Comparison of Outcomes of Allogeneic Hematopoietic Cell Transplantation for Multiple Myeloma Using Three Different Conditioning Regimens

Abstract

Background:

Allogeneic hematopoietic cell transplant (allo-HCT) is a potentially curative therapy for patients with multiple myeloma as it provides graft-versus-myeloma effect alongside a myeloma-free graft. Although reduced-intensity conditioning regimens decrease non-relapse mortality (NRM), there is a paucity of data with regard to the ideal conditioning regimen in myeloma.

Methods:

We conducted a retrospective comparison of three different preparative regimens used for allo-HCT for multiple myeloma at our institution in recent clinical trials: Busulfan/Fludarabine (BuFlu), Fludarabine /Melphalan 100mg/m2 (FM100), and Fludarabine/Melphalan 140mg/m2 (FM140). NRM, progression-free survival (PFS) at 3 years, and overall survival (OS) at 3 years were the primary endpoints. Secondary endpoints included time to engraftment, and the incidence of grade II-IV acute graft-versus-host disease (aGVHD), and chronic graft-versus-host disease (cGVHD).

Results:

A total of 73 patients received allo-HCT with these regimens. NRM at 3 years was seen in 3 (21%), 5 (28 %), and 6 (24%) patients in BuFlu, FM100 and FM140, respectively. Three-year PFS in the BuFlu, FM100 and FM140 groups was 16% (HR 1.2; 95% CI 0.6–2.1), 26% (HR 0.6; 95% CI 0.3–1.2), and 11% (ref), respectively. Three-year OS in the BuFlu, FM100 and FM140 groups was 39% (HR 1.1; 95% CI 0.5–2.2), 43% (HR 0.7; 95% CI 0.3–1.4), and 32% (ref), respectively. High-risk cytogenetics and relapsed disease prior to allo-HCT were found to be independent predictors of inferior OS on multivariate analysis, with HR of 2.1 (P = 0.02) and 2.6 (P = 0.004), respectively. In contrast, the preparative regimen did not emerge as a predictor of PFS or OS.

Conclusions:

Durable clinical remission can be achieved in 11–25% of patients multiple myeloma with the use of allo-HCT, without any significant difference in the safety or efficacy of the conditioning regimen. High-risk cytogenetics and relapsed disease prior to transplant were associated with inferior PFS and OS.

Introduction:

The treatment landscape for multiple myeloma has evolved rapidly in recent years. Introduction of proteasome inhibitors (PI)^1–4, immunomodulatory drugs (IMiD)^5,6, and antibody therapies^7,8 have significantly prolonged life expectancy of patients with multiple myeloma (MM), especially when combined with high-dose chemotherapy and autologous hematopoietic stem cell transplantation (auto-HCT)^9,10. Unfortunately, even with newer agents and auto-HCT, patients still experience an eventual relapse¹¹.

Studies with myeloablative allogeneic hematopoietic stem cell transplantation (allo-HCT) have suggested that cure might be possible for a small subset of patients^12–14, which has been attributed to the graft versus myeloma effect^15,16. Unfortunately, myeloablative allo-HCT has significant non-relapsed mortality (NRM), such as infections and GvHD that can be as high as 53% at 100 days¹⁷, which has limited its use despite improvements in survival with better supportive care¹⁸. Attempts to reduce the NRM have led to the use of reduced-intensity conditioning (RIC) regimen allo-HCT. Studies have shown that RIC have decreased NRM and improved overall survival compared to myeloablative regimens¹⁷, though with higher relapse rates¹⁹. At this time, the optimal RIC regimen for multiple myeloma is still unknown.

This article investigates three different RIC regimens in MM patients who underwent allo-HCT at our single institution.

Methods:

We conducted a retrospective comparison of three different preparative regimens used for allo-HCT for multiple myeloma at MD Anderson Cancer Center in recent clinical trials. The three regimens used at our institution were: Busulfan/Fludarabine (BuFlu), Fludarabine /Melphalan 100mg/m2 (FM100), Fludarabine/Melphalan 140mg/m2 (FM140). Data regarding FM100 and FM140 were obtained from a prospective Phase II randomized clinical trial comparing the two regimens, while data for the BuFlu regimen were extracted from a separate single-arm prospective study. Eligibility for both clinical trials included age >18, ECOG PS 0–1, and adequate organ function. Both clinical trials were approved by the Institutional Review Board and were conducted according to the Declaration of Helsinki.

Preparative Regimens and Supportive Care:

The BuFlu regimen consistent of busulfan (Bu) 80 mg/m2 per day on day −13 and −12 in outpatient clinic. Bu was subsequently dosed to achieve a target AUC of 16k or 20k on days −6, −5, −4, and −3 based on pharmacokinetic studies. Fludarabine was given 40 mg/m2 intravenously on days −6, −5, −4, and −3. The FM100 and FM140 regimens consisted of the same dose of fludarabine 30 mg/m² given intravenously on days −4, −3, −2, and −1. Melphalan was dosed at 100 mg/m² or 140mg/m2 given intravenously on day −1. Patients receiving grafts from matched unrelated donors (MUDs) all received anti-thymocyte globulin (ATG) 1.5 mg/kg infused over 6 hours for 3 days (day −3, −2, and −1). All patients received similar infectious prophylaxis with standard antimicrobials. Filgastrim 5 μg/kg was started 7 days following allo-HCT until ANC > 0.5 × 10 ⁹ /L for 3 consecutive days. GVHD prophylaxis was given similarly in all three groups which consisted of Tacrolimus Day −2 or −1 alongside methotrexate 5mg/m2 on day 1, 3, 6, and 11. Rituximab was administered weekly for four doses on days −5, +2, +9, and +16 in the last 23 patients treated with FM based regimens to reduce the risk of GVHD. Donor lymphocyte infusion (DLI) was given at the discretion of the treating physician at time of relapse. Maintenance chemotherapy following transplant was allowed at the discretion of the treating physician.

International Myeloma Working Group (IMWG) uniform response criteria was used to assess response before and after transplantation. Primary endpoints were non-relapsed mortality (NRM), progression-free survival (PFS) at 3 years, and overall survival (OS) at 3 years. Survival at 3 years was chosen as it reflects the length of follow-up in the most recent of the two clinical trials. Secondary endpoints included time to engraftment, grade II-IV acute graft-versus-host disease (aGVHD), and chronic graft-versus-host disease (cGVHD). PFS was defined as the time from the day of allo-HCT to progression or death. OS was defined as the time from the day of allo-HCT to death. Observations were censored at the time the patient was last known to be alive without disease progression for PFS, and the time last known to be alive for OS. NRM was defined as death in the absence of relapse or persistence of disease. High-risk chromosomal abnormalities were defined as: del (17p), t(4;14), t(14;16), del (1p), amp (1q), del 13 by (conventional karyotyping only), and plasma cell leukemia. Categorical variables were compared using χ2 test or Fisher’s exact. Kaplan-Meier method was used to estimate the probability of OS and PFS. The cumulative incidence method accounting for competing risks was used to estimate the incidence of NRM and GVHD. Predictors of OS and PFS were evaluated on univariate and multivariate analysis using Cox’s proportional hazards regression analysis. Statistical significance was defined at the 0.05 level. Analyses were primarily conducted using STATA 14.0.

Results:

Among 73 patients who received allo-HCT from 2005 to 2015 with these regimens, there were 22, 23 and 28 patients who received BuFlu, FM100 and FM140 conditioning, respectively. Median follow-up in the BuFlu group was 31 months compared to 84 months in the FM groups.

Baseline Characteristics:

The median age at transplant was 53 years, which was similar among all three regimens (Table 1). Patients receiving BuFlu were more likely to harbor high-risk cytogenetic abnormalities (82%) as compared to FM100 (30%) and FM140 (43%) (Table 1). Durie Salmon stage and the percentage of patients achieving a VGPR or greater prior to transplant was also similar across three regimens (Table 1). Sixty-seven patients (92%) received prior auto-HCT. The majority of patients received allo-HCT at disease relapse (62%) as opposed to consolidation following auto-HCT in first remission (38%), which was similar among all three groups (Table 1). The majority of patients received a matched related donor (MRD) transplant as opposed to matched unrelated donor (MUD). The FM100 arm had a higher rate of MRD transplants 91% as compared to FM140 (57%) and BuFlu (73%) (Table 1). There were 14 patients who received maintenance therapy post-transplant which consisted of: lenalidomide (6), bortezomib (5), carfilzomib(1), carfilzomib/pomalidomide (1), and dexamethasone (1). There were six patients who received DLI.

Table 1.

Baseline characteristics of patients receiving three conditioning regimens: Busulfan/Fludarabine (BuFlu), Fludarabine/Melphalan100 (FM100), and Fludarabine/Melphalan 140 (FM140).

	BuFlu N=22	FM100 N= 23	FM140 N=28	p value
Median Age (Range)	54.5	53	52	0.2
Sex (%)				0.2
Male	14 (64)	11 (48)	16 (57)
Female	8 (36)	12 (52)	12 (43)
Donor typr				0.03
MRD	16 (73)	21 (91)	16 (57)
MUD	6 (27)	2 (9)	12 (43)
Cytogentics				0.001
Standard-risk	4 (18)	16 (70)	16 (57)
High-risk	18 (82)	7 (30)	12 (43)
Disease status at HCT				0.7
1st Remission	7 (32)	10 (43)	11 (39)
Relapsed	15 (68)	13 (57)	17 (61)
> PR to Pre-HCT Rx	17 (77)	16 (69)	19 (68)	0.7

Outcomes:

All patients achieved neutrophil engraftment, with the median time to engraftment being 12 days (range 10–24 days), which was similar across all three groups. At day 100, cumulative incidence (CI) of grade II-IV aGVHD was 4/22 patients (18%) in BuFlu, 5/23 (22%) in FM100, and 6/28 (21%) in FM140. The three-year risk of cGVHD was also similar among the three groups: BuFlu 6/22 (27%), FM100 9/23 (39%), and FM140 8/28 (29%). 100-day NRM was seen in 0, 0 and 1 (4%) patients in BuFlu, FM100 and FM140, respectively (p=0.5). 1-year NRM was seen in 2 (9%), 2 (9%) and 4 (14%) patients in BuFlu, FM100 and FM140, respectively (p=0.7). The most common cause of NRM was GVHD. There were four deaths from aGVHD, two each in BuFlu and FM140. There were four deaths from cGVHD, two each in FM100 and FM140 groups, respectively.

In terms of post allo-HCT response, 5 (23%), 4 (18%) and 8 (36%) patients achieved nCR/CR/sCR, VGPR, and PR with BuFlu; 10 (43%), 4 (17%), 7 (30%) with FM100; and 9 (32%), 6 (21%), 9 (32%) with FM140. The overall response rates (CR+VGPR+PR) were: BuFlu 17 (77%), FM100 21 (91%), and FM140 24 (86%). The 3-year PFS in the BuFlu, FM100 and FM140 groups was 16% (HR 1.2; 95% CI 0.6–2.1), 26% (HR 0.6; CI 0.3–1.2), and 11% (ref), respectively (p=0.2; Figure 1A). The 3-year OS in the BuFlu, FM100 and FM140 groups was 39% (HR 1.1; CI 0.5–2.2), 43% (HR 0.7; CI 0.3–1.4), and 32% , respectively (p=0.5, Figure 1B).

Figure 1.

The (A) PFS and (B) OS by conditioning regimen.

Fourteen patients went on to receive post-transplant maintenance. Ten patients (71%) progressed while on maintenance with a median time to progression of 8.3 months. There were two patients who died while on maintenance therapy. At time of analysis, two patients were still receiving maintenance with lenalidomide with ongoing response of sCR, 43 months and 99 months post allo-HCT.

Six patients received DLI, 4 after disease progression, and 2 for persistent disease after transplant. Of the four DLI given at disease progression, three failed to respond and died within two months. One patient achieved a VGPR and progressed after 23 months. Of the two patients who received DLI for persistent disease, one with high-risk cytogenetics progressed after 4.7 months, while the other with standard-risk cytogenetics progressed 31.8 months after DLI.

We evaluated the role of preparative regimen, donor type, Durie-Salmon stage, cytogentics, response to pre-transplant therapy, and disease status at transplant on PFS and OS in univariate analyses. On univariate analysis for PFS, high-risk cytogenetics, less than VGPR to pre-transplant therapy, and relapsed disease at allo-HCT were significant predictors for worse PFS. Of these factors, high-risk cytogenetics and relapse disease remained significant predictors for worse PFS in a multivariate analysis (Table 2). Similarly, for OS, both high-risk cytogenetics and relapsed disease at allo-HCT remained significant predictors on both univariate and multivariate analyses (Table 2). Preparative regimen was not associated with PFS or OS when added to the multivariate analysis.

Table 2.

Multivariate Analysis for PFS and OS

	PFS			OS
	HR	95% CI	p	HR	95% CI	p
Cytogentics
Standard-risk	Ref.			Ref.
High-risk	2.3	1.3–4.1	0.004	2.1	1.1–4.1	0.02
Disease status at HCT
1st Remission	Ref.			Ref.
Relapsed	2.5	1.5–4.4	0.001	2.6	1.3–5.1	0.004

Patients (n=15) with standard risk cytogenetics who received allo-HCT while in first remission had a median PFS and OS of 28 and 78 months, respectively (Figure 2A and B). In contrast, patients (n=24) with high-risk cytogenetics and relapsed disease status at allo-HCT had a median PFS and OS of 3 and 8 months, respectively (Figure 2A and B). The inferior OS and PFS in patients with high-risk cytogenetics and relapsed disease status prior to allo-HCT was seen across all three conditioning regimens.

Figure 2.

The (A) PFS and (B) OS Categorized by Cytogenetic Risk and Disease Status (First remission vs Relapse).

Discussion:

In this retrospective analysis, we compared three preparative regimens used for allo-HCT for multiple myeloma. There were no significant differences in NRM, GVHD, or survival between the three regimens.

Several studies have shown a significant improvement in early TRM after allo-HCT with RIC regimens, and in patients who received allo-HCT in recent years^20,21. Most of these regimens have used a combination of fludarabine or low-dose total body irradiation (TBI) with an alkylating agent, either melphalan or busulfan^22–24. Our results are consistent with outcomes reported in some of the recent allo-HCT studies that used similar conditioning, and reported a 1–3 year NRM of 22–27%, 2–3-year PFS 18–34%, and 2–3 year OS of 36–50%^20,21. A recent single-center study from The University of Michigan reported on 22 patients with high-risk cytogenetics or early relapse who underwent an allo-HCT with a conditioning regimen of fludarabine and busulfan. Their reported NRM at 3 years was 29%, PFS at 3 years was 15%, and OS at 3 years was 33%²⁵.

The regimens used in this report range from RIC (FM100) to myeloablative (FluBu) as defined by Bacigalupo et al²⁶. However, NRM was comparable in the 3 groups, thereby confirming the relative safety of myeloablative regimens, such as BuFlu, in recent years perhaps due to better patient selection and supportive care.

In our study, patients who were transplanted after relapse, or had high-risk cytogenetics had shorter PFS and OS, compared to patients transplanted in first remission or with standard-risk cytogenetics. Several previous single-center and registry studies have shown that patents transplanted with relapsed or refractory disease had worse outcomes. Bashir et al from our group reported 5-year PFS and OS of 15 and 21% after allo-HCT for myeloma²⁷. They reported that patients receiving allo-HCT for relapsed disease had a significantly shorter PFS and OS. Similarly, Freytes et al. from CIBMTR reported 3-year PFS and OS of 6 and 20%, respectively, in patients who received a RIC all-HCT in a salvage setting²⁸.

Several studies have suggested that allo-HCT can overcome high-risk cytogenetic abnormalities^29,30. Schilling et al, in a multicenter analysis on 101 patients with advanced disease and multiple chromosomal abnormalities, showed that only del(17p13) was associated with a shorter event-free survival due to increased relapse incidence²⁹. Similarly, Roos-Weil et al from the French cooperative group SFGM-TC reported the outcome of 143 myeloma who underwent an allo-HCT. They did not see any difference in PFS, OS or progression rate between patient with or without high-risk chromosomal abnormalities³⁰. In contrast, other studies showed worse outcome in patients with high-risk cytogenetics after an allo-HCT²⁷. Our results are consistent with reports that show an adverse impact of HR cytogenetics on the outcomes from allo-HCT. Other important issues related to allo-HCT are the role of DLI^31,32 and post allo-HCT maintenance therapy^33,34, which could not be adequately addressed in our study due to the small number patients receiving these interventions.

Some of the limitations of our study include small sample size, variable follow up, heterogeneous patient population treated over a decade, and lack of data on minimal residual disease (MRD), since the test had not been standardized until recently.

In conclusion, Bu/Flu, FM100, and FM140 demonstrated similar outcomes and toxicity when used as conditioning regimens for multiple myeloma. High-risk cytogenetics and disease relapse were independent predictors of worse outcome in allo-HCT.

ABBREVIATIONS:

allo-HCT

allogeneic hematopoietic cell transplant

auto-HCT

autologous hematopoietic cell transplant

HR

hazard ratio

MDACC

MD Anderson Cancer Center, OS, overall survival

PFS

progression-free survival

PR

Partial response

PS

performance status

SD

Stable disease