Hairy cell leukemia

Evaluation of the long-term outcome in 121 patients

Authors


Abstract

BACKGROUND:

Historically, the first treatment choices for hairy cell leukemia (HCL) were splenectomy and alpha-interferon. Recently, purine analogues (pentostatin and cladribine) changed radically the treatment modality, inducing complete and durable responses in the majority of patients.

METHODS:

The authors analyzed the outcome of different lines of therapy in 121 HCL patients followed in their institute from 1986 to 2008, with a median follow-up of 105 months. Patients were divided into subgroups according to the number of treatments; Group A included 121 patients who underwent a front-line therapy, Group B patients (n =53) were treated with 2 lines, Group C patients (n = 34) with 3 lines, Group D patients (n = 17) with 4 lines, and Group E patients (n = 8) with 5 lines.

RESULTS:

In Group A, 92 (77%) patients obtained a complete response (CR), 23 (18%) a partial response, and the remaining 6 (5%) a minor or no response; median duration of response was 2.7 years. In Group B, 53 relapsed patients achieved a second CR rate of 73.5%; median duration of response was 2.5 years. Group C contained 34 patients in a second relapse, with a CR rate after the third line of treatment of 70.5% (median duration of response, 2.2 years). In Group D, 11 (64.7%) patients obtained a CR (median duration of response, 1.6 years), and in Group E 4 (50%) of 8 patients achieved a CR (median duration of response, 1.3 years).

CONCLUSIONS:

This study confirms the high risk (>40% of all patients) of retreatment of HCL patients and the need to maximize primary response. Cancer 2010. © 2010 American Cancer Society.

Hairy cell (HC) leukemia (HCL) has always attracted interest out of proportion to its frequency. Over the years, the reasons for this high level of interest have varied, but have always been underpinned by hematologists' fascination with the distinctive cytology and biology of HCs. After recognition of the disease as a distinct entity in the 1950s,1 interest initially centered on correct diagnosis. From the start, the diagnosis was recognized to have important therapeutic implications. Somewhat counterintuitively, given the activated appearance of the HC and the frequent presence of cytopenias, the disease was recognized to behave in an indolent fashion and to be resistant to aggressive chemotherapy.

Before the introduction of purine analogues, treatment with splenectomy and interferon-alpha was a common approach that led to clinical and hematologic responses, but they were rarely complete, and median survival was only 4 years.1, 2 Patients who undergo splenectomy have a better 5-year overall survival than those who do not; however, the 5-year survival is only 55% to 60%.3, 4 Treatment of HCL changed dramatically approximately 20 years ago, when it was shown that patients who had been given 2′-deoxycoformycin or 2-chlorodeoxyadenosine had achieved durable complete remission (CR).5, 6 Treatment with 2-chlorodeoxyadenosine or 2′-deoxycoformycin resulted in 80% to 90% CR, with a relapse rate of 25% to 30% at 3- to 5-year follow-up.7 In past years, several reports8-14 of long-term follow-up studies with a single purine analogue have confirmed that the overall CR rates are very similar, ranging from 76% to 98%, and the relapse rates are equally comparable, showing that about 30% to 40% of patients would relapse after 5 to 10 years of follow-up. These reports indicate that both purine analogues are highly effective and lead to excellent overall survival; however, it appears that neither purine analogue is likely curative, because molecular studies suggest that all patients have evidence of minimal residual disease while in CR.15 Recently, Else and coworkers16 reported their data on both agents. They reviewed data on 233 patients treated with either 2′-deoxycoformycin or 2-chlorodeoxyadenosine after a median follow-up of 16 years. The authors have extensive long-term experience with both agents, and concluded that there was no significant difference in outcome between the 2 agents. They estimated that the overall CR rate is 80%, and the median relapse-free survival is 16 years. In patients who relapse, second response rates are high, and the remissions are quite durable. For those patients still in remission at 5 years, about 25% will relapse by 15 years.

The conclusions from these long-term follow-up studies involving purine nucleoside analogues show that enormous progress has been made with prolongation of patient survival. Substantial evidence of residual disease after therapy may account for the 40% relapse rate over time. The survival curve for patients reported in long-term studies has not reached a plateau.

The actual therapeutic problem is related to the concept that this disease has not been eradicated or cured; there is a need for continued study in this field to define the optimal therapeutic regimen; in particular, the need for additional biologic or immunologic agents should be investigated.

Herein, we retrospectively review our institutional series of 121 HCL patients with a long follow-up, explaining the presence of particular subsets of patients with the need of several treatments during the outcome.

MATERIALS AND METHODS

Patients

Our clinical database of consecutive patients with HCL (n = 129) was retrospectively reviewed. Of these, 121 patients were included in this study. We analyzed the outcome of the different lines of therapy in 121 HCL patients followed in our institute from 1986 to 2008, with a median follow-up of 105 months. This was a retrospective review of the data after treatment was completed and was not a clinical trial. The patients were not treated uniformly but according to the optimal therapeutic approach used in the different periods. Table 1 summarizes the characteristics of all 121 patients at diagnosis.

Table 1. Characteristics of 121 Patients
CharacteristicValue
Sex, men/women103/18
Age, y 
 Median55
 Range27-84
White blood count (109/L) 
 Median3.0
 Range0.9-25.7
Hemoglobin (g/dL) 
 Median11.8
 Range5.3-16.6
Platelets (109/L) 
 Median80.0
 Range10.0-255
Splenomegaly 
 Yes58
 No63

The diagnosis of HCL was established by morphology, flow cytometry, and immunohistochemical analysis of peripheral blood, bone marrow, and/or spleen specimens. HCL variant was excluded by the study.

Treatment

Patients were divided into subgroups according to the number of treatments. Group A included all the 121 patients who received the first line of therapy; among these, 68 patients received only the first-line treatment. Group B was represented by 53 patients treated with 2 lines of treatment; Group C had 34 patients who underwent 3 lines; Group D included 17 patients with 4 lines of therapy; and Group E was represented by 8 patients treated 5 times.

Table 2 shows the different therapeutic approaches in the 5 subgroups. 2-Chlorodeoxyadenosine was given in 3 different schedules according to the different periods: 1) by continuous intravenous infusion at a dose of 0.1 mg/kg/d over 7 consecutive days; 2) by 2-hour intravenous infusion for 5 consecutive days at 0.14 mg/kg/d; and 3) by 2-hour intravenous infusion at a dose of 0.14 mg/kg once a week for 5 consecutive weeks. 2′-deoxycoformycin was administered at 4 mg/m2 intravenously every 2 weeks until maximal response, consolidated with 2 further doses. For both treatments, prophylaxis was given with cotrimoxazole against Pneumocystis carinii pneumonia. Rituximab was administered at the conventional dose of 375 mg/m2 weekly for 4 consecutive weeks.

Table 2. Summary of the Different Therapeutic Approaches of Each Successive Course of Therapy
TreatmentGroup A, n=121Group B, n=53Group C, n=34Group D, n=17Group E, n=8
  1. 2-CdA indicates 2-chlorodeoxyadenosine; 2′-DCF, 2′-deoxycoformycin.

2-CdA754123113
2′-DCF10542
Alpha-interferon33512
Rituximab2524
Splenectomy311

Response Criteria

Response was evaluated at 3 months after treatment by morphology and immunophenotyping of peripheral blood and bone marrow biopsy. The criteria for response were those of the Consensus Resolution.17 Briefly, a CR was defined as a morphological absence of HCs in the blood and bone marrow and normalization of any organomegaly and cytopenias. A partial response (PR) required normalization of peripheral counts, together with at least 50% reduction in organomegaly and bone marrow HCs, and <5% circulating HCs. All other responses were considered as nonresponses.

In addition, relapse after CR was defined as the reappearance of HCs in the peripheral blood or bone marrow, development of cytopenias, and/or splenomegaly on physical examination. Relapse after PR was a >50% increase of residual disease.

RESULTS

Figure 1 explains the response and the mean duration of response of the different lines of therapy. In this report, we included CR and PR patients in the median duration of response evaluation.

Figure 1.

Evaluation of response and median duration of response are shown in the 5 lines of treatment. Range: *0.52-14.88; **0.23-9.39; ***0.67-8.39; ****0.64-1.98. NR indicates nonresponders; PR, partial response; CR, complete remission.

Initially, of the 121 patients, 75 received 2-chlorodeoxyadenosine, 10 2′-deoxycoformycin, 33 alpha-interferon, and 3 splenectomy, resulting in 77% CR, 18% PR, and 5% nonresponders (NRs). Fifty-three patients relapsed, with a median duration of response of 2.7 years (range, 0.52-14.88 years). In particular, the median duration of response of CR patients was 4.1 years (range, 1.10-14.88 years). They were retreated with 4 different lines of therapy: 41 with 2-chlorodeoxyadenosine, 5 with 2′-deoxycoformycin, 5 with alpha-interferon, and 2 with rituximab, resulting in 73.5% CR and 15% PR; 11.5% of patients were NRs. The median duration of response was 2.5 years (range, 0.23-9.39 years). Thirty-four patients underwent third-line of therapy; in particular, 23 patients received 2-chlorodeoxyadenosine, 4 2′-deoxycoformycin, 5 rituximab, 1 splenectomy, and 1 alpha-interferon, resulting in 70.5% CR and 20.5% PR; 9% of patients were NRs. Seventeen patients relapsed, with a median duration of response of 2.2 years (range, 0.67-8.39 years). They were retreated with 4 different lines of therapy: 11 with 2-chlorodeoxyadenosine, 2 with 2′-deoxycoformycin, 2 with rituximab, and 2 with alpha-interferon, resulting in 64.7% CR and 23.5% PR; 11.8% of patients were NRs. Eight of 17 patients experienced disease progression, after a median duration of response of 1.6 years (range, 0.64-1.98 years). These 8 patients received a fifth line of therapy, and after treatment (3 2-chlorodeoxyadenosine, 4 rituximab, and 1 splenectomy) only 1 relapsed, with a duration of response of 1.3 years.

Table 3 summarizes the responses in the different lines of treatment. The ability to attain CR decreased with each course of therapy. Although relapse rates remain relatively stable after each course of treatment (43.8% after first line, 64.1% after second line, 50% after third line, 47% after fourth line), there is a reduction in median duration of response with successive courses of treatment.

Table 3. Response According to the Different Lines of Treatment
GroupNo. of PatientsCR (%)PR (%)NRs (%)
  1. CR indicates complete remission; PR, partial response; NR, nonresponder.

A12192 (77)23 (18)6 (5)
B5339 (73.5)8 (15)6 (11.5)
C3424 (70.5)7 (20.5)3 (9)
D1711 (64.7)4 (23.5)2 (11.8)
E84 (50)2 (25)2 (25)

There were 15 deaths in this series (12.5%); 5 deaths (4.1%) were HCL-related, after refractory disease. One (0.9%) patient died because of a secondary malignancy, 6 (5%) patients died as a consequence of cardiovascular events, and 1 patient (0.9%) died in an accident. For 2 (1.6%) patients, the cause of death remains unknown.

DISCUSSION

The last 2 decades have represented an exciting period for clinicians involved in the care of patients with HCL. Four new drugs, alpha-interferon, 2′-deoxycoformycin, 2-chlorodeoxyadenosine, and rituximab, have been identified as highly effective in the treatment of this rare disease. The advances in the treatment of HCL over the past 20 years have resulted in survival curves similar to those for the appropriate age-related cohorts. Although prolonged remissions are usually seen after treatment with purine analogue drugs, a proportion of patients are primarily resistant to these agents, and it is probable that a significant proportion of those entering remission will ultimately relapse.

Relapses occur in up to 40% of patients in several long-term follow-ups, regardless of the purine analogue that was used as first-line therapy.8, 11-14, 16, 18-20 After relapse, treatment with an additional course of purine analogues is recommended. It is acceptable to repeat the same treatment that was given initially, although changing to a different purine analogue will yield the same results. Although the ability to obtain a CR decreases with each course of therapy, the probability of achieving CR does not seem to be affected by the agent that is used.14, 16

The present analysis extends the follow-up of 121 patients consecutively treated at the Seràgnoli Institute with different therapeutic approaches (70% with purine analogues as front-line treatment) according to the period (from 1986 to 2008). Also in the further lines of treatment, the purine analogues were the most used drugs: 86% (second line), 79% (third line), and 77% (fourth line); only with the fifth line, the subset of 8 heavily treated patients, were purine analogues used in only 37.5%. With a median follow-up of 105 months, this series shows the real scenario in terms of clinical outcome of HCL patients, represented by several progressions/relapses in >40% of all study population; extrapolating, there is a small subset of patients with refractory disease.

The ability to achieve CR decreased with each course of therapy: from 77% of the front line to 50% of the fifth line (second line, 73.5%; third line, 70.5%; fourth line, 64.7%). Although relapse rates remained relatively stable after each course of treatment (43.8% after first line, 64.1% after second line, 50% after third line, 47% after fourth line), there was a constant reduction in median duration of response with successive courses: from 2.7 years after the first, to 1.3 years after the fifth different therapeutic approach. Globally, 53 (44%) patients of our entire study population presented at least 1 progression/relapse and, particularly, there is a real selected subgroup of HCL patients with multirelapsed disease.

Although a substantial number of patients have remained disease-free for a prolonged period of time, >1/3 eventually relapsed. Retreatment with purine analogues is associated with a high response rate; however, the relapse-free survival curve does not appear to reach a plateau, and some patients become refractory to this class of agents. Recently, monoclonal antibodies and immunotoxins have emerged as a useful strategy for patients who have relapsed or refractory HCL.21-25 The real need is to maximize primary responses to prolong remissions. Patients who have a complete hematological remission are frequently found to have minimal residual disease by immunohistochemical and immunophenotypic studies. Few preliminary data have reported that the combined use of rituximab after a purine nucleoside analogue, or given concomitantly with these agents, may produce a higher percentage of CR without evidence of residual disease and cumulative toxicity.26-28 This chemotherapy-immunotherapy combination could increase the economic implications but, at the same time, in terms of cost/benefit balance, it is important to evaluate this aspect over a long period of time, because it could potentially be possible to reduce the number of further treatments for the patients.

There is a need for continued study in this field to better define the optimal therapeutic regimen. Duration of therapy, optimal dose and schedule of drug administration, and the need for additional biologic or immunologic agents should be investigated. In those patients with resistant or relapsed disease, we also need to explore new agents as well as targeted immunotoxin conjugates.

CONFLICT OF INTEREST DISCLOSURES

Supported in part by BolognAIL.

Ancillary