What Is the Optimal Initial Treatment for Chronic Lymphocytic Leukemia?
What Is the Optimal Initial Treatment for Chronic Lymphocytic Leukemia?
The past decade has seen significant advances in our understanding of the effects of cytogenetic abnormalities and other biologic factors on the natural history of chronic lymphocytic leukemia (CLL). Deletion of 17p13 and 11q22—corresponding to loss of the p53 and ataxia telengiectasia mutated (ATM) tumor-suppressor genes, respectively—is associated with the need for earlier therapy and a poor long-term prognosis.[1,2] The mutational status of the immunoglobulin heavy chain variable region (IgVH) also predicts time to treatment and long-term survival.[3-6] Patients whose CLL exhibits unmutated (germline) IgVH have a markedly inferior prognosis to that of patients with mutated IgVH. Expression of CD38 and zeta-associated protein (ZAP)-70 correlates with unmutated IgVH, the need for earlier therapy, and inferior long-term survival.[3,5-9]
|What is the ideal initial treatment for patients with chronic lymphocytic leukemia?|
These invaluable prognostic tools help clinicians predict the biologic behavior of a patient's CLL over time. Risk-stratification provides critical information on the percentage of high-risk patients in a clinical trial and is essential for proper interpretation of the results of clinical treatment studies in CLL. However, the use of these prognostic factors in selecting the optimal initial therapy for the individual CLL patient is less clear. While the impact of adverse cytogenetic abnormalities on response to therapy has been documented,[1,10] the effect of IgVH mutational status, CD38, and ZAP-70 on response to treatment is less clear. This review will pay particular attention to the efficacy of the various upfront treatment options in patients with poor-risk features such as del(17p13) and del(11q22).
When to Initiate Treatment
The National Cancer Institute (NCI) Working Group on CLL established guidelines for initiating treatment in 1996. These indications included autoimmue and nonautoimmune cytopenias, bulky or symptomatic lymphadenopathy or organomegaly, disease-related B-symptoms or fatigue, and rapid lymphocyte doubling time. However, CLL is often diagnosed by routine complete blood count examination, and most patients are asymptomatic with a normal hemoglobin and platelet count at diagnosis. Such asymptomatic patients should be followed expectantly and should receive treatment only upon disease progression.
Several studies in asymptomatic patients failed to show a benefit in overall survival (OS) with early chlorambucil (Leukeran) therapy. However, chemoimmunotherapy regimens combining monoclonal antibody therapy with purine analogs achieve much higher complete response (CR) rates than chlorambucil.[13-15] Thus, early treatment needs to be reconsidered, particularly in patients with high-risk biologic or genetic markers predicting a poor long-term prognosis.
Ongoing efforts by the German CLL Study Group (GCLLSG) and several US cooperative groups will address whether early intervention with chemoimmunotherapy can improve long-term survival in high-risk CLL. High-risk, previously untreated CLL patients in the United States who do not meet NCI 1996 treatment criteria but wish to be considered for early therapy should be referred to the Intergroup study randomizing such patients to observation or early chemoimmunotherapy.
Alkylating agents served as initial therapy for CLL for decades, and chlorambucil is still given as first-line therapy, particularly for older patients and patients who cannot tolerate purine analogs. Chlorambucil is typically given as a single pulse dose of 40 mg/m2 orally every 28 days. The use of concurrent steroids did not improve long-term survival; thus, chlorambucil is typically administered as a single agent.
While continuous daily dosing attained superior results, higher doses resulted in increased myelosuppression and frequent dose reductions, particularly in older or frailer patients. Thus, less-intensive pulse dosing should generally be used. Its primary advantages are a well established toxicity profile and low cost, whereas its disadvantages are a low CR rate, even in previously untreated patients, and the risk of myelodysplasia with extended use.
The introduction of purine analogs in the 1980s dramatically altered the treatment of CLL.[17-21] Fludarabine was approved for alkylator-refractory CLL, but several large, prospective, randomized studies demonstrating improved response rates and progression-free survival (PFS) established fludarabine as standard initial therapy for CLL in the United States.[22-24] These studies are summarized in Table 1.
A multicenter European study randomized 196 evaluable patients to fludarabine or the alkylator-based regimen CAP (cyclophosphamide, doxorubicin [Adriamycin], prednisone). The overall response (OR) rate favored fludarabine (60% vs 44%) in both relapsed (n = 96, 48% vs 27%) and previously untreated (n = 100, 71% vs 60%) patients, although the difference in the untreated group was not statistically significant. Fludarabine achieved a longer median duration of response and a tendency toward improved OS in previously untreated patients.
A randomized, multicenter study confirmed these findings in 509 previously untreated CLL patients. Patients were randomized to receive fludarabine at 25 mg/m2 IV daily for 5 days every 28 days, chlorambucil at 40 mg/m2 orally every 28 days, or fludarabine at 20 mg/m2 IV daily for 5 days and chlorambucil at 20 mg/m2 orally every 28 days, for up to 12 cycles. Patients who failed to respond or relapsed were allowed to cross over to the other arm. The combination arm was closed due to excessive toxicity. Fludarabine achieved superior CR, OR, median duration of remission, and median PFS (20%, 63%, 25 months, 20 months) than chlorambucil (4%, 37%, 14 months, 14 months). However, due to the crossover design, no statistically significant OS difference (66 vs 56 months) was observed.
A multicenter French study randomized 938 patients with previously untreated Binet stage B or C CLL to fludarabine, CHOP (cyclophosphamide, doxorubicin HCl, vincristine [Oncovin], prednisone), or CAP. While fludarabine achieved better response rates than CAP, OS was identical (67–70 months) in all three groups.
Most recently, the GCLLSG randomized 180 patients over age 65 to standard IV fludarabine for up to 6 cycles or chlorambucil at 0.4 to 0.8 mg/kg orally on days 1 and 15 every 28 days for up to 12 cycles. Patients in the fludarabine arm experienced more hematologic toxicity (63% vs 39%) but had a similar infection rate (5%) and better quality of life than patients treated with chlorambucil. Fludarabine induced higher CR rates (13% vs 0%) and OR rates (85% vs 68%), but there was no difference in PFS or OS. Thus, while single-agent fludarabine achieves superior response rates and PFS than alkylator-based regimens, OS is unchanged, due to the ability to salvage alkylator failures (Table 1).
Combining Fludarabine and Alkylator Therapy
Studies of the combination of fludarabine and cyclophosphamide (FC) in relapsed CLL have demonstrated promising clinical activity and acceptable toxicity.[20,21] Based on these results, the GCLLSG randomized 375 previously untreated CLL patients (age ≤ 65 years) to standard fludarabine or FC (fludarabine at 30 mg/m2 IV and cyclophosphamide at 250 mg/m2 IV daily for 3 days) every 28 days for 6 cycles. The OR rate (94% vs 83%), CR rate (24% vs 7%), median PFS (48 vs 20 months), and treatment-free survival (37 vs 25 months) favored FC, but no difference in OS was observed. FC was superior to fludarabine in IgVH unmutated patients (median PFS = 52 vs 20 months) and del(11q22) patients (median PFS = 36 vs 16 months). However, patients with del(17p13) did poorly regardless of treatment arm, with a median PFS and OS of 11 and 16 months, respectively. Thus, FC mitigated the poor prognosis associated with unmutated IgVH status and del(11q22), but not of del(17p13).
The US Eastern Cooperative Oncology Group (EGOG) randomized 278 patients to single-agent fludarabine or FC. FC consisted of fludarabine at 20 mg/m2 IV on days 1 through 5 and cyclophosphamide at 600 mg/m2 IV on day 1 every 28 days for up to 6 cycles. FC achieved a superior OR (74% vs 59%), CR (23% vs 5%), and median PFS (32 vs 19 months), but no OS advantage was seen. Information on genetic and molecular prognostic factors was available on 235 patients. CR and OR were not affected by cytogenetics, IgVH mutational status, CD38, ZAP-70, or p53 mutational status. There was no statistically significant difference in median PFS between IgVH-mutated (29.6 months) and -unmutated patients (20.8 months), and FC was superior to fludarabine in both groups. However, the presence of del(17p13) or del(11q22) negatively affected PFS, with hazard ratios (HRs) of 3.43 and 1.90, respectively. This study suggested that cytogenetics, but not IgVH mutational or ZAP-70 status, predicts the clinical outcome of fludarabine-based therapy.
Finally, the UK CLL4 study randomized 777 patients to oral chlorambucil, fludarabine, or FC. Patients randomized to FC enjoyed a superior OR, CR, and 5-year PFS (94%, 38%, 36%, respectively), compared to patients who received chlorambucil (72%, 7%, 10%) or fludarabine (80%, 15%, 10%). FC was superior in all age groups, including patients over 70 years old. However, patients with del(17p13) had markedly inferior CR/nodular partial response (nPR) and OR rates compared to patients without loss of 17p13, whether therapy was fludarabine (22% and 44% vs 38% and 79%) or FC (25% and 38% vs 65% and 97%).
Thus, FC achieves superior response rates and PFS compared to single-agent fludarabine (Table 1). However, no OS advantage for upfront FC has been observed to date. Furthermore, patients with del(17p13), which occurs in 5% of CLL patients at diagnosis, do poorly even with FC. Therefore, clinical research to identify effective initial therapy regimens in this cytogenetic group remains a priority.