Non-myeloablative allogeneic hematopoietic stem cell transplantation for adults with relapsed and refractory mantle cell lymphoma: a single-center analysis in the rituximab era.

Relapsed and refractory (rel/ref) mantle cell lymphoma (MCL) portend a dismal prognosis. Allogeneic hematopoietic stem cell transplantation (allo-HSCT) represents the only potentially curative therapy in this setting. We analyzed the survival outcomes of 29 recipients of non-myeloablative allo-HSCT for rel/ref MCL, and studied possible prognostic factors in this setting. The cumulative incidences of disease progression and non-relapse mortality at 3 years were 28% (95% confidence interval (CI): 13-46%) and 29% (95% CI: 13-47%), respectively. The cumulative incidence of grade II-IV acute GvHD at days +100 and +180 was 34% (95% CI: 18-52%) and 45% (95% CI: 26-62%), respectively. With a median follow-up in survivors of 53 (range 24-83) months, the 3-year overall survival (OS) and PFS were 54% (95% CI: 38-76%) and 41% (95% CI: 26-64%), respectively. In vivo T-cell depletion with alemtuzumab (n=6) was associated with inferior 3-year PFS (0% vs 51%, P=0.007) and OS (17% vs 64%, P=0.014). Conversely, a second-line international prognostic index (sIPI) at transplantation equal to 0 (no risk factors) was associated with an improved 3-year PFS (52% vs 22%, P=0.020) and OS (71% vs 22%, P=0.006) compared with sIPI ⩾1. Performing an allo-HSCT before 2007 was associated with a decreased 3-year OS (25% vs 76%, P=0.015) but not with a significantly inferior PFS (17% vs 59%, P=0.058). In this single-center series, we report encouraging results with allo-HSCT for patients with rel/ref MCL. High alemtuzumab doses should probably be avoided in this context.

INTRODUCTION

Mantle cell lymphoma (MCL) comprises approximately 6% of all non-Hodkgin lymphoma (NHL) and typically portends a poor long-term prognosis. Recent advances in the treatment of MCL have resulted in improved survival. Sequential high-dose chemotherapy followed by autologous stem cell transplantation or hyper-fractionated chemotherapy have lead to higher complete remission (CR) rates and remission duration exceeding 5 years in recent series.[1-4] Additionally, the introduction of novel drugs in the relapsed setting now offers effective therapeutic options.[5-11] Despite these improvements, patients with MCL have the worst long-term prognosis of any B cell NHL. Patients who relapse after intensive first line therapy have limited options to achieve durable disease control with conventional and novel therapies.[12, 13] Allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains the only potentially curative treatment, with the predominant mechanism of action attributed to potential graft-versus-lymphoma (GVL) effects.[14-18] Retrospective studies, most of them from registry data, have shown a mean progression-free (PFS) and overall (OS) survival of 25-40% and 30-50% at 3 years, respectively,[19-22] with more favorable results observed in single-center studies,[23-25] Transplant-related mortality (TRM) has ranged from 25 to 40% at 3 years.[19-22] We sought to identify potential prognostic factors for patients with relapsed and refractory (rel/ref) MCL undergoing non-myeloablative (NMA) or reduced intensity conditioning (RIC) allo-HSCT in the post-rituximab era.

METHODS

Patients

In this retrospective single center study, we analyzed 29 patients with rel/ref MCL who underwent non-myeloablative (NMA) or reduced intensity conditioning (RIC) allo-HSCT at Memorial Sloan Kettering Cancer Center (MSKCC) between April 1999 and May 2013. Written informed consent for treatment was obtained from all patients and donors. Approval for this retrospective analysis was obtained from the MSKCC Institutional Review and Privacy Board. All patients had biopsy proven MCL as defined by the World Health Organization criteria including immunohistochemical analysis for cyclin D1 and/or cytogenetic analysis by either conventional karyotyping or fluorescence in situ hybridization (FISH) for t(11;14)(q13;q32).

Eligibility criteria for transplant included availability of a human leukocyte antigen (HLA)-matched or single-allele-mismatched donor or appropriate cord blood (CB) graft. Double-unit CB (DUCB) grafts were 4-6/6 HLA-A,-B antigen, −DRB1 allele matched to the recipient with a cryopreserved total nucleated cell (TNC) dose > 1.5 × 107/kg/unit as previously described.[26] Unit-unit HLA-match was not considered in unit selection. Additional criteria included absence of active infection, lack of cardiac, pulmonary, hepatic or renal dysfunction that would preclude administration of the cytoreductive regimen. HLA matching was performed with DNA sequence-specific oligonucleotide typing for HLA-A, -B, -C, DRB1 and –DQB1 loci.

Conditioning and transplantation procedure

All patients received a NMA or RIC regimen. The predominant NMA regimen consisted of cyclophosphamide 50 mg/kg (day -6), fludarabine 25 mg/m2 for 5 days (from day -6 to -2) and TBI 200 Gy (day -1). In recipients of DUCB, the dose of fludarabine was 30 mg/m2 for 5 days. In 17 patients, rituximab was administered at 375 mg/m2 day -8 or -7 and weekly for 4 doses beginning from day +21. In this NMA regimen group, recipients of MRD or MUD were treated on (n=8) or as per (n=7) a prospective phase II clinical trial (clinicaltrials.gov NCT00425802),[27] while recipients of DUCB (n=5) were treated on a parallel prospective phase II clinical trial (NCT00387959).

The RIC regimen consisted of fludarabine 25 mg/m2 for 5 days (day -8 to -4) followed by melphalan 70 mg/m2 for 2 days (from day -3 to -2), with (n=6) or without (n=2) alemtuzumab (20 mg flat dose for 4 days (day -8 to -5). Six patients were treated on (n=5) or as per (n=1) a prospective phase II clinical trial (NCT00027560).[28] One patient received cyclophosphamide 750 mg/m2 and fludarabine 30 mg/m2 for 3 (day -7 to -5). Donor stem cells were infused at day 0 after at least 24-48 hours from the completion of chemotherapy.

GVHD prophylaxis and supportive care

For patients receiving PBSC grafts, GVHD prophylaxis consisted in cyclosporine alone (n=6),[28] or in association with methotrexate (n=1) or mycophenolate (n=4). The other patients received tacrolimus-based prophylaxis with mycophenolate mofetil (MMF, n=1) or methotrexate plus sirolimus (n=12).[29] All recipients of DUCB grafts (n=5) received cyclosporine and MMF, as previously described.[30] Equine (30mg/kg total dose) or rabbit (5mg/kg total dose) antithymocyte globulin (ATG) was used in 1 recipient of a matched related donor, 5 recipients of matched unrelated donors and 2 recipients of a mismatched related donor. Recipient of DUCB and patients receiving alemtuzumab did not receive ATG. Patients were managed clinically according to standard institutional guidelines, including antimicrobial prophylaxis. Monitoring of cytomegalovirus (CMV) reactivation in peripheral blood initially by CMV pp65 antigenemia assay, and later by CMV PCR assay (beginning in November 2011), was performed regularly through day 100 when either the patient or donor was CMV seropositive. Preemptive therapy was instituted in patients with documented CMV viremia per institutional standard.

Study definitions

Mantle Cell Prognostic Index (MIPI), second line international prognostic index (sIPI) and Hematopoietic Stem Cell Transplant Comorbidity Index (HCT-CI) were calculated retrospectively according to published methods.[22, 31-33] Second line IPI was calculated assigning 1 point for each of the following risk factors: age >60; serum LDH >1 normal value; performance status (ECOG score) ≥ 2; Ann Arbor stage ≥ 3; extranodal involvement > 1 site. Remission quotient was calculated as the number of months from diagnosis to allo-HSCT divided by the number of previous lines of therapy.[34] Standard definitions were used to assess disease response to therapy.[35] Time to neutrophil recovery was defined as the first of 3 consecutive days post transplant with an absolute neutrophil count (ANC) ≥500/μl. Platelet engraftment was defined as the first of 3 consecutive days at platelet count > 20.000/μl and at least 7 days without platelet transfusion support. GVHD was diagnosed clinically with histologic confirmation when appropriate. Acute and chronic GVHD were graded according to the International Bone Marrow Transplant Registry and the National Institutes of Health consensus criteria, respectively.[36, 37]

Study endpoints and statistical analysis

Analyses were performed as of December 31th, 2014. OS was defined from the date of transplant to death from any cause. PFS was defined from the date of transplant to disease progression or relapse, death, or last follow-up, whichever came first. Non-relapse mortality (NRM) comprised all deaths not related to disease progression or relapse. The following variables were assessed for their effects on OS and PFS: sIPI, time from diagnosis to allo-HSCT, age at transplant, prior lines of therapy, remission quotient, prior auto-HSCT, remission status before allo-HSCT, Positron Emission Tomography (PET) status pre allo-HSCT, HCT-CI, type of donor, use of alemtuzumab, and use of ATG. Ki-67 proliferation index was available only in a small fraction of patients, and was therefore not included in the analysis. Univariate analyses were performed using the log-rank test. Kaplan-Meier (KM) survival and a permutation-based logrank test were used to compare PFS and OS based on alemtuzumab use and sIPI. The cumulative incidence of aGVHD and of cGVHD, progression of disease (POD) and non-relapse mortality (NRM) incidence were calculated using the competing risk method. All analyses were conducted using the R statistical package (version 3.1.1).

RESULTS

Engraftment and GVHD

All patients had previously received rituximab. Twenty-four patients received peripheral blood stem cells (PBSC) from an HLA-matched or single-allele-mismatched sibling (n=12) or unrelated donor (n=12). Double unit cord blood grafts were used in 5 patients. Complete pretransplant characteristics of the 29 patients are described in Table 1. All but one DUCB patient engrafted. The median times to neutrophil and platelet engraftment were 11 days (range 0-60 days) and 16 days (range 0-152 days), respectively. The cumulative incidence of grade II-IV aGVHD for the entire cohort at day +100 and +180 were 34% (95%CI: 18-52%) and 45% (95%CI: 26-62%), respectively. One patient developed grade IV aGVHD. Median time to onset of aGVHD was 74 days (range 17-400 days). The cumulative incidence of cGVHD at 1 and 3 years was 24% (95%CI: 10-41%) and 35% (95%CI: 17-53%), respectively. Three patients experienced severe cGVHD by NIH criteria. The sub-type of cGVHD was classic in 3 patients and overlap syndrome in 8 patients.

Table 1

Patients characteristics (N=29)

Patient Characteristic	N (%)
Male, n (%)	25 (86%)
Median age, y (range)	37 (34-71)
Stage ≥3 at diagnosis, n (%)	28 (96%)
Blastic variant	2 (7%)
Extranodal involvement, n (%)	26 (90%)
MIPI at SCT
low risk, n (%)	22 (81%)
intermediate risk, n (%)	4 (16%)
high risk, n (%)	1 (1%)
IPI score at SCT (sIPI)
0 no risk factors, n (%)	19 (68%)
>= 1 risk factor, n (%)	9 (32%)
Median time from initial diagnosis, mo (range)	53 (6-200)
Median no. of prior lines of chemotherapy SCT (range)	5 (1-6)
LDH elevated at SCT (%)	12 (41%)
Remission Quotient *<11, n (%)	13 (45%)
Extranodal sites at SCT, n (%)	4 (14%)
Previous autoSCT, n (%)	13(45%)
PET positive before SCT, n (%)	9 (37%) (Missing data, n = 5)
Disease status at SCT
CR, n (%)	17 (59%)
PR, n (%)	9 (31%)
SD, n (%)	3 (10%)
Chemosensitive disease beforeSCT, n (%)	26 (90%)
HCT-CI≥3, n (%)	5 (17%)
Source of cells (PBSC), n (%)	24 (83%)
Source of cells (DUCB), n (%)	5 (17%)
Rituximab during conditioning, n (%)	18 (62%)
ATG, n (%)	10 (34%)
Campath, n (%)	6 (21%)
Conditioning regimen
Fludarabine/Melphalan/ Alemtuzumab	6 (21%)
Fludarabine/Melphalan	2 (7%)
Cyclophosphamide/Fludarabine/TBI	20 (69%)
Cyclophosphamide/Fludarabine	1(3%)

MIPI: Mantle Cell International Prognostic Index; SCT: Stem Cell Transplantation; IPI: International Prognostic Index; sIPI: second line International Prognostic Index; LDH: Lactate Dehydrogenase; PET: Positron Emission Tomography; CR: Complete Remission; PR: Partial Remission; SD: Stable Disease; HCT-CI: Hematopoietic Stem Cell Transplantation Comorbidity Index; PBSC: Peripheral Blood Stem Cells; DUCB: Double Umbilical Cord Blood; ATG: Anti-thymocyte Globulin; TBI: Total Body Irradiation.

Reactivation of cytomegalovirus and EBV

In patients at risk (donor and/or host CMV seropositive), CMV reactivation was documented in 9 of 19 patients. All 6 patients treated with alemtuzumab were CMV seropositive and 3 had CMV reactivation. No patients developed CMV end-organ disease. None of the patients reactivated EBV.

Non relapse mortality

The NRM of the entire cohort at 3 years was 29% (95% CI:13-47%). Five of 9 deaths were attributed to GVHD while the remaining deaths were secondary to: Enterobacter cloacae spp. septicemia, Pseudomonas aeruginosa septicemia, probable progressive multifocal leukoencephalopathy and one acute ischemic stroke.

Progression of disease and survival

The cumulative incidence of POD at 3 years was 28% (95%CI: 13-46%). Of the 10 patients who progressed, 7 died of MCL. The median time to POD was 12 months after allo-HSCT (range 1-85 months). Two patients received donor leukocyte infusion (DLI) for POD without response. One patient died of aGVHD following DLI. With a median follow up in survivors of 53 months (range 24-83 months), the 3-year PFS and OS are 41% (95%CI: 26-64%) and 54% (95%CI: 38-76%), respectively (Figure 1). At last follow up, 10 patients were alive.

Figure 1

Kaplan-Meier estimate of OS and PFS of the study cohort (n=29).

Analysis of pre-transplant factors revealed that in vivo T-cell depletion with alemtuzumab was associated with inferior 3-year PFS (0% vs. 51%, p=0. 0.007) (Figure 2A). Additionally, a sIPI at transplant equal to 0 was associated with an improved 3-year PFS (52% vs. 22%, p=0.020) compared to higher sIPI values (Figure 3A). Similarly, 3-year OS was significantly reduced with alemtuzumab (17% vs. 64%, p= 0.014) (Figure 2B) and also observed for patients with sIPI ≥ 1 (22% vs. 71%, p=0.006) (Figure 3B). Performing an allo-HSCT before 2007 was associated with a decreased 3-year OS (25% vs. 76%, p=0.015) but not with a significantly inferior PFS (17% vs. 59%, p=0.058). Of note, the use of ATG did not influence survival outcomes (3-year PFS 50% vs. 36%, p=0.671; 3-year OS 60% vs. 51%, p=0.579). The other factors analyzed were not significant on PFS and OS (Table 2).

Figure 2

Kaplan-Meier estimate of PFS (A) and OS (B) according to alemtuzumab use.

Figure 3

Kaplan-Meier estimate of PFS (A) and OS (B) according to sIPI.

Table 2

Univariate analysis of PFS and OS risk factors.

Variable	Value	Numberofpatients	PFS 3-yearestimate	P-value	OS 3-yearestimate	P-value
Age	<60	17	0.51 (0.32-0.83)	0.054	0.51 (0.31-0.83)	0.464
	≥60	12	0.25 (0.09-0.67)		0.58 (0.36-0.94)
Alemtuzumab use	No	23	0.51 (0.34-0.77)	0.007	0.64 (0.46-0.88)	0.014
	Yes	6	0		0.17 (0.03-0.99)
ATG oralemtuzumab	No	13	0.53 (0.31-0.89)	0.231	0.67 (0.45-0.99)	0.248
	Yes	16	0.31 (0.15-0.65)		0.43 (0.24-0.76)
HCT-CI	<3	24	0.45 (0.29-0.71)	0.576	0.53 (0.36-0.78)	0.68
	≥3	5	NA		0.6 (0.29-0.99)
PET positivebefore allo-HSCT	No	15	0.37 (0.19-0.75)	0.826	0.63 (0.41-0.97)	0.692
	Yes	9	0.44 (0.21-0.92)		0.44 (0.21-0.92)
Prior auto-SCT	No	16	0.5 (0.31-0.82)	0.629	0.61 (0.41-0.91)	0.363
	Yes	13	0.29 (0.12-0.7)		0.43 (0.22-0.84)
Prior lines oftherapy	<3	17	0.4 (0.22-0.73)	0.435	0.56 (0.35-0.87)	0.459
	≥3	12	0.42 (0.21-0.81)		0.50 (0.28-0.88)
Remission quotient	<11	16	0.38 (0.2-0.71)	0.865	0.49 (0.3-0.82)	0.669
	≥11	13	0.46 (0.26-0.83)		0.62 (0.4-0.95)
sIPI	0	19	0.52 (0.33-0.81)	0.020	0.71 (0.53-0.97)	0.006
	≥1	9	0.22 (0.07-0.75)		0.22 (0.07-0.75)
Time fromdiagnosis to allo-HSCT	<2 years	12	0.5 (0.28-0.88)	0.867	0.67 (0.45-0.99)	0.394
	≥2 years	17	0.35 (0.19-0.67)		0.46 (0.28-0.78)
Type of donor	Related	12	0.33 (0.15-0.74)	0.264	0.5 (0.28-0.88)	0.344
	Unrelated	17	0.45 (0.26-0.78)		0.56 (0.36-0.88)
Year of allo-HSCT	< 2007	12	0.17 (0.05-0.59)	0.064	0.25 (0.09-0.67)	0.017
	≥2007	17	0.59 (0.4-0.88)		0.76 (0.59-0.99)

PFS: Progression Free Survival; OS: Overall Survival; ATG: anti-thymocyte Globulin; HCT-CI: Hematopoietic Stem Cell Transplantation Comorbidity Index; PET: Positron Emission Tomography; HSCT: hematopoietic Stem Cell Transplantation; auto-SCT: autologous Stem Cell Transplantation; sIPI: second line International Prognostic Index;

DISCUSSION

In this study we report the MSKCC experience of NMA/RIC allo-HSCT for MCL and demonstrate a 3-year OS and PFS of 53% and 46%, respectively. This compares favorably with previous studies.[19-22] We also observed a 3-year incidence of NRM of 26% and 3-year incidence of POD of 30%, similar to previously reported series.[19-22]

Analysis of prognostic factors revealed interesting results. We found that in vivo T-cell depletion with alemtuzumab was associated with inferior PFS and OS, mostly attributable to POD. This can be explained by a reduced GVL effect with in vivo T-cell depletion. The importance of a GVL effect in MCL is demonstrated indirectly by the relapse risk associated with allografts performed with T-cell depletion. Morris et al. reported a relapse incidence of 50% at 3 years in a small cohort of patients with MCL treated with an alemtuzumab-based RIC regimen.[38] A recent study from the British Society for Blood and Marrow Transplant reported a significantly increased incidence of POD in 51 patients receiving an alemtuzumab-based conditioning regimen versus 17 non-alemtuzumab treated patients.[20] Perez-Simon et al. reported that patients treated with an alemtuzumab-based regimen required additional immunotherapy with DLI to achieve disease control compared to patients treated without alemtuzumab.[39] Alemtuzumab is a monoclonal antibody directed against the CD52 antigen present not only on lymphocytes but also on monocytes, offering explanation for more potent immunosuppressive properties compared to other methods of in vivo T-cell depletion. Alemtuzumab impairs not only the activity of lymphocytes but also dendritic cells,[40] and cell-mediated cytotoxicity of natural killer (NK) cells.[41] This likely compounds impairment of the GVL effect.[42, 43] It should be noted that we used a higher dose of alemtuzumab compared to other centers, potentially contributing a dose-dependent effect on the incidence of POD.[20]

In contrast, we did not observe an effect of ATG on PFS or OS. Le Gouill et al. reported a trend toward higher relapse rate when ATG was used in this setting; however the difference was not statistically significant.[19] The existence of a GVL effect in lymphoid malignancies has been previously demonstrated, at least for indolent histology lymphoma such as follicular lymphoma,[44, 45] and chronic lymphocytic leukemia.[46, 47] In MCL, DLI has been effective in reducing the incidence of POD after allo-HSCT in some series.[14, 38, 48, 49] Moreover, reduction of immunosuppressive therapy and presence of cGVHD have also been associated with disease control in this setting.[50] Tam et al. reported a reduced disease progression in MCL patients developing cGVHD after allo-HSCT (5% versus 46% at 6 years). A recent multicenter study from CIBMTR showed a significantly decreased risk of POD in patients with evidence of cGVHD after RIC allo-HSCT.[51]

Moreover, we observed a statistically significant association between PFS, OS and a sIPI equal to 0 before allo-HSCT. Second-line IPI has never been tested in MCL patients undergoing allo-HSCT; nonetheless a possible role of this prognostic factor in the context of allo-HSCT for lymphoma has already been reported.[32] Interestingly, a positive PET pre allo-HSCT was not associated with a worse PFS or OS compared to a negative PET. This is in line with a prior report from our group regarding pre allo-HSCT PET for indolent and aggressive lymphomas.[52] Finally, performing an allo-HSCT before 2007 was associated with a lower OS but not with an inferior PFS. This is probably associated with a general improvement in supportive care for allo-HSCT patients as already reported in literature.[53]

CMV reactivation, monitored with antigenemia before November 2011 and PCR thereafter, was similar to the incidence reported for conventional grafts. None of the patients developed CMV disease, including those treated with alemtuzumab where a higher risk is expected. A low EBV reactivation rate was consistent with what is observed in conventional grafts. Rituximab could have further reduced this incidence.

Considering the retrospective nature of the study we should consider the presence of potential biases. The small patient population likely impaired the ability to detect outcome difference in characteristics that have demonstrated significance in other studies such as MIPI,[54] blastic morphology,[55] Ki67 proliferative index.[54] The heterogeneity and the small size of the study population limited the power of the analysis. Interestingly, the majority of patients had a low MIPI or sIPI score suggesting a possible patient selection. Nevertheless, this is one of the largest single center studies evaluating MCL and allo-HSCT.

In conclusion, our study showed that RIC allo-HSCT is an effective curative strategy in patients with relapsed and refractory MCL. These findings would support other studies suggesting a GVL effect in MCL, and furthermore that T-cell depletion should be probably avoided. Moreover, considering that disease relapse after allo-HSCT is a major cause of death for patients affected by MCL, prospective studies should focus on safely decreasing the incidence of POD. This goal is potentially achievable with the growth of pathway specific targeted agents that could be used as post-allograft maintenance therapy.[5-10]