|Home | About | Journals | Submit | Contact Us | Français|
Autologous blood or marrow transplantation (ABMT) for low-grade lymphomas can prolong event-free survival but requires long-term follow-up. We report one of the longest follow-ups to a prospective transplantation study in such diseases. On a phase II study, 80 patients with low-grade, transformed, or mantle cell lymphoma received ABMT with 4-hydroperoxycyclophosphamide (4-HC) purging as part of initial or salvage therapy. Diagnoses included nontransformed follicular lymphoma in 63% and transformed lymphoma in 15%. With 16.6 year median follow-up for survival, actuarial 10-year event-free and overall survivals were 34% (95% confidence interval, 25 – 46%) and 45% (35 – 57%). Median event-free and overall survivals were 3.0 and 8.0 years. Early non-relapse mortality incidence was 8%; myelodysplasia or leukemia incidence was 4%. Most relapses occurred within 3 years, with a median time to diagnosis of relapse of 1.8 years (range, 0.1 – 15.6 years). On multivariate analysis, age > 50, ≥ 3 prior chemotherapy regimens, and ABMT after relapse were associated with significantly inferior survival. Fifteen patients (19%) were event-free > 15 years after transplantation, raising the possibility of a plateau in the progression-free survival curve. Thus, 4-HC-purged ABMT can produce extended remissions in a subgroup of patients with indolent lymphomas.
High-dose therapy with autologous blood or marrow transplantation (ABMT) can prolong event-free survival (EFS) in patients with follicular lymphoma [1-4]. Numerous retrospective and prospective studies of ABMT for indolent lymphomas have been published. Although a few studies have suggested a plateau in the progression-free survival curve [4-6], others have not , casting doubt on its curative potential. Given the long natural history and continuous relapse pattern that typify this group of diseases, mature data are required in order to assess the overall therapeutic impact.
Both lymphoma surviving the conditioning regimen and lymphoma re-introduced at the time of BMT appear responsible for relapse [8-10], although their relative contributions are unknown. Tumor cells commonly can be found circulating in the blood of patients with lymphoid malignancies [11,12]. Moreover, syngeneic BMT and purged ABMT appear to be associated with lower relapse rates than unpurged ABMT [5,9,10], further highlighting the significance of autograft contamination. Diverse strategies have been used to clear or purge the autograft of tumor cells, including positive stem cell selection, in vitro treatment with monoclonal antibodies or cytotoxic drugs, and in vivo monoclonal antibodies such as rituximab [9,13], although the benefit of graft purging remains unproven [2,14,15]. The cyclophosphamide congener 4-hydroperoxycyclophosphamide (4-HC), and the related compound mafosfamide, have been widely used as cytotoxic purging agents. Inactivation of cyclophosphamide occurs almost completely via tissue aldehyde dehydrogenase (ALDH1), which converts the metabolic pathway intermediate aldophosphamide to the non-cytotoxic metabolite carboxyphosphamide . Normal engrafting hematopoietic stem cells are protected against 4-HC by their high expression of ALDH1 . We report the results of a single-institution, phase II study of 4-HC purged ABMT for patients with a history of indolent or mantle cell lymphoma. With a median follow-up of over 16 years in surviving patients, this is one of the longest prospective series reported in this group of diseases to date.
Between 1989 and 1994, 80 adults with biopsy-proven non-Hodgkin's lymphoma received ABMT on a phase II clinical trial (J8913) at Johns Hopkins University. The study was approved by the Johns Hopkins Institutional Review Board, and all participants gave written informed consent. The following diagnoses were eligible: nodular poorly differentiated, mixed, or large cell lymphoma (follicular lymphoma, grades 1-3, respectively); diffuse intermediate differentiated lymphoma (mantle cell lymphoma); diffuse well-differentiated lymphoma including chronic lymphocytic leukemia (chronic lymphocytic leukemia /small lymphocytic lymphoma); or diffuse poorly differentiated (diffuse small cleaved cell) lymphoma (some cases of marginal zone and mantle cell lymphomas). Transformed lymphomas were not excluded. Additional requirements included minimal or no apparent disease at the time of BMT; ≤ 10% morphologic bone marrow involvement by lymphoma; white blood cell count ≥ 3000/μL; platelet count ≥ 100,000/μL; creatinine ≤ 2.0 mg/dL; total bilirubin ≤ 2.0 mg/dL; ejection fraction ≥ 45%; and forced vital capacity or forced expiratory volume in 1 second of ≥ 75% predicted.
Pathology specimens in each case were reviewed by members of the Johns Hopkins Department of Pathology prior to study entry. Diagnoses were translated into current World Health Organization terminology by re-review of all available pathology reports, with slide re-review in selected cases.
Autografts were derived from bone marrow harvesting. A minimum of 2 × 108 nucleated bone marrow cells/kg with a target of 4 × 108 nucleated bone marrow cells/kg was collected. The cells were treated ex vivo with 4-HC (60 μg/mL for 30 minutes) and cryopreserved until infusion. The methods for marrow processing, 4-HC purging, and infusion have been previously published [17,18].
The preparative regimen consisted of either cyclophosphamide (50 mg/kg/d iv for 4 days with uroprotection) and total body irradiation (TBI; 300 cGy/day for 4 days), or busulfan (16 mg/kg orally over 4 days, with pharmacokinetic adjustments)  and cyclophosphamide (50 mg/kg/d iv for 4 days) . Dosing was based on the lesser of ideal or actual body weight. Busulfan-cyclophosphamide was initially given in those unable to receive TBI and later became the institutional preference. The autograft was infused on the day after TBI or, following busulfan-cyclophosphamide conditioning, two days after cyclophosphamide (day 0).
Supportive care was delivered according to institutional standard or approved research protocols and included prophylactic vancomycin, norfloxacin, fluconazole, acyclovir (if herpes simplex virus seropositive), and trimethoprim-sulfamethoxazole or dapsone. Fungal infections were treated with Amphotericin B with or without 5-flucytosine. The general platelet goal was ≥ 20,000/uL in the absence of bleeding or anticoagulation. Ten patients received three weeks of recombinant human interleukin-3 beginning day 0 for augmentation of count recovery. Granulocyte or granulocyte-macrophage colony stimulating factor was not routinely given.
Primary induction failure was defined as nonresponse to first-line therapy or disease progression within two months of its completion. Lymphoid aggregates in the bone marrow or an atypical marrow lymphocytosis without immunologic proof of clonality was categorized as suspicious for involvement by lymphoma. Retreatment with the same chemotherapy regimen, or a change in regimen due to nonresponse, was regarded as a separate therapy; a change for other reasons in a responding patient or combined modality therapy was considered a single regimen. Remission refers to complete or partial remission. Neutrophil recovery was defined as the first of three consecutive days with an absolute neutrophil count of ≥ 500/uL. Platelet recovery was defined as the first day with a platelet count of ≥ 20,000/uL, measured at least seven days from platelet transfusion.
Dates and causes of death were determined from chart and study database review and searches of the Social Security Death Index and National Death Index. For the purpose of this analysis, deaths that occurred after relapse or progression were attributed to lymphoma unless otherwise noted. In three such patients the date of relapse was unknown and was imputed using the median time between relapse and death in the other patients.
EFS and overall survival (OS) rates with corresponding 95% confidence intervals (CI) were calculated from the date of transplantation using the Kaplan-Meier method. An event was defined as relapse or progression, myelodysplastic syndrome (MDS) or acute leukemia, or death from any cause. Radiographic confirmation of remission was not required given the lack of routine imaging beyond 5 years after BMT. Univariate Cox proportional hazards models were used to determinate associations between clinical variables and survival, with variables having a P < 0.10 then included in a multivariate Cox model that included a term to adjust for the time from diagnosis to BMT. A hazard ratio (HR) > 1 indicates an increased risk of failure associated with having the variable relative to its reference category.
Early non-relapse mortality (NRM) was defined as nonrelapse death within 101 days after BMT. Cumulative incidences of relapse or progression, NRM, and MDS or leukemia were estimated using competing-risks analyses ; death without relapse or progression was considered a competing risk for relapse or progression, relapse or progression was a competing risk for NRM, and death was a competing risk for MDS or leukemia. Prognostic factors for competing-risk endpoints were analyzed using proportional hazards models for competing risks . All P-values are two-sided. Statistical analyses were performed with R, version 2.8.0  and represent data through March 22, 2010.
Characteristics of the patients and transplants (n = 80) are summarized in Table 1. The most common diagnosis was follicular lymphoma without transformation (63%), with transformed lymphomas representing 15% of cases. Similar proportions received BMT in first or second remission (36% and 38%, respectively) with 21% being in third or greater remission. Over two-thirds (68%) received busulfan-cyclophosphamide conditioning, with the remainder receiving cyclophosphamide-TBI.
The median time to neutrophil recovery was 29 days (range, 16 – 71 days); in one case recovery date was unknown, and four additional patients had no or no confirmed neutrophil recovery before death or relapse. In patients whose neutrophil recovery was documented, the median time to neutrophil recovery was not significantly different in those transplanted in first remission versus after relapse (29 and 30 days, respectively; P = 0.22). The median time to platelet recovery was 37 days (range, 25 – 70 days); in twelve patients the recovery time was unknown because of transition to outside follow-up, and nine additional patients had no or no confirmed recovery before death or relapse with a median time to event of 58 days (range, 35 – 360 days).
Nonfatal, unexpected toxicities in the early post-transplant period included several cases of significant upper gastrointestinal bleeding, an intracranial bleed, and a self-limited case of idiopathic, severe denervating polyneuropathy consistent with Guillan-Barré syndrome. Severe pulmonary toxicity occurred in four patients; this was fatal in two patients and possibly contributed to death in the two others. There were five documented cases of hepatic veno-occlusive disease, one of which was fatal.
Causes of death after transplantation are summarized in Table 2. Early transplant-related mortality occurred in six patients. Five additional deaths within the first year were related or possibly related to transplantation, all in patients transplanted beyond first remission. The cumulative incidence of NRM was 14% in the first year, with an 8% incidence of early NRM. Three cases of MDS or acute myeloid leukemia were identified (two after busulfan-cyclophosphamide and one after cyclophosphamide-TBI), with a cumulative incidence of 4% (95% CI, 0 – 9%); these were diagnosed at 5.0 years, 6.4 years, and 14.4 years after BMT and occurred in patients transplanted in first remission without subsequent relapse of lymphoma.
Disease and survival outcomes for the group overall are shown in Figure 1. At last assessment, 28 patients (35%) were alive, including 9 whose disease relapsed or progressed after BMT. Forty-one patients had relapsed or progressed. Fifty-two patients had died, 32 after having relapsed or progressive lymphoma (Table 2).
With a median follow-up of 16.6 years in patients without events (range, 6.7 – 20.1 years) and 3.5 years (range, 0.1 – 20.1 years) overall, the actuarial EFS at 5 years and 10 years after BMT was 45% (95% CI, 35 – 57%) and 34% (25 – 46%) respectively; the median EFS was 3.0 years (Figure 1A).
Of the 28 surviving patients, all but one (96%) had greater than 10 year follow-up for vital status after transplantation, and 25 patients (89%) had greater than 15 year follow-up. With a median follow-up of 16.6 years in surviving patients (range, 6.7 – 20.1 years) and 7.7 years (range, 0.1 – 20.1 years) for all patients, the actuarial OS at 5 years and 10 years after BMT was 57% (95% CI, 48 – 69%) and 45% (35 – 57%) respectively; the median OS was 8.0 years (Figure 1A).
The estimated cumulative incidence of relapse or progression was 39% (95% CI, 28 – 50%) at 5 years and 48% (95% CI, 37 – 59%) at 10 years on competing-risk analysis (Figure 1B). Of the 38 patients whose relapse date was confirmed, the median time to diagnosis of relapse after BMT was 1.8 years (range, 0.1 – 15.6 years). Twenty-eight of these patients (74%) had relapse identified within the first 3 years. Nine patients (one with mantle cell lymphoma, the remainder with follicular low-grade lymphoma) relapsed more than 7 years after BMT, and 3 (all with follicular low-grade lymphoma) relapsed more than 12 years after BMT. The median time from diagnosis of relapse after BMT to death was 1.7 years (range, 0.1 – 12.1 years).
Disease-specific survival estimates are shown in Figure 2 and Table 3. Of the 19 patients who were event-free, 9 were transplanted in first complete or partial remission and 14 had a documented history of follicular low-grade lymphoma including 3 with transformations (Figures 2A through 2D). Of the 12 patients with transformations, all from follicular low-grade lymphoma, 3 were event-free at 15.6 – 17.8 years after BMT, 2 had nonrelapse deaths (at 1 year and 10.3 years), and the remainder died of lymphoma at 0.1 – 4.3 years after BMT (Figures 2E and 2F). Of the five mantle cell lymphoma patients, one remained event-free at 18.4 years after BMT and four relapsed at 1.6 – 9.6 years. Fifteen patients (19%) were event-free more than 15 years after BMT, of which nine were transplanted in first remission, five in second remission, and five in third or greater remission.
A post-hoc analysis of variables in relation to transplantation outcomes is presented in Table 4. On univariate analysis, no significant association was found between survival outcomes and time from diagnosis to BMT, having transformed lymphoma, having bone marrow involvement prior to transplantation, or receiving busulfan-cyclophosphamide versus cyclophosphamide-TBI conditioning. On unadjusted analysis, age > 50 years at BMT was associated with a tendency toward greater risk of relapse or progression (HR 1.73, 95% CI 0.9 – 3.32, P = .10) without a significantly greater risk of NRM (HR 1.48, 95% CI 0.61 – 3.6, P = .38). On multivariate analysis, age > 50 years was associated with a statistically significantly inferior event-free and overall survival, as was having 3 or more prior chemotherapy regimens and BMT after relapse. The results of univariate and multivariate analyses of event-free and overall survival were similar in patients with follicular lymphoma inclusive of transformations (data not shown); this included, on multivariate analysis, age and number of prior chemotherapy regimens being significantly associated with event-free and overall survival, and BMT after relapse being associated with a significantly inferior EFS (HR 3.83, 95% CI 1.23 – 11.1, P = .02) and tendency toward inferior OS (HR 2.56, 95% CI 0.82 – 8.33, P = 0.11) compared with BMT in first complete or partial remission.
We report encouraging long-term outcomes of patients with B-cell lymphoma treated with high-dose therapy and 4-HC-purged ABMT. With a median follow-up of over 16 years in surviving patients, this is one of the most mature transplantation series reported in this group of diseases.
Most cases of relapse or progression were identified within the first 3 years. The EFS (45% at 5 years) is similar to that of other studies [24-26]. Few relapses were diagnosed more than 10 years after transplantation although one relapse was diagnosed at more than 15 years. The occurrence of late relapses in our study underscores the importance of long-term data when assessing transplantation outcomes for these histologies.
It is recognized that a fraction of such patients can enjoy long-term clinical remissions after conventional-dose or intensified therapies. Some mature studies of ABMT [4-6,24,26] have reported a plateau or a possibility of plateau in the disease-free survival curve for follicular lymphomas, but not others . The pattern of relapse might be best thought of as biphasic with an early high risk period followed by a much lower risk late period . If a biphasic model is appropriate, our results suggest the possibility that many, or perhaps most of the 19% of patients who remain in continuous complete remission more than 15 years after transplantation will ultimately die of causes other than lymphoma. The favorable outcome of such patients may be due to an inherently better biology of disease (although a subset were transplanted after one or more relapses), a beneficial impact of dose intensity, or a combination thereof.
Limitations of this study include the biases inherent to a single-institution study and the heterogeneity of diagnoses and prior therapies. The latter has been addressed in part by a separate analysis of follicular lymphoma outcomes. For follicular lymphoma patients and the group overall, having fewer than 3 prior chemotherapy regimens was associated with improved outcomes, as was transplantation in first remission. This is consistent with a number of other studies [5,7,27], as is the observed negative impact of older age [5,7,9,25,27]. However, in a more recent phase II trial of ABMT with pre- and post-transplantation rituximab for low-grade or mantle cell lymphomas, we found no statistically significant difference with transplantation in first remission versus after relapse . Several randomized trials of high-dose therapy and ABMT [1-4], but not all , have demonstrated prolongation of EFS in subgroups of patients with follicular lymphoma. The observed impact on OS has been variable and may depend on the timing of transplantation, as randomized trials have found a survival benefit in patients with follicular lymphomas in the relapsed  but not frontline setting [1,4,29]. Because patients are often followed clinically rather than radiographically years after transplantation, some patients believed to be disease-free may have asymptomatic active disease.
Our one-armed prospective study of bone marrow purging and high-dose therapy in consecutive patients finished accrual before the rituximab era. As such, it cannot address whether and how this ABMT strategy compares to transplant or non-transplant strategies in a similar patient population treated with more contemporary regimens. The study does, however, shed light on issues that may help to guide or interpret future therapies. First, concern has been raised that in vitro purging may increase the rate of secondary hematologic neoplasms . Nevertheless, high-dose cyclophosphamide treatment, including treatment of marrow cells in vitro, was not associated with an alarming incidence of secondary MDS or leukemia in this study; our observed incidence (4%) is in keeping with or lower than in other series . Short exposures to cyclophosphamide or its metabolites even at high doses may be of less concern than chronic exposures. This study with long-term follow-up that includes lightly pretreated patients may provide an upper bound for the risks of MDS and acute leukemia associated with high-dose therapy. Second, a substantial subset of patients achieved long-term clinical remissions including patients with transformed lymphoma. As therapies that include maintenance of various sorts are evaluated in the treatment of such patients, these results provide one of the longest-term follow-ups of ABMT without the use of monoclonal antibody therapy.
The impact of autograft purging remains undefined. 4-HC purging may have contributed to delayed engraftment in some patients in our study. The time to count recovery is similar to other studies of in vitro-purged BMT [30,31]. Although polymerase chain reaction has been used to assess physical removal of tumor cells from stem cell grafts such as is achieved with column separation or monoclonal antibodies, this approach is not readily applicable to the assessment of the effects of chemotherapeutic agents since cell death may not be immediate [32,33]. In early studies of leukemia and lymphoma, fewer colony forming unit-granulocyte-macrophages (CFU-GM) postpurging, as a measure of 4-HC or mafosfamide activity, were associated with improved disease outcomes [15,33]. Some other studies of lymphoma have also indirectly suggested a benefit to purging [9,15,34]; others have not [2,14]. Analyses of autograft purging are partly confounded by the inclusion of numerous different methodologies which may or may not have similar efficacies. It is well recognized that eradication of residual disease from the autograft is prognostically significant [35,36]; this could reflect more favorable disease biology, a true benefit of the purging procedure, or both.
These results demonstrate that very durable clinical remissions could be achieved in a subset of patients after high-dose therapy and ABMT. Although the impact purging played in the outcome will remain unclear, these data do suggest that ABMT may produce extended disease-free survival and possibly cure in a subset of patients with indolent lymphomas.
Supported by the National Institutes of Health (P01 CA015396 to R.J.J., K23 CA124465 to Y.L.K.) and National Cancer Institute Lymphoma SPORE (P50 CA09688 to R.F.A.). Presented in part at the 2003 American Society of Hematology Annual Meeting.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The authors declare no competing financial interests.