For clinicians, the most practical way to approach the treatment of non-Hodgkin's lymphoma is based on the clinical behavior of the disease. Hematopoietic stem cell transplantation (HSCT) has been the backbone of the therapeutic options for treating aggressive (intermediate- and high-risk) disease that has relapsed or does not respond to initial therapy. Its role in the treatment of indolent lymphomas has been more commonly applied in patients with multiply relapsed or refractory disease. In some cases, such as chemotherapy- sensitive, relapsed aggressive NHL, HSCT results in a durable remission and is superior to conventional salvage therapy.[1,2] There may be a role, however, for an earlier HSCT in patients with poor prognostic factors at the time of initial diagnosis. In this review, we will summarize the treatment of indolent and aggressive NHL with HSCT and comment on the evolving role of new hematopoietic stem cell transplant- based therapies. Indolent Non-Hodgkin's Lymphoma The role of HSCT in treating indolent lymphoma has been limited not only by the heterogeneity of the disease and its chronicity with many long-term survivors, but also by high rates of transplant-related mortality with allogeneic HSCT and high rates of relapse after autologous HSCT. Autologous Transplants
Autologous transplants in this setting have an observed regimenrelated toxicity rate that is similar to that seen with other HSCT-treated diseases. In most cases, the regimenrelated mortality rate is about 5% to 10%. However, most published studies report a continual relapse rate with no obvious plateau in the survival curves. Thus, autologous HSCT is not curative therapy for the majority of indolent lymphomas treated with current regimens.[3,4] Relapses are thought to be due to both the inability to eradicate the tumor cells and tumor contamination of the infused stem cell product. In Vitro Purging in Autologous Transplantation
Numerous groups, including Freedman et al,[5] have reported that patients who receive autologous grafts that are not contaminated with lymphoma cells by polymerase chain reaction (PCR) analysis have a statistically higher initial disease-free survival rate, compared to patients with PCR-contaminated stem cells. As a result, considerable effort has gone into studying the best way to purge grafts of tumor contaminant cells. The earliest such studies centered on positive selection of CD34 cells. Other approaches have focused on tumor elimination by antitumor antibodies and complement or chemotherapy exposure with drugs such as 4-hydroperoxycyclophosphamide or mafosfamide. Initially, improved disease-free survival was reported in patients who underwent autologous bone marrow purging with a cocktail of anti-B cell monoclonal antibodies.[6] Greater than a 3 log depletion of follicular lymphoma cells was achieved, and no lymphoma cells could be detected in 50% of treated patients. Patients who had PCR-detected residual lymphoma cells in their stem cell product were initially more likely to relapse posttransplant, with a relapse rate of 39% (vs 5% in PCR-negative patients) after a median follow-up of 23 months (P < .00001). With time, however, late relapses are being observed.[5,6] That said, none of these techniques addresses the disease in the patient. In vitro techniques are expensive, labor intensive, and in some cases delay engraftment. In addition, techniques such as CD34 selection have been reported to be associated with an increased risk of infection.[7,8] To date,none of these in vitro approaches have reduced the overall risk for relapse after HSCT, suggesting that residual disease left in the patient after autologous HSCT remains a major problem contributing to relapse. Indeed, Apostolidis et al[9] showed that achieving a minimal tumor burden in the patient prior to HSCT, rather than purging the graft alone, is the most important factor to determine longterm outcome. In Vivo Purging in Autologous Transplantation
Recently, interest has focused on in vivo purging. Antibodies such as rituximab(Drug information on rituximab) (Rituxan), given prior to high-dose chemotherapy, appear to increase the sensitivity of lymphoma cells to chemotherapy. Rituximab has been studied by numerous investigators, in many cases administered with mobilization regimens for collecting peripheral blood stem cells (PBSC) in mantle cell and indolent NHL patients.[10-12] In a small pilot study, Magini et al[11] showed that 93% of patients receiving rituximab with chemotherapy had a PCR-negative stem cell graft, compared to 40% of control cases. When granulocyte colonystimulating factor (G-CSF, Neupogen) alone is used for mobilization of PBSC with rituximab, two or more doses of rituximab should be given before PBSC collection to get the least contaminated product. Thus, several groups have shown that rituximab can render a stem cell product free of contaminating tumor cells by PCR while having no adverse effects on the collection of an adequate number of stem cells or on engraftment posttransplant. Rituximab has also been administered post-HSCT. Brugger et al[13] treated patients with follicular and mantle cell NHL with four weekly doses of rituximab after autologous HSCT. Following total-body irradiation and high-dose chemotherapy, the complete response rate was 44%, and after the addition of rituximab to the transplant regimen, this rate initially increased to 57% and continued to improve over time. By 1 year, it was 88%, and by 2 years, all follicularNHL patients and 90% of mantle cell lymphoma patients were in complete remission. Following high-dose therapy, 48% of evaluable patients had no evidence of disease by PCR. Immediately after rituximab therapy, this parameter improved to 80%, and at 6 months posttransplant, it was 100%. Leukopenia and infections were reported. Horwitz et al[14] studied the administration of rituximab alone posttransplant in more aggressive NHL or transformed B cell NHL, reporting grade 3/4 neutropenia in 9 of 20 patients. Flinn et al[10] also reported late infection problems, with three deaths in the first year posttransplant as well as neutropenia, disseminated herpes zoster, and atypical mycobacterial infections. Thus, rituximab therapy appears to be deliverable with good results in this setting. However, increasing evidence suggests that some patients (at least 25% to 45%)[10,13,14] also develop transient neutropenia after transplant, and there may be an increased risk of infection. Randomized controlled studies of rituximab use in an autologous transplant setting are lacking. In the future, other antibodies such as CD22 and CD40 will be similarly studied. It may be that the most effective therapy will involve a combination of antibodies, similar to what has been reported for in vitro purging. Targeted Therapy in Autologous Transplants
Another area of active research has been to give targeted therapy, ie, using radiolabeled antibody, combined with chemotherapy and followed by autologous stem cell rescue. This targeted radiotherapy is based on the fact that hematologic malignancies are sensitive to radiation. In addition, the antibody is not internalized, nor does it need to activate an immune response to generate an antitumor effect. Based on the isotope tagged to the antibody and its penetration, the radiolabeled antibody does not need to reach every malignant cell for it to be effective. Different radiolabeled antibodies to anti-CD20 have been used in an autologous transplant setting.[15-19] In a phase I/II study of iodine-131 tositumomab (anti-CD20, Bexxar), etoposide(Drug information on etoposide), and cyclophosphamide(Drug information on cyclophosphamide) (Cytoxan, Neosar), Press et al[17] reported that the maximum tolerated dose of the anti-CD20 monoclonal antibody was 25 Gy, with etoposide, 60 mg/kg, and cyclophosphamide, 100 mg/kg. The reported time to engraftment and toxicity data were similar to historical results with total-body radiation, etoposide, and cyclophosphamide therapy. Overall survival at 2 years was 83%, and the progression-free survival rate at 2 years was 68%. Approximately 21% developed a human antimurine antibody. The results have been encouraging, but which radioisotope and regimen will prove to be the most effective and practical to deliver remains unknown. Many of the radiolabeled antibody studies published to date have been restrictive in their eligibility requirements, for example, requiring no evidence of enlarged spleen and low tumor burden of no more than 500 cc. Randomized studies are needed to show an increased benefit in patients, for example, with increased survival and disease-free survival associated with the radiolabeled antibody- containing regimens. Longterm follow-up is also needed to address the issue of whether secondary malignancies are more likely to occur with intensified radiolabeled targeted therapy. Unique Problems With Autologous Transplant
Unique issues arise with respect to transplanting low-grade lymphoma. Fludarabine (Fludara) is commonly used as conventional therapy to treat the disease, and its use may affect the ability to mobilize stem cells.[20] Autologous transplants are also problematic in this setting, given the high rate of secondary cancers reported, including myelodysplastic syndrome (MDS)/acute myelogenous leukemia (AML), for which incidence rates are as high as 20%.[9,21] Timing of Autologous Transplant
Recently, investigators have placed much emphasis on moving autologous HSCT up earlier in patients with poor prognostic factors. Colombat et al[22] treated 29 patients, with 7 in first complete remission at a median follow-up of 6 years. The overall survival was 64%, and the actuarial event-free survival was 55%. Tarella et al[23] treated 46 patients with advanced low-grade NHL; 17 had small lymphocytic lymphoma, 29 had follicular lymphoma, and 10 also had histologic transformation. Patients received tumor debulking by two courses of APO (doxorubicin [Adriamycin], prednisone(Drug information on prednisone), vincristine [Oncovin], methotrexate(Drug information on methotrexate), asparaginase(Drug information on asparaginase) [Elspar], mercaptopurine(Drug information on mercaptopurine) [Purinethol]) and two courses of DHAP (dexamethasone, high-dose cytarabine(Drug information on cytarabine) [Ara-C], cisplatin [Platinol]), sequential administration of highdose etoposide, methotrexate, and cyclophosphamide with stem cell harvest, and high-dose mitoxantrone(Drug information on mitoxantrone) (Novantrone) and melphalan(Drug information on melphalan) (Alkeran) with PBSC infusion. Ten follicular lymphoma patients had ex vivo purging of stem cells. At a median follow-up of 4.3 years, the estimated 9-year overall survival was 84% and progression-free survival was 45%. Follicular lymphoma patients had longer survival without evidence of disease-59% vs 17% for small lymphocytic lymphoma patients. These results indicate that there may be longer progression-free survival after autologous transplant if it is part of the up-front therapy. Thus, we need to address the issue of whether a randomized study should be conducted in patients with intermediate and high scores on the International Prognostic Index (IPI) of risk factors, comparing the addition of a front-line autologous HSCT after conventional therapy. Indeed, front-line autologous transplant may be very relevant for the long-term benefit of patients. As oncologists, we tend to treat patients with multiple cycles of different therapeutic regimens and have in the past relied heavily on alkylating agents. As we treat patients more intensely for longer periods of time, we can expect to see more secondary cancers-especially MDS/AML. The use of autologous transplant upfront and the decreasing use of multiple regimens of therapy initially may prolong the time to the development of MDS. If we conduct such studies in the future, we need to use molecular correlating studies (eg, microarray assays) to help identify patients that do benefit from upfront transplants. Immunotherapy and Biologic Modifier Therapy After Autologous Transplant
Indolent NHL may be a better target for immunotherapy approaches, in both autologous and allogeneic settings. Several different approaches are being used to study the addition of immunotherapy to an autologous transplant to reduce relapse rates. These strategies include treatment with immune stimulators such as dose-intensive interleukin-2 (IL-2, Proleukin) therapy,[24] IL-2-incubated stem cells with sequential IL-2,[25,26] and low-dose IL-2 with or without rituximab.[27] In indolent lymphoma, vaccination with idiotypespecific vaccines after transplant is an alternative approach to be further studied. Again, no randomized studies have shown evidence of efficacy with immunotherapy in this setting, but the randomized Southwest Oncology Group study of dose-intense IL-2 after an autologous transplant is still ongoing. Other biologic modifiers such as bcl-2 antisense for maintenance therapy after autologous transplant also need to be investigated. Myeloablative and Nonmyeloablative Allogeneic Transplants
Allogeneic transplants offer the advantage of a "clean" stem cell product and a graft-vs-lymphoma effect. With a myeloablative allogeneic transplant, there is the risk of developing graft-vs-host disease (GVHD) and a high regimen-related mortality rate. International Bone Marrow Transplant Registry data for myeloablative regimens shows a transplantrelated mortality rate of 40% to 50%, with an event-free survival rate of 49%. The high regimen-related mortality rate has limited the use of myeloablative allogeneic transplants. It should be noted, however, that there is a plateau in survival in these allogeneic transplant recipients. Because of the high associated upfront mortality rate, myeloablative allogeneic transplants have never shown a survival advantage. In particular, patients with a history of multiple therapeutic regimens also seem to have an increased risk of complications. T-cell depletion or antithymocyte globulin (Thymoglobulin) therapy is being used by many centers to cut down on GVHD incidence, but these approaches usually result in higher replase rates and may not be the answer unless further manipulations of the stem cell infusion allow us to safely and effectively separate the cells that cause GVHD from those that provide the graft-vs-lymphoma effect. The intent of myeloablative allogeneic transplant is to use high-dose therapy to help eradicate the underlying disease, prevent graft rejection, and produce a graft-vs-lymphoma response to maintain disease control. In indolent NHL, a graft-vslymphoma effect is documented by lower relapse rates after an allogeneic transplant and by the evidence of tumor control by infusion of donor lymphocyte cells following a myeloablative transplant. Thus, if the risk associated with a myeloablative allogeneic transplant can be lowered, the lower risk for disease progression may eventually lead to a superior outcome. The source of donor stem cells appears to make a difference in initial mortality rates, with PBSC being superior to bone marrow.[28] Alternatively, nonmyeloablative transplants have been studied based on the theory that the most important part of an allogeneic transplant is the donor cell graft-vs-lymphoma effect. Khouri et al[29] treated 15 patients, 11 of whom had engraftment of donor cells and 8 of the 11 who achieved a complete remission. Five of six patients (83.3%) with chemotherapysensitive disease have survived, compared with two of nine (22.2%) with refractory or untested disease (P = .04). Thus, patients with lower tumor burden and chemotherapy-sensitive disease may be more effectively treated with a nonmyeloablative approach and have the longest duration of responses. However, caution is needed in selecting patients for nonmyeloablative therapy. In many diseases, higher doses of radiation and chemotherapy have been associated with a reduced risk of relapse.[ 30] Nonmyeloablative therapy probably needs to be considered mainly in patients with a diagnosis that is very sensitive to the graft-vslymphoma effect, older patients, and patients with comorbid medical problems. Indolent NHL may be a good target for reduced-intensity nonmyeloablative allogeneic stem cell transplant.[31] Tandem transplants with autologous transplant followed by nonmyeloablative allogeneic transplant has also been studied.[32] In one trial, 11 of 13 patients achieved a complete response postallograft, including 9 patients with a partial response after the autograft. Seven also received additional donor lymphocytes. Seven patients developed acute GVHD (grade 2-4) and two developed chronic extensive disease. Between 210 and 340 days postallografting, 2 patients have relapsed, 10 are alive, and 5 are in complete remission. Five have died-two from GVHD and progressive disease, two from GVHD and infection, and one from disease progression. Chronic Graft-vs-Host Disease
Chronic GVHD will remain a problem after a nonmyeloablative transplant. It therefore behooves us to remember that in a myeloablative setting, significant chronic GVHD is associated with a 50% nonrelapse mortality rate.[33] The pathophysiology of GVHD is poorly understood and needs to be further studied. The identity of the antigenic targets for the immune reactive cells have not been well clarified; nor is it clear which populations of cells mediate the ongoing immune responses seen in chronic GVHD. In a coisogenic murine model, T cells from donor mice transplanted into mice with only a three-amino acid difference in their Dr molecule developed chronic GVHD.[34] Some investigators believe that recipient alloantigens provide the stimulus for the graft-derived T cells that have already undergone selection and maturation in the donor thymus environment, causing GVHD. Alternative explanations have included flawed T-cell reconstitution, dominant autoantigens driving the system, and improper thymic selection of de novo donor cells undergoing maturation in the new host that does not result in tolerance. Many of our therapies for chronic GVHD, though, only control disease but do not delete the pathogenic clone. The incidence of GVHD appears to be similar between nonmyeloablative and myeloablative transplant recipients, although long-term follow-up is lacking.[35] Older patients have decreased thymus function, and thus, as we age we generate an environment more conducive to the development of chronic GVHD. The preferred source for nonmyeloablative transplants studies are PBSCs, which are associated with a higher risk of developing chronic GVHD, a longer disease duration, and a greater amount of therapy required to treat GVHD.[36] Also, many nonmyeloablative transplants depend on additional donor lymphocyte infusions for disease control. In general, 80% of patients who receive donor lymphocytes and respond develop GVHD. Better understanding of graft selection and the mechanism of chronic GVHD, as well as alternative approaches to treating GVHD that do not affect lymphoma control, are needed. Optimizing Allogeneic Transplant Therapy
Interest is also beginning to focus on optimizing immune responses against lymphoma cells in an allogeneic transplant. Numerous laboratories are studying the generation of minor histocompatible antigen-specific clones to treat residual disease after allogeneic transplant for a hematologic malignancy. However, the concept of using restricted antigenspecific T-cell clones as effective treatment after transplant may be flawed. The reason that the graft-vstumor effect is so prominent in an allogeneic setting most likely is due to its polyclonal recognition of tumor cells that results in disease control. Limited recognition of tumor cells by antigen-specific clones will most likely lead to escape of recognition of the tumor cells by various mechanisms, and will not result in effective long-term disease control. Further investigation of the ability to amplify the polyclonal tumor response after an allogeneic transplant,to prevent relapse and to determine which patients require this intervention, needs to be undertaken. In addition, what is lacking in humans is the ability to separate the cells that cause GVHD from the cells that cause a graft-vs-tumor effect, and we need to study the differences in the antigen recognition repertoire of these cells. Summary
The optimal therapy for indolent lymphoma and the timing of various transplant entities remains unclear. It is conceivable that early autologous transplant for patients with intermediate and poor IPI risk factors may maximize survival and quality of life, with nonmyeloablative allogeneic transplants being left to later, when disease recurs. Quality of life in patients with chronic GVHD is poorly studied in the setting of nonmyeloablative transplants, and for patients with indolent NHL, quality of life is more of an issue early in the disease than disease eradication, thus making this type of approach an interesting one for investigation in future randomized trials. Indolent disease is more likely to be a good target for nonmyeloablative immunotherapy than aggressive disease that will quickly outgrow the ability of the immune system to control it. How much bulky low-grade NHL tumor burden patients can have at the time of nonmyeloablative transplant is still poorly defined, but common sense and previous experience would argue that chemotherapy-sensitive disease and a lower tumor burden are associated with the best long-term outcomes.