses for IP include as a prognostic biomarker, a predictive biomarker, a clinical tool (e.g., to identify patients at risk for relapse, toxicity, or response; to guide therapeutic decisions), or a regulatory tool for determining patient stratification, patient selection and risk-based treatment assignment.

When considering using MRD assessment in trial design as an enrichment or stratification factor, various trial designs could be considered: enrichment design, biomarker stratified or biomarker strategy design.³⁴ In the enrichment design, all patients are assessed for eligibility but only those who are biomarker positive are randomized onto the trial. This approach does not provide any information about the biomarker-negative population and the indication for use of the biomarker assay would be limited to the selected population. In the biomarker stratified design, all patients are assessed for eligibility and the biomarker assay is used to stratify patients into separate arms. This approach has the advantage of providing information about both the biomarker-negative and the biomarker-positive populations and can evaluate the predictive and prognostic attributes of the biomarker. Finally, the biomarker strategy design allows for evaluation of the adequacy of the biomarker to guide therapy. In this design, all patients regardless of biomarker status are randomized to a traditional therapeutic approach or to a biomarker-guided treatment approach in which patients who are biomarker-negative receive standard therapy and those who are biomarker-positive receive additional or investigational treatment.

A number of factors are considered during the regulatory assessment of a biomarker.³⁵ First is the degree of risk introduced by the use of the biomarker. If the biomarker is being used in the selection/stratification process there is less risk, but if it is being used as a surrogate endpoint for regulatory approval, the risk is greater if the biomarker does not perform as expected. A second factor is the underlying biological rationale for the biomarker and understanding of its position in the disease pathway. Factors associated with the assay must be considered, including the analytical validation of performance characteristics (i.e., reliability, reproducibility, sensitivity, specificity). Another factor is the types of data available to assess the strength of the association of the biomarker with its proposed clinical outcome (e.g., PFS, OS). Finally, reproducibility of the data must be demonstrated in test and confirmatory data sets.

From a regulatory perspective, MRD is currently considered a prognostic biomarker. While the use of MRD testing in patient selection and risk-based treatment assignment may be reasonable, it must be recognized that issues related to the assay performance could introduce more risk. Consultation with the Centers for Devices and Radiological Health (CDRH) may be needed depending on the risk posed by use of the assay and the assignment of patients based on the assay. At this time, while MRD is more commonly being included as a secondary or exploratory endpoint, the field is not at the point of using it as a surrogate endpoint. To get to the point where MRD could be used as a surrogate endpoint, a number of issues must be resolved. More information is needed to determine the threshold that best correlates with clinical benefit and the optimal time points to assess MRD. The role of the disease setting must be taken into account as it is not clear whether what we know about MRD in the newly diagnosed setting translates to other settings (e.g., relapsed/refractory, smoldering). Additionally, as noted above, it is evident that other important clinical factors such as cytogenetics, extramedullary disease, and residual disease on advanced imaging impact survival outcomes. It remains to be determined how to incorporate these factors with MRD assessment. The degree of MRD improvement that is clinically meaningful has yet to be determined, thus the design of a trial with MRD as the primary endpoint is difficult from a statistical perspective. Finally, it remains to be determined how to handle the problem of missing data associated with MRD assessments performed in prior studies.

A review of the FDA’s internal databases between 2014–2016 evaluated 34 original or supplemental applications submitted to the Division of Hematology Products.³⁶ Of those, 38% included MRD data for CML, CLL, ALL and myeloma. The outcome of the FDA’s review of these applications was 3/13 were not proposed for inclusion in the USPI, 6/13 were deemed as adequate for inclusion of MRD in the U.S. prescribing information and 4/13 were deemed as inadequate. Reasons for excluding MRD data from the U.S. prescribing information were missing data, inconsistent testing across samples sources (i.e., blood vs bone marrow), high test failure rates due to inability to detect a clonal rearrangement, lack of test validation in the disease setting, incomplete test characteristics data (i.e., limit of detection) and incomplete planned statistical analysis. Thus it is critical that data collection and assay performance characteristics be of sufficient rigor and completeness to allow for a comprehensive review by the FDA. It is recommended that plans for MRD assessment be discussed with the FDA in advance.

Milestones and Deliverables:

Future annual meetings are planned to discuss development and implementation of MRD and IP testing in myeloma. The ongoing efforts of the i²TEAMM will provide key data regarding the validity of MRD as a prognostic biomarker but may not be sufficient to establish MRD as a surrogate endpoint. Continued discussions with regulatory agencies are critical as clinical trials utilizing MRD as a decision tool are planned. Currently the field is hindered by the lack of standardization with respect to IP assessment and consensus is needed to determine the source (marrow vs blood), markers/cell populations and time points to be assessed. The development of a standardized NGF panel for IP would allow routine incorporation of IP as an exploratory endpoint in ongoing/future clinical trials. It is also recognized that molecular-based methodologies that evaluate the immune microenvironment are promising and also should be pursued. The myriad issues related to regulatory approval of MRD and/or IP as endpoints will require considerable efforts by the myeloma research community to overcome.

Conclusion:

Significant progress has been made in the standardization of MRD analysis. Correlation between achievement of MRD negativity and survival outcomes has been documented across many different patient populations, including newly diagnosed transplant eligible, newly diagnosed transplant ineligible and relapsed/refractory. However, MRD status has not yet been established as a surrogate endpoint. Furthermore, there is insufficient evidence to support the use of MRD status to guide treatment decisions, outside of the context of a clinical trial. Further efforts are needed to develop consensus guidelines for standardizing immune profiling and to more routinely incorporate these studies into prospective clinical trials.

Highlights.

• Summary from the second annual BMT CTN Myeloma Intergroup MRD/IP workshop.

• MRD is currently considered a prognostic biomarker from a regulatory perspective.

• Additional prospective studies are required to develop MRD as a surrogate endpoint.

• Standardization of immunophenotyping/immune profiling studies is needed.

Abbreviations:

ASCT

autologous stem cell transplant

BM

bone marrow

CR

complete response

CyTOF

cytometry by time of flight

DPd

daratumumab/pomalidomide/dexamethasone

DRd

daratumumab/lenalidomide/dexamethasone

IMiD

immunomodulatory drug

MFC

multiparametric flow cytometry

MGUS

monoclonal gammopathy of undetermined significance

MM

multiple myeloma

ND

newly diagnosed

PB

peripheral blood

RR

relapsed/refractory

scRNA-seq

single cell RNA-seq