Salvage Therapy for CLL and the Role of Stem Cell Transplantation

Patients with chronic lymphocytic leukemia (CLL) who relapse following previous treatment with alkylator agents and are then re-treated with alkylator combinations have overall response rates ranging from 22%–62%, but most responses are of short duration, with few patients achieving complete remission (CR) (Table 1 ).^1,–5 Response rates to purine analogs in these circumstances are higher and of longer duration. In the Intergroup study comparing primary treatment with chlorambucil and fludarabine, at crossover, those patients who failed chlorambucil had an overall response rate to fludarabine of 46%, whereas only 7% of patients who failed fludarabine responded to chlorambucil, suggesting there is little role for alkylator therapy in patients who have failed previous purine analogs.⁶ 

A large number of trials have investigated the response to fludarabine in previously treated CLL, with overall response rates ranging from 13% to 59% and CR rates of 3% to 37% ^5,7,–12 Similar response rates have been seen in patients treated with the other nucleoside analogs 2-chlorodeoxyadenosine^13,14 and deoxycoformycin (Table 1 ).^15 –17 

Although the purine analogs induce higher overall response rates and longer duration of remission than alkylator-based therapies, these agents are not curative and patients will relapse and eventually become refractory to therapy. The molecular mechanisms whereby CLL cells become resistant to chemotherapy are poorly understood, but mutations and deletions of p53 become more frequent as the disease progresses, rising from 7% at diagnosis ¹⁸ to almost 50% by the time patients have become fludarabine refractory.¹⁹ Regardless of the mechanism, once patients become refractory to both alkylator agents and nucleoside analogs, outcome is poor, with median survival of less than one year.²⁰ A number of strategies have been explored to attempt to improve the outcome in these patients

The rationale for combining nucleoside analogs with alkylators includes the activity of both classes of agents in this disease, the lack of overlapping toxicity and the demonstration of synergy between these groups of agents. The mechanism of synergy appears to be induction of DNA damage by alkylator agents with subsequent inhibition of DNA repair by the nucleoside analogs. Results of combination chemotherapy in previously treated CLL patients are shown in Table 2 . A number of studies have examined the use of the combination of fludarabine with cyclophosphamide in previously treated patients, with the largest series reported from the MD Anderson Cancer Center.²¹ Although the CR rate overall with this regimen was 25%, this decreased to 15% in patients previously treated with alklyating agents, and was only 3% in patients already refractory to fludarabine. The major toxicity of the combination was myelosuppression. Administration of cyclophosphamide at 300 mg/m² daily for three days appears to reduce the risk of myelosuppression without compromising the efficacy of cyclophosphamide in combination with fludarabine. Cyclophosphamide has also been used in combination with deoxycoformycin and shown activity in previously treated patients with CLL.²² Addition of anthracycline to fludarabine and cyclophosphamide appears to increase the response rate in relapsed CLL patients with CLL. The combination of fludarabine, cyclophosphamide and mitoxantrone (FCM) produced CR in 30%–50% of previously treated patients and the use of this combination allowed harvesting of peripheral blood stem cells.²³ 

Very modest activity is seen in previously treated CLL when rituximab is administered at standard doses. When the drug was used at higher doses or at three times weekly dosing schedules, the overall response rate increased to 40%–45%, but few patients achieved CR. ^24,25 In the treatment of CLL, more interest has been focused on the use of rituximab in combination with chemotherapy. Rituximab appears to downmodulate expression of anti-apoptotic factors, making the CLL cells more susceptible to chemotherapy. Encouraging results of fludarabine in combination with rituximab have been reported in previously untreated patients, but in the relapsed setting most experience has been with the combination of fludarabine, cyclophosphamide and rituximab (FCR).

In 177 previously treated CLL patients the overall response rate to FCR was 73%, with 25% achieving CR, irrespective of whether the patients had previously received any of these agents either alone or in combination (Table 3 ). Patients who were fludarabine refractory still had overall response rates of 58%, but the CR rate was decreased to 6%.²⁶ On multivariate analysis the pre-treatment factors significantly associated with achieving complete remission were higher platelet count, fewer numbers of prior treatments and lower β2-microglobulin (β2M), whereas treatment failure was associated with fludarabine refractoriness and renal impairment. The median follow-up of all patients was 28 months, and median time to progression for patients achieving CR, nodular complete remission and partial remission was 39, 33 and 15 months, respectively. The estimated overall survival is 42 months, and on multivariate analysis the factors predictive of poor survival were failure of response to therapy, elevated β2M and adverse cytogenetics.

Alemtuzumab (Campath-1H), a humanized anti-CD52 monoclonal antibody, is approved for CLL patients who are fludarabine refractory and in the pivotal study of alemtuzumab in 93 patients with fludarabine refractory CLL, the overall response rate was 33%, although only 2% of patients achieved CR.²⁷ The median time to progression for responders was 9.5 months, and there was an improvement in survival in responding patients at 32 months, compared to 16 months for all patients. Of note, alemtuzumab appears to have activity against cases that are unresponsive to chemotherapy due to the presence of p53 mutations.²⁸ This finding was confirmed in a study of 36 patients with fludarabine-refractory CLL treated with alemtuzumab, 15 (42%) of whom had p53 mutations or deletions.¹⁹ Alemtuzumab has poor activity against bulky lymphadenopathy, but demonstrates impressive efficacy in clearing the peripheral blood and bone marrow compartments of disease. Studies are examining the role of alemtuzumab treatment after chemotherapy to clear residual disease and this may be an effective way to obtain autologous peripheral blood stem cells free of contamination with CLL.

Alemtuzumab has shown benefit when used in combination with fludarabine and 5 of 6 patients who were previously documented to be refractory to both agents responded to the combination, with one achieving CR.²⁹ This combination is being examined in ongoing studies.

Patients with relapsed CLL should ideally be treated in the setting of clinical trials and a number of promising agents are currently being investigated. Most promising of these include the cyclin-dependent kinase inhibitor, flavoperidol, which has shown impressive responses using a modified dosing schedule.³⁰ Other novel agents include the histone deacetylase inhibitor depsipeptide, the cyclin-dependent kinase inhibitor UCN-01, the protein kinase C modulator bryostatin 1, and the antisense oligonucleotide oblimersen and the small molecule gossypol. A number of novel monoclonal antibody agents including engineered anti-CD20, anti-HLA-DR, anti-CD40 and anti-CD23 are being investigated. Agents with demonstrated efficacy in other disease settings, such as lenalidomide are also being investigated in CLL.

Although the median age at presentation of CLL patients is 65 years, 40% are younger than age 60, and 12% are younger than age 50 at presentation. A number of clinical and biological features can be used to identify high-risk patients before they become refractory to chemotherapy. In addition to symptom palliation, young patients with poor risk CLL, who almost invariably die of their disease, are being offered therapies such as stem cell transplantation (SCT) to prolong survival and potentially cure their disease. To date, no studies in CLL have compared the role of standard chemotherapy with transplantation, although a retrospective matched-pair analysis suggested a survival advantage for autologous SCT over conventional therapy.³¹ A pilot study from the Medical Research Council enrolled previously untreated patients and followed them prospectively to asses the feasibility of performing autologous SCT.³² As shown in Figure 1 , only 65 of 115 patients (56%) entered into the study on an intent to treat basis proceeded to autologous SCT. Only 1 patient died of early transplant-related complications and the CR rate after transplantation was 74% (48 of 65). The predicted 5-year overall survival was 77.5% and 5-year disease-free survival was 51.5%. None of the variables examined at study entry were found to be predictors of either overall or disease-free survival. Sixteen of 20 evaluable patients achieved a molecular CR in the first 6 months following transplantation. Detectable molecular disease by PCR was highly predictive of disease recurrence. Of concern, 5 of 65 (8%) patients developed post-transplant acute myeloid leukemia/myelodysplastic syndrome.

A number of studies have reported the outcome following autologous SCT for CLL (Table 4 ).^32,–36 Early treatment-related mortality is low after autologous transplantation, but there appears to be a high incidence of opportunistic infections in patients undergoing autologous SCT for CLL compared to other patient populations. Whether this is due to a greater degree of immune incompetence in patients with CLL or is secondary to the immune suppressive effects of fludarabine and other therapy is unknown. Occurrence of second malignancies is of concern and occurred in 19% of patients at a median of 35 (range 1 to 138) months, with secondary MDS occurring in 9% at a median of 36 (range 11 to 87) months after autologous SCT.³⁶ 

As in other disease settings, patients appear to have a better outcome when they are treated early in the course of the disease and in the presence of low tumor burden. This suggests that selected patients should be transplanted as soon as treatment is indicated and perhaps even, for selected patients, as consolidation therapy of a first complete or partial remission. One difficulty in selecting high-risk patients is that these patients may also have an adverse outcome after transplant. The immunoglobulin gene mutation status maintains its poor prognostic significance after autologous transplantation,³⁷ although this can be overcome using allogeneic transplant.³⁸ The timing of the harvesting of the cells and whether they should be harvested in first remission and kept until later in the treatment course also needs to be further investigated. It is not always possible to collect enough CD34⁺ cells, especially in heavily pre-treated patients, and it is recommended that an interval of at least 3 months should be allowed between the last dose of fludarabine and leukapheresis.³⁹ The major problem after autologous SCT remains relapse of disease. On long-term follow-up of 137 patients who underwent autologous SCT with chemosensitive disease, progression-free survival at 6 years was 30% (± 4%) with no evidence of a plateau.³⁶ Multiparameter flow cytometric analysis and PCR are being used to investigate whether persistence of minimal residual disease will predict which patients will relapse following transplant in CLL.

Allogeneic SCT for CLL is associated with significant morbidity and mortality, both from regimen-related toxicity as well as from graft-versus-host disease (GVHD) and infection. Despite this, there is evidence that patients who survive can have long term disease control (Table 5 ).^40,–43 In a report of Registry data, treatment-related mortality following allogeneic SCT in CLL patients was 46%, the mortality from GVHD being 20%.⁴⁰ The outcome of 25 patients with CLL who underwent allogeneic SCT at the Fred Hutchinson Cancer Center have been reported.⁴³ Twenty-one donors were HLA matched siblings, 1 was a partially mismatched sibling and 3 were syngeneic. Fourteen patients developed grades 2–4 acute GVHD and 10 developed clinical extensive chronic GVHD. Late clearance of CLL cells was associated with the development of chronic GVHD in 1 patient. Two patients had recurrent CLL. Non-relapse mortality at day 100 was unacceptably high at 57% for the 7 patients conditioned with busulfan and cyclophosphamide and 17% for the 18 patients conditioned with total body irradiation-containing regimens. Actuarial survival at 5 years for the 25 patients was 32%. All patients who received busulfan and cyclophosphamide died within 3 years of transplant. For the 14 patients transplanted since 1992 who received total body irradiation, actuarial 5-year survival is 56%, suggesting that long-term disease-free survival might be achieved in this disease.

There are no direct studies comparing the outcome of autologous compared to allogeneic SCT in the treatment of CLL. Studies from the MD Anderson suggest improved outcome after myeloablative allogeneic compared to autologous transplant.³³ In 14 patients with CLL refractory to, or relapsed after chemotherapy with fludarabine, 13 (93%) achieved a CR post-transplant. At the time of reporting, 9 patients (64%) remained alive and in CR with a median follow-up of 36 months,⁴⁴ suggesting that allogeneic hematopoietic transplantation can induce durable remission even in patients with CLL refractory to fludarabine. In a phase II study at Dana-Farber Cancer Institute, 162 patients with high-risk CLL were enrolled in a study in which 25 patients with an HLA-matched sibling donor underwent T cell-depleted allogeneic transplant and 137 patients with no sibling donors underwent B cell-purged autologous transplantation.³⁶ The 100-day mortality was 4% in both autologous and allogeneic SCT groups, although later treatment-related mortalities had a major impact on outcome. With a median follow-up of 6.5 years there was no difference in overall survival, and the 6-year overall survival was 58% after autologous and 55% after allogeneic SCT, although progression-free survival was significantly longer following autologous SCT than T cell-depleted allogeneic SCT.

The major advantage of the use of allogeneic SCT appears to be the potential for a graft-versus-leukemia effect. This effect can potentially be exploited by infusion of donor lymphocytes following allografting. A number of case reports in patients with CLL have suggested that a graft-versus-leukemia effect is operative in this malignancy, either following infusion of donor lymphocytes or after cessation of immunosuppressive therapy. A number of studies are underway addressing the issue of the number of lymphocytes required and the optimal timing of donor lymphocyte infusions after allogeneic SCT in this and other hematologic malignancies. Preclinical studies attempting to develop strategies to exploit maximal graft-versus-leukemia effect without concomitant GVHD are underway.

A major advance in reducing the short-term morbidity and mortality of SCT has been the introduction of non-myeloablative or reduced-intensity conditioning regimens. Although these procedures are commonly known as “mini-SCT,” this misnomer significantly underestimates the risk of GVHD associated with the procedure. Most frequently these regimens include fludarabine in the conditioning regimen and exploit the immunosuppressive as well as anti-leukemic properties of this agent. In this setting the majority of the anti-leukemia effect results from the graft-versus-lymphoma effect and not from the chemotherapy. GVHD can be decreased by the use of alemtuzumab in the conditioning regimen, although this may also be associated with an increased risk of relapse of disease, requiring later infusion of donor lymphocytes.⁴⁵ 

Reduced-intensity conditioning regimens appear to be associated with a decreased mortality after allogeneic transplantation, and allow transplantation in older patients, making this approach applicable to increased numbers of CLL patients.^46,47 Patients in these studies were often heavily pretreated and refractory to therapy. Despite this the majority demonstrated donor engraftment and the CR rate was high. Survival is improved in patients transplanted while they still had sensitive disease. These studies provide perhaps the strongest direct evidence to date for a graft-versus-leukemia effect that can be exploited in the management of CLL. A major focus of ongoing research is the amount of pre-transplant and post-transplant immunosuppression required to establish stable mixed chimerism and eventual full donor chimerism following non-myeloablative SCT. It should be stressed that these procedures are currently investigational in nature, and although the acute morbidity and mortality appears significantly lower when compared to high-dose conditioning regimens with allogeneic transplantation, the long-term results with regard to morbidity of chronic GVHD and disease control are not yet available.

To examine whether reduced-intensity conditioning decreases treatment-related mortality after allogeneic SCT for CLL, the outcome after allogeneic transplantation of 73 patients with reduced-intensity conditioning was compared with that of 82 matched patients from the European Bone Marrow Transplant Registry database who had undergone standard myeloablative conditioning for CLL during the same time period. Patients undergoing reduced-intensity conditioning had a significant reduction of treatment related mortality, but a higher incidence of relapse. There was no significant difference in event-free or overall survival between the two groups. ⁴⁸ 

The outcome of 64 patients with advanced CLL after non-myeloablative SCT from the Fred Hutchinson Cancer Research Center multi-institutional protocol using related (n = 44 ) or unrelated donors (n = 20) was reported recently.⁴⁷ As shown in Table 6 , the median age was 56 (range 44–69) years. The majority of patients were fludarabine refractory. Treatment-related mortality at 100 days was 11%, and 22% by 2 years, with significant GVHD. At a median follow-up of 24 months, 39 patients were alive, with 25 in CR. Two-year overall survival was 60% and disease-free survival was 52%. Although complications were higher in the patients with unrelated donors, there were higher CR and lower relapse rates than with related HCT, suggesting more effective graft-versus-leukemia activity from the unrelated donors.

In conclusion, high-dose therapy and autologous hematopoietic SCT are feasible in the many younger patients with poor-risk CLL, but the utility of this approach remains unproven. Monitoring of minimal residual disease may direct whether such an approach should be considered. Myeloablative allogeneic SCT is associated with increased morbidity and mortality and should be restricted to patients with very poor prognosis. Although no direct comparisons of myeloablative and non-myeloablative transplants have been performed for patients with CLL, given their older age, it seems most reasonable to consider non-myeloablative conditioning regimen transplants as the approach of choice for CLL patients in whom allogeneic transplant is being considered.

Although high-dose therapy is associated with a high CR rate and appears to result in long-term event-free survival and eradication of PCR-detectable minimal residual disease in a cohort of these patients, the follow-up of most clinical trials is too short to assess whether the use of SCT can cure CLL, and most series are now reporting continued relapses after this approach. It is possible that autologous transplantation will result in significant prolongation of overall and disease-free survival compared to conventional therapy, but this must always be considered in the context of improved outcome using conventional therapy (discussed by Professor Hallek in “Chronic lymphocytic leukemia (CLL): first-line treatment” in this volume). Furthermore, as data on reduced-intensity allogeneic transplantation become more mature, it may replace autologous transplantation as the procedure of choice in the future.

Future approaches to the management of this disease must take into account the balance between the increased morbidity and mortality of transplantation in CLL with the curative potential that these approaches potentially offer. In the absence of any other treatment modalities currently capable of improving outcome in this disease, the treatment of choice for younger patients with poor risk CLL may indeed be allogeneic or autologous SCT. Such treatment should be considered only in the setting of well-designed clinical trials assessing the impact of this treatment on outcome in these patients.