Disease biology alters the response to frontline bortezomib, lenalidomide and dexamethasone in Multiple Myeloma

Multiple Myeloma (MM) is a heterogeneous incurable disease. While majorities of patients have indolent courses with good response to upfront induction, some patients progress rapidly despite treatment. No consistent correlation between the intensity of therapy and long-term disease outcome currently exists [1]. Complete response has been demonstrated to strongly correlate with improved progression-free survival [2,3].

that can be utilized in selecting regimens to achieve the best response. Current standard initial therapy for patients with MM depends on the cytogenetic risk of disease [5,6] and eligibility for Autologous Stem Cell Transplantation (ASCT) [7]. However, there is no general agreement as to the standard induction regimen.
A three-drug regimen has been demonstrated to be more effective as compared to doublet regimens. The addition of bortezomib to lenalidomide and dexamethasone (VRd) as compared to lenalidomide and dexamethasone (Rd) in the SWOG S0777 trial resulted in signifi cantly improved Progression-Free Survival (PFS) and Overall Survival (OS) [8]. Approximately 525 patients were included in phase 3 multicenter study; they were randomly assigned to VRd or lenalidomide and dexamethasone (Rd) only. VRd resulted in higher rates of overall response (82 versus 72 %) [8,9]. Moreover, VRd continues to represent an appropriate standard of care based on the longer-term followup data [8].
Similarly, the risk of disease progression or death was signifi cantly lower with daratumumab plus lenalidomide and dexamethasone than lenalidomide and dexamethasone alone [10]. Herein, three-drug regimens are the mainstay of initial therapy for MM while two-drug regimens may still be of importance in frail patients who may not tolerate standard three-drug regimens [11].
Over the last 20 years, several new agents, such as Immunomodulatory agents (IMiDs), Proteasome Inhibitors (PIs), monoclonal Antibodies (mAbs), have been FDA-approved. These agents have been incorporated into clinical guidelines and have transformed our approach to the treatment of MM patients [12].
More recently, the benefi t of adding a fourth drug is being evaluated in several clinical trials. The phase 2 trial (GRIFFIN) randomized 207 patients to VRd with or without daratumumab (D-VRd versus VRd). It showed that D-VRd resulted in a higher overall response rate (99 versus 92 percent) and deeper response (63% reached stringent Complete Remission (sCR) and CR). However, toxicity was greater with D-VRd with higher rates of neutropenia and upper respiratory tract infections [15].
The association between depth of response and the longterm outcome remains a debated topic in MM, however, the relationship between a complete response and progressionfree survival has been more consistent [16]. A signifi cant correlation between the achievement of Complete Response (CR) or Very Good Partial Response (VGPR) and improved PFS, and thus OS has been demonstrated in eight out of ten studies included in a meta-analysis [3].
Additionally, the achievement of response less than a Very Good Partial Response (VGPR) after initial induction is an adverse prognostic factor for Progression-Free Survival (PFS) in MM [2]. Therefore, achieving CR or VGPR is a valid surrogate marker of the treatment effi cacy [2].
While debating between three-drug regimens versus four-drug regimens aiming to achieve the deepest response to induction regimen for newly diagnosed MM, balancing effi cacy and toxicity is a challenge due to advanced age and comorbidities associated with the myeloma population.
Biomarkers that can guide the selection of frontline therapy based on prediction for effi cacy would be clinically useful in determining the best approach to achieve maximal benefi t in an individualized approach rather than a trial-and-error approach.
In this study, we hypothesized that a selective subgroup of patients with newly diagnosed MM, which share common biological markers, may not benefi t from frontline treatment with VRd.

Study aims
Based on previous studies demonstrating the best initial response to be strongly correlated with survival outcomes [2,3,16], this study aims at evaluating the initial response rate to 3-6 cycles of frontline VRd as a surrogate marker for treatment effi cacy. Additionally, we aimed to evaluate whether disease burden, biology, or presenting clinical picture impact response to VRd.
The primary outcome of this study was to identify if certain characteristics of myeloma are associated with higher or lower initial response rates to VRd. Results from our study may become a future platform in selecting an induction regimen through a personalized approach.

Study design
This is a cross-sectional study including participants with newly diagnosed MM. Participants were selected from the cohort of patients with newly diagnosed MM treated in our institution from April 2011 until May 2018, and, followed for at least the fi rst 6 months after initiation of therapy.
We classifi ed our cohort according to response to therapy several variables that could potentially affect response to therapy and analyzed the correlation between different variables and response to therapy.
We classifi ed our cohort based on gender, age at diagnosis, stage of disease [5], disease biology in terms of Free Light Chain analysis (FLC), and immunofi xation (IgG, IgA, or IgM monoclonality). We also classifi ed our patients by their risk strata using Fluorescence in Situ Hybridization (FISH) and cytogenetics (standard versus high risk) [5].
Disease burden defi ned as the percent of bone marrow involvement with malignant plasma cells. CRAB criteria, hypercalcemia defi ned as serum calcium >11 mg/dL, renal insuffi ciency defi ned as serum creatinine >2 mg/dL, anemia defi ned as hemoglobin level < 10 g/dl, presence of bone lesions defi ned as one or more osteolytic lesions ≥5 mm in size by PET-CT or MRI, and the presence or absence of extramedullary disease.
For response assessment, we use the International Myeloma Working Group (IMWG) response defi nition [17].
For MM staging, we used the International Staging System) [18,19]. Our Institutional Review Board (IRB) approved the research protocol. We excluded patients who had evidence of other malignancies that required active treatment.

Statistical analysis
Data were analyzed using Statistical Analysis Software (SAS) 9.4.
Among the cohort of patients treated in our institution at the selected period, we included in our analysis only the patients who received the bortezomib-based regimen, and the cohort who received VRd for at least 3-6 cycles.
Data collected included age at diagnosis, gender, presence of CRAB at the initial presentation, presence of extramedullary disease at the time of diagnosis, ISS stage, and percent of bone marrow involvement at diagnosis, albumin level at diagnosis, FLC restriction, FLC ratio, and risk classifi cation for each participant.
We classifi ed our cohort based on response status into three groups (responders, partial responders, and non-responders).
We analyzed the correlation between each different variable and response status. For categorical dichotomous variables, we  (96 subjects) received VRd and 20% (24 subjects) received Vd due to intolerance and fragility. All participants included in this study were able to fi nish 3-6 cycles of VRd.

Patients' characteristics
The median age at diagnosis was 60.5 (ranging from 39-73 years), 62% of participants were males and 38% were females.
Thirty-seven percent were lambda restricted and 63% were kappa restricted. The majority of our cohort had IgA or IgG monoclonality on immunofi xation, 31% were IgA and 68% were IgG. 41% had > 60% bone marrow involvement by malignant plasma cells. Response to therapy evaluated after the third cycle for all patients with 50% achieved VGPR or better, 45% achieved PR, and fi ve patients had disease progression ( Table   1).

Analysis of VRd treated cohort
We had 96 subjects who received at least 3 cycles of frontline VRd. Thirty-six subjects had high-risk disease and 45 had standard risk with missing risk strata on 15 subjects. In regards to response assessment, 50% (48 of 96 subjects) were responders, 45% (43 of 96 subjects) were partial-responders and 5% were non-responders.
When analyzing response status by gender in our VRd treated cohort, the 96 subjects were 36 females and 60 males. Despite having a numerically higher response rate in females with 56% responders, it was not statistically signifi cant (P-value 0.31).
In addition, there was no statistically signifi cant difference in the mean age at diagnosis among the three outcomes (P-value 0.114).
Analyzing disease biology and its correlation with response to VRd, we showed that neither FLC restriction (Kappa or lambda) nor the FLC ratio (involved /uninvolved FLC) showed a statistically signifi cant difference in the response to VRd partial-responders were in the IgG group, and all fi ve nonresponders in the VRd cohort were IgG subtypes ( Table 2).

Subgroup analysis
An unplanned subgroup analysis was performed on the IgG and IgA group to analyze the correlation between risk stratifi cation and response outcomes in those subgroups. We also analyzed the correlation between FLC restriction and the presence of bone lesions at the time of diagnosis. We selected to do subgroup analysis for those variables due to a discordance of prevalence in the study cohort, the majority of our cohort had osteolytic bone lesions at diagnosis and about two-thirds had kappa restriction.
Among subjects with IgG type myeloma, neither risk of disease (P-value= 0.52), FLC restriction (P-value= 1.0), nor presence of osteolytic bone lesions at diagnosis (P-value=0.51) correlated with treatment response.
Additionally, in the IgA type myeloma, our data did not show a statistically signifi cant difference between risk strata, FLC restriction, or osteolytic lesions at diagnosis and treatment outcomes (P-values= 1.0, 1.0, and 0.56 respectively).
Other variables including disease burden, percent of bone marrow involvement with malignant plasma cells, presence of CRAB criteria at diagnosis, presence of extramedullary disease, and ISS stage failed to show a statistically signifi cant correlation with treatment outcomes.
In conclusion, our results showed an association between immunoglobulin monoclonality in MM and initial response to VRd. IgG type myeloma showed to have a less optimal response to frontline VRd, characterized by a higher rate of partial response and a lower rate of VGPR or CR in comparison to IgA myeloma patients.

Discussion
Multiple myeloma is an incurable disease [20,21], achieving complete response is the goal of therapy as it correlates with improved overall survival and progression-free survival [2,3,16]. VRd is the preferred frontline induction regimen to date due to its tolerability and its ability to induce a deep response in patients with newly diagnosed MM regardless The risk stratifi cation (high risk versus standard risk) between treatment outcomes also failed to show a statistically signifi cant difference P-value=0.52.
However, when analyzing the disease biology in terms of serum immunofi xation results describing the immunoglobulin type (monoclonal IgA, IgG, IgM, or IgD), we showed clinically and statistically signifi cant difference between responders, partial-responders, and non-responders (P-value=0.0006).
Remarkably, 79% (23 of 29 subjects) in the IgA group achieved VGPR or better compared to 37% (23 of 63 subjects) of the IgG group who achieved VGPR or better. The majority of our  of if followed by autologous stem cell transplantation or not [9,22]. We studied several factors that might affect the initial response to VRd. We summarized potential factors into patient demographics, disease burden at diagnosis, disease stage, and biology.
The main fi nding, we report from this cross-sectional study is that disease biology in regards to immunoglobulin monoclonality (heavy chain subtypes) has a statistically signifi cant association with response outcomes.
Although the impact of serum immunofi xation was addressed in previous retrospective studies and did not show an association with disease outcomes [23][24][25], IgD myeloma was associated with worse outcomes due to a higher incidence of renal insuffi ciency [26]. In a large multicenter retrospective study, which addressed the same question of the effect of paraprotein on myeloma outcomes, no difference was observed between IgA and IgG in overall survival. However, it showed a shorter PFS for IgA type myeloma. [24] This study concluded that worse PFS was contributed to renal insuffi ciency and specifi cally higher levels of FLC excretion, which was more prominent in the IgA group. Their data showed that patients with IgG or IgA subtype with the same level of urinary FLC excretion had the same incidence of renal failure and poor survival.
We are reporting a different outcome, IgG paraprotein is associated with suboptimal response to frontline using VRd, but showed no signifi cant difference when we stratifi ed subjects based on FLC type in our subgroup analysis.
Our study implicates that the immunoglobulin subtype of myeloma could be another factor that needs to be taken into account as we select a preferred regimen for optimal response.
Quadruplet regimen with targeted monoclonal antibodies in combination with VRd as frontline therapy in newly diagnosed multiple myeloma is superior to VRd in Overall Response Rate (ORR) and PFS, however, it may carry a higher toxicity [15,27].
Whether an upfront quadruplet regimen would provide a better response rate in the IgG subgroup as compared to triplet VRd is not known. While balancing benefi ts and toxicities in selecting therapeutic regimens for patients with multiple myeloma is crucial, further confi rmation of our fi ndings is critical.
A subgroup analysis using the previous study to analyze the outcome in IgG and IgA myeloma, using frontline D-VRd versus VRd, or a future large randomized prospective study comparing the response to D-VRd versus VRd based on immunoglobulin subtype of multiple myeloma could help validate our fi ndings.
Our data could also potentially affect how we stratify therapeutic regimens in various individuals, preferring quadruplet regimen to VRd with IgG type MM, especially if it is in the context of molecularly high-risk disease.
There are certain limitations to our study. The small sample size in the VRd-treated cohort limits our conclusion; perhaps a more collaborative effort with additional institutions to increase our sample size would strengthen the validity of our results.
Designing the study as a cross-sectional retrospective study limits our results. Although we could evaluate the association between immunoglobulin monoclonality and treatment outcomes, we are unable to state a causal relationship between IgG type MM and poor response to frontline VRd.
We achieved the primary aim of our study which is IgG subgroup of MM patients was associated with lower response rates to frontline VRd. Our data raises awareness of the importance of disease characteristics and biology in treatment response. As we aim at personalized medicine and designing individualized treatment algorithms in MM, future studies evaluating the effi cacy of various treatment regimens for MM should always incorporate disease characteristics and biology as one the important parameters in the analysis.

Conclusion
We conclude that IgG type MM was associated with suboptimal response to frontline VRd. Our limited single institutional study suggests that immunoglobulin monoclonality could be an important parameter to be considered as we stratify therapeutic regimens in multiple myeloma. Further studies would help confi rm our fi nding. MM remains an incurable disease and a disease that occurs in older individuals who commonly carry underlying comorbidities, being able to balance treatment-related benefi ts and risks is crucial. Our fi ndings could be important as we aim towards personalized medicine.

Disclosure
The data that support the fi ndings of this study are available from the corresponding author upon reasonable request.