• T-cell large granular lymphocyte leukemia;
  • methotrexate;
  • cyclosporin A;
  • pentostatin;
  • cyclophosphamide


  1. Top of page
  2. Abstract
  6. Acknowledgements


To the authors' knowledge, there is no standard treatment for patients with T-cell large granular lymphocyte (LGL) leukemia. Available data are limited by patient numbers and coexisting pathologies.


The authors report on the use of immunosuppressants (cyclosporin A [CSA] and low-dose oral methotrexate [MTX] given continuously) and cytotoxic agents in the treatment of 29 patients with T-cell LGL leukemia age over the past 20 years.


The overall response rate (ORR) to MTX (n = 8 patients) was 85.7% (complete hematologic response [CHR] rate, 14.3%; partial response [PR] rate, 71.4%) with dose-dependent responses observed and safe usage of doses >10 mg/m2 per week in 2 patients. The ORR to CSA (n = 23 patients) was 78.2% (CHR rate, 30.4%; PR rate, 47.8%). The median time to response for both agents was 1 month. Toxicity, although it was minor in most patients and was more common in the CSA group, included second malignancies in 5 patients. An ORR of 67% (all CHR) was attained with pentostatin (n = 4 patients); recurrences developed after a median of 4.6 years. Successful retreatment with pentostatin was possible but with increasing drug resistance. Cyclophosphamide induced CHR that lasted >7 years with bone marrow clearance in 1 of 4 patients. Alemtuzumab induced a PR in 1 patient who had refractory disease.


Both MTX and CSA were efficacious in the treatment of T-cell LGL leukemia but generally required long-term maintenance therapy. The authors highlight the risks of second malignancies and persistence of bone marrow disease. Although MTX and CSA were effective as first-line therapy, alemtuzumab and pentostatin merit further investigation, particularly for refractory disease. Cancer 2006 © 2006 American Cancer Society.

T-cell large granular lymphocyte (LGL) leukemia is a lymphoproliferative disorder of clonal CD3-positive cytotoxic T cells that was described first by Brouet et al.1 and defined by McKenna et al.2 The median age ay onset is 55 years with equal gender distribution. One-third of patients are asymptomatic.5 Clinical manifestations usually are because of neutropenia with recurrent infections and mouth ulcers. Eighty-five percent of patients are neutropenic,6 with severe neutropenia (<0.5 × 109/L) seen in 50% of patients. Most deaths attributable to this leukemia are caused by bacterial sepsis in patients with severe neutropenia. More rarely, thrombocytopenia (20%) and/or anemia (50%) may occur.6, 7 Splenomegaly is common, whereas lymphadenopathy is rare. Other associated findings include positive rheumatoid factor status (with or without arthritis), hyper-γ-globulinemia, livedo reticularis, and other autoimmune conditions.6, 7

Indications for the treatment of LGL leukemia include neutropenia with recurrent infections, transfusion-dependent anemia, symptomatic thrombocytopenia, and (more rarely) progressive lymphocytosis and/or splenomegaly. An understanding of the mechanisms of cytopenias would help to rationalize treatment. However, those mechanisms remain elusive. Possible mechanisms include 1) the dysregulation of Fas/Fas ligand pathways. Leukemic LGL cells express high levels of Fas and Fas ligand and are resistant to Fas-mediated death in vitro.8 The correction of cytopenias in T-cell LGL leukemia has been associated with disappearance and/or reduction in serum Fas ligand levels9, 10; and 2) immune-mediated phenomena, which include direct antiglobulin test (DAT)-positive hemolytic anemia,11 pure red cell aplasia (PRCA), and (rarely) immune thrombocytopenia.12 Neutrophil antibodies, although they often are positive, are of doubtful significance because of increased levels of circulating immune complexes.5 Overlapping syndromes of LGL leukemia, myelodysplasia (MDS), and paroxysmal nocturnal hemoglobinuria13, 14 also favor an immune basis for this disease, with response type to antigenic stimulation determining the predominant disease phenotype. It is believed that persistent antigenic stimulation15–17 underlies pathogenesis. One possible source of such stimulation is a retrovirus or protein homologous to the BA21 epitope of HTLV-I.18 Neither bone marrow (BM) infiltration5, 7 nor hypersplenism can account for the degree of cytopenias; whereas cytokines, such as tumor necrosis factor α and interferon γ, which are known to inhibit erythropoeisis, are elevated in LGL leukemia.5

Unlike B-cell malignancies, there is no gold standard for the treatment of patients with T-cell LGL leukemia. Published data have been limited to case reports, retrospective series, and a few prospective series with small patient numbers and a variety of agents. Response criteria and coexisting disorders, such as MDS and PRCA, also were variable. Treatment approaches have included growth factors, splenectomy, cytotoxic therapy (alkylating agents, purine analogues, steroids), and immunosuppressive drugs (cyclosporin A [CSA], methotrexate [MTX], and [rarely] antithymocyte globulin). Therefore, it is difficult to compare data from these studies directly and, in particular, to compare retrospective data with prospective data.

Correction of cytopenias with therapy may be achieved without eradication of the clone,19–26 which often is resistant to treatment. Furthermore, complete normalization of the neutrophil count is not essential for symptomatic improvement. In this report, we describe the efficacy and safety of drugs in a cohort of 29 patients with well documented T-cell LGL leukemia who were treated over the past 2 decades, and we evaluate the role of these drugs in patients with new, recurrent, and/or refractory disease.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Twenty-nine patients (12 men and 17 women) with pure, isolated T-cell LGL leukemia who required treatment were identified from data bases at the Royal Marsden Foundation National Health Service Trust (London, United Kingdom), Rikshospitalet University Hospital (Oslo, Norway), and Royal Berkshire Hospital (Reading, United Kingdom). Twenty-four untreated patients from these data bases for whom some follow-up data were available were used to compare survival data. Patient note review was performed. The diagnosis of T-cell LGL leukemia was based on peripheral blood morphology, immunophenotype, and demonstration of T-cell clonality by polymerase chain reaction (PCR) or Southern blot analyses.

Pretreatment patient characteristics are summarized in Tables 1 and 2. Survival was estimated from the date of diagnosis to the date of death from any cause or last follow-up.

Table 1. Patient Characteristics
Patient No.Age, YearsGenderClinicalCytopeniaRhF (A)L/S/LN*DATStatusOS, Months
  • RhF indicates rheumatoid factor; OS, overall survival; RhF, rheumatoid factor positive; L, hepatomegaly; S, splenomegaly; LN, lymphadenopathy; DAT, direct antiglobulin test; E, anemia, N, neutropenia; P, thrombocytopenia; + A, coexisting rheumatoid arthritis; ND, not done; NS, not stated; RhA, rheumatoid arthritis; Hb, hemoglobin; FBC, full blood count; ITP, immune thrombocytopenia; URT, upper respiratory tract infection.

  • *

    Letters in parentheses indicate radiologic evidence only.

  • The most severe cytopenia(s) and/or main indication(s) for treatment.

134MaleSymptomatic anemia, transfusion dependentENSS, LNegativeDead178.7
268FemaleArthritis, recurrent sore throats, mouth ulcersNRhF + A(S)PositiveAlive43.1
377FemaleFever, sweats, arthralgiaNRhF + ASNegativeAlive96.2
427FemaleRecurrent sore throatsNNegative(S)NegativeAlive112.4
557FemaleAnergia, poor healing, mouth ulcersN, E, PRhF + ASNDDead224.7
654FemaleMouth ulcersN, E, PRhF + ANoneNegativeAlive101.2
737FemaleEasy bruising and epistaxisPNegativeNoneNegativeAlive93.8
874MaleAsymptomatic, routine FBC showed lymphocytosisN, E (lymphocytosis)NegativeNoneNegativeAlive39.3
962MaleAsymptomatic neutropenia, anorexia, depression, weight lossP (marked lymphocytosis)NSS, LNDDead45.9
1048MaleRecurrent skin infectionsNNegative(S)NegativeAlive55.7
1126MaleMouth ulcersN, E, PRhFSPositiveAlive80.5
1273MaleRecurrent skin infectionsNNegativeNonePositiveAlive49.6
1366FemaleMouth ulcers, recurrent skin infectionsNRhF + AS (LN)NegativeAlive158.5
1456FemaleOphthalmic shingles, mouth ulcers, nose bleedsN, E, PNegativeNoneNegativeAlive41.4
1531FemaleMouth ulcers, sinusitisN, ENSSNSDead70.7
1657MalePyelonephritis, splenomegaly, anemia (Hb 3 gm), ITP-likeN, E, PNegativeS, (LN)NegativeAlive68.5
1731FemaleLymphocytosis, URTI, skin infections (pregnancy)N, ENegative(S)NegativeAlive29.0
1878MaleSymptomatic anemiaN, E, PRhF + AL, S, LNNegativeDead12.4
1962FemaleRecurrent infections and neutropeniaNRhF + ASNegativeAlive14.0
2054FemaleRecurrent infections and neutropeniaN, ENo RhASNDAlive36.7
2145FemaleHemolytic anemiaENegativeSNegativeAlive24.3
2261FemaleProgressive neutropeniaNRhF + ANoneNegativeAlive37.7
2335MaleAbdominal painN, PNo RhASNDAlive12.9
2467FemaleProgressive granulocytopeniaNRhF + ANoneNDAlive35.4
2550FemaleRecurrent infections and neutropeniaNNegativeNoneNegativeAlive42.7
2678MaleRecurrent pneumonias and neutropeniaNRhF + ASNDAlive8.0
2779MaleRecurrent infections and neutropeniaN, ENegativeNoneNegativeAlive5.3
2834MaleRecurrent infections and neutropeniaNND + ASNegativeAlive18.0
2982FemaleRecurrent infectionsNNegativeNoneNDAlive30.1
Table 2. Patient Characteristics by Treatment Group
 No. of patients (%)
  • CSA indicates cyclosporin A; MTX, methotrexate.

  • *

    P = .0002.

 No. of patients2923824
 Male:female ratio12:178:154:414:10
 Median age, y (range)57 (26–82)57 (26–82)58.5 (26–73)51.5 (17–79)
 Median × 109/L (range)*0.3 (0–4.9)0.3 (0–4.9)0.30 (0.1–0.8)1.75 (0.2–7.7)
 ≤1 × 109/L25198 
 ≤0.5 × 109/L20187 
 Median × 109/L (range)174 (3–649)184 (15–649)195 (30–297)188 (148–378)
 ≤150 × 109/L962 
 ≤50 × 109/L421 
 Median g/dL (range)10.9 (8.5–17.3)11 (8.6–13.9)12.9 (10.4–15.8)13.7 (10.4–17.3)
 ≤11 g/dL1491 
 ≤10 g/dL650 
 Median × 109/L (range)2.6 (0.4–42.9)2.6 (0.4–42.2)2.20 (1.5–12.8)6.0 (0.8–14.6)
Rheumatoid factor
 Rheumatoid factor positive11/23 (47.8%)11/20 (55%)3/7 (43%)0/5 (0%)
 Rheumatoid arthritis9/24 (37.5%)9/22 (40.9%)2/8 (25%) 
Adenopathy/organomegaly [radiologically only]
 Spleen19 [4] (66)15 [4] (65)5 [2] (63)4 [0]
 Liver3 [0] (10.7)2 [0] (8.7)0 [0] (0)1 [0]
 Lymph node3 [2] (10.7)2 [1] (8.7)1 [1] (12)0 [0] (0)


Low-dose, oral, weekly MTX (5–13.5 mg/m2) was used. Concomitant, low-dose, oral folic acid (5–10 mg daily) was given to prevent mouth ulceration. Liver and renal functions were monitored, and lung function testing was performed biannually. Treatment was continued indefinitely in responders. Patients who received initial, concomitant prednisolone or growth factor therapy were included as long as those agents were tailed off within the first 2 months.

Oral CSA therapy (1–6.7 mg/kg daily) was given in 2 divided doses. Renal function, blood pressure, and CSA levels were monitored. Although no target therapeutic level was set, values were maintained at ≤250 ng/mL in the absence of toxicity. Treatment was continued indefinitely in responders. Patients who received concomitant prednisolone or growth factor therapy were included as long as those agents were tailed off within the first 2 months.

Single-agent cyclophosphamide (50–100 mg orally daily or 250–500 mg intravenously [i.v.] weekly) was given until maximum response then stopped.

Pentostatin (4 mg/m2 i.v.) was administered every 1 to 2 weeks until maximum response and then stopped. Patients received prophylaxis against Pneumocystis carinii pneumonia (PCP), fungal infections, and viral infections.

Alemtuzumab was given i.v. as a dose-escalating regimen for 1 week followed by 30 mg daily for 6 weeks. Appropriate prophlylaxis was provided.

Assessment of Response

Response criteria are summarized in Table 3. BM infiltration was not included; however, where BM samples were assessable, responses are described. The time to response (TTR) was defined from the start of the maximum evaluated dose to achievement of ≥50% improvement in cytopenia(s). Differences in pretreatment and posttreatment blood counts were compared by using the Student t test for paired data. Kaplan–Meier survival curves were constructed, and differences were compared by log-rank testing or proportional hazard (Cox) regression analysis.

Table 3. Criteria for Response Used in this Study*
  • *

    Responses had to be sustained for ≥2 consecutive months.

Complete hematologic response
 All of the following must be true:
  Normal blood count (Hb ≥13 g/dL for men, ≥11 g/dL for women, neutrophils >2 × 109/L, lymphocytes <3.5 × 109/L, platelets >150 × 109/L)
  Absence of organomegaly
  Absence of symptoms
Partial response
 All of the following must be true:
  Improvement of organomegaly of ≥50% and
  Blood lymphocytes <15 × 109/L
  Neutrophils ≥2 × 109/L or 50% improvement from baseline
  Platelets ≥100 × 109/L or 50% improvement from baseline
  Hemoglobin ≥12 g/d: for men or ≥11 g/dL for women or 50% improvement from baseline, not supported by transfusion
  Absence of symptoms
Minimal response
 Improvement in organomegaly, cytopenia(s), and or lymphocytosis falling short of a partial response
No response
 Any response that does not include the above but without worsening in cytopenia or organomegaly
Progressive disease
 Worsening of cytopenia or organomegaly


  1. Top of page
  2. Abstract
  6. Acknowledgements

No patient was identified with other coexisting hematologic malignancies. Among the treated patients (Tables 1, 2), the median age at diagnosis was 57 years (range, 26–82 years), neutropenia (<1 × 109/L) was present in 86% of patients and was severe in 80% of patients; thrombocytopenia was observed in 31% of patients and was an isolated phenomenon in 7% of patients; and anemia (hemoglobin, <10 g/dL) was documented in 21% of patients but was the sole cytopenia in only 2 patients. Twenty-one percent of patients showed pancytopenia. Circulating LGLs were seen in all patients. The predominant phenotype was CD3-positive/CD8-positive/CD4-negative (94%). One patient (Patient 9) who had marked lymphocytosis showed an uncommon phenotype with CD4-positive cells that coexpressed natural killer antigens. Features among nontreated patients are summarized in Table 2.

Treatment and Response

Twenty-three patients received CSA, 8 patients received MTX, 4 patients received pentostatin, 4 patients received cyclophosphamide, and there was some cross-over of patients. Responses are summarized in Table 4. Improvement in fatigue often preceded hematological response.

Table 4. Overall Summary of Response
TreatmentNo. of patientsMedian TTR (range), monthsDuration of treatment (range), monthsMedian follow-up (range), monthsNo. of CHRsNo. of PRsORR (%)
  • TTR indicates time to response; CHR, complete hematologic response; PR, partial response; ORR, overall response rate; MTX, methotrexate; NE, not evaluable; CSA, cyclosporin A; DCF, deoxycoformycin.

  • *

    One patient only.

MTX8 (1 NE)1.0 (0.8–1.5)22.2 (3–31.8)22.8 (9.23–32.9)1585.7
CSA231.0 (0.5–16.5)23.1 (4–98.7)31.8 (0–149.6)71178.2
DCF4 (1 NE)0.75 (0.5–1.0)3.25 (1–10)55.52067
Cyclophosphamide40.75*1.625 (1.25–10)931033

An overall response rate (ORR) (partial responses [PRs] plus complete hematologic responses [CHRs]) of 86% and a 14% CHR rate were observed in the largely pretreated MTX group. The posttreatment neutrophil count increased significantly (P = .007) and was ≥1 × 109/L in all responders. A concomitant but nonsignificant reduction in lymphocytes (P = .14) was observed. MTX doses >10 mg/m2 were required to achieve this response in 2 patients. Although a CHR was achieved in only 1 patient, many of the remaining patients continued to show improvements in cell counts. Patients responded despite the number and nature of cytopenias. This included 2 patients with moderate-to-severe pancytopenia (Patients 11 and 14). Persistent, mild thrombocytopenia precluded a CHR in 2 patients.


CSA produced a 92% response rate (5 CHRs and 7 PRs) in treatment-naive patients (n = 13). In pretreated patients (n = 10), the response rate was 60% (2 CHRs, 4 PRs, 2 minor responses, and 2 nonresponses). The ORR was 78% (30% CHR rate). Therapeutic levels were variable: A CHR was achieved with CSA levels as low as 40 to 50 ng/L, and there were response failures despite CSA levels of 150 to 200 ng/L. One patient (Patient 8) required combination treatment with CSA and erythropoietin to achieve a response. Sequential analysis of CSA levels failed to identify any ideal therapeutic level. In 3 instances, continued responses were seen despite discontinuing CSA. Significant increases in neutrophil count (P = .004) were observed, with normalization in 6 patients (46%), again, associated with a nonsignificant reduction in lymphocyte counts (P = .124). Chi-square analysis indicated no association between response to CSA and the presence of splenomegaly (P = .27) or rheumatoid factor (P = .98); however, women had a significantly better response compared with men (P = .04).


Four patients received pentostatin, but only 3 of those patients were evaluable (Patient 15 died with a severe infection shortly after the first dose of pentostatin). Two patients (67%) achieved durable CHRs of 1.5 years (Patient 9) and 7.5 years (Patient 1). Both patients developed recurrent disease off therapy and were retreated with pentostatin: One of those patients (Patient 9) achieved 2 subsequent CHRs of 9 months' and 8 months' duration before showing only minimal response to a 4th course of the same treatment.30


Of 3 evaluable patients who received cyclophosphamide for a median of 1.6 months (range, 1.25–10 months), the only responder, who received a total of 10 months of therapy, showed an initial response after 3 weeks and achieved a CHR that lasted >7 years. That patient also achieved clearance of BM infiltration and, thus, had a true complete response.


Splenectomy (n = 5 patients) induced marked but often transient responses in platelets; however, neutropenia did not improve, and increasing lymphocytosis was seen. Eighty-three percent of patients showed neutrophil responses to granulocyte-colony stimulating factor (n = 6 patients). Minimal or no response was observed with other agents that were used infrequently, including thalidomide, mitomycin, and chlorambucil; combined cyclophosphamide, vincristine, and predisolone; combined cyclophosphamide, vincristine, doxorubicin, and prednisolone; combined mitozantrone, doxorubicin, and etoposide; and danazol. In particular, cytotoxic combination chemotherapies largely were ineffective and caused toxicity without benefit. Alemtuzumab in a single patient induced a rapid though transient PR.

BM Responses

BM response was assessed in 8 patients. Three patients (Patients 1, 9, and 13) who received pentostatin (n = 2 patients) and cyclophosphamide (n = 1 patient), respectively, attained histologic clearance of BM disease and, thus, represented the only 3 true complete responses in this series. BM appearances pre-MTX or pre-CSA and post-MTX or post-CSA treatments were evaluated in 5 patients with variable and inconsistent findings but with detectable residual disease in all patients.

Response Timing and Duration

Most rapid responses were observed with pentostatin and cyclophosphamide (median TTR. 0.75 months) (Table 4). The median TTR was 1 month for both MTX and CSA, although the range was greater for CSA (0.5–16.5 months) compared with MTX (0.8–1.5 months; P = .09). Maintenance treatment with CSA (median duration, 7.7 months; range, 0–98.7 months) and MTX (median duration, 22.8 months; range, 2.8–31.2 months) generally was required to sustain response. In 3 patients (Patients 4, 7, and 20), CSA was discontinued without recurrence of the cytopenia after 4.5 years, 1.4 years, and 0.8 years, respectively. However, in Patient 4, the effects of exogenous and endogenous progesterone may have influenced blood counts.29 Patient 11, who showed a good response to MTX despite 8 previous lines of therapy, eventually relapsed with neutropenia after 31 months and has achieved a second PR with a combination of MTX and very-low-dose CSA. He went on to survive a nonmyeloablative stem cell transplantation.

Toxicity and Survival

Adverse events generally were minor and did not require cessation of therapy, except in 2 patients on CSA, in whom therapy had to be discontinued because of persistent headaches in 1 patient and skin cancer, hypertension, and gum hyperplasia in the other patient (Patient 2). Side effects from MTX were observed in 2 patients who suffered minor folic acid-responsive mouth ulcers and intolerance caused by marked fatigue. Patient 12 developed Epstein–Barr virus-negative, diffuse, large B-cell lymphoma during MTX treatment at a dose of 13.5 mg/m2.

Three patients who received CSA developed second malignancies, including anal carcinoma, basal cell carcinoma, and hepatocellular carcinoma. There were 5 deaths (28%), with none related directly to therapy. Three patients died from neutropenic infective complications that were related to disease progression, 1 patient died from alcohol-related complications, and 1 patient on long-term CSA therapy died from hepatocellular carcinoma. The median overall survival for all 53 patients was 14.5 years (Fig. 1), and there was no significant difference between treated and untreated patients (P = .79). Treated and untreated patients showed significant differences only in pretreatment neutrophil counts and rheumatoid factor status. There was no statistical association between age at diagnosis, gender, rheumatoid factor status, or neutrophil count and survival.

thumbnail image

Figure 1. Overall survival is illustrated for all patients (treated and untreated; n = 53) with T-cell large granular lymphocyte leukemia.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  6. Acknowledgements

In the current study, we evaluated treatment options for patients with T-cell LGL leukemia. The cross-over of treatments reflects the refractory nature of this disease to many agents and its chronic course. Splenectomy seems to be applicable to patients who have thrombocytopenia and significant splenomegaly; and prednisolone appears to be useful as an introductory adjunct, allowing rapid albeit transient responses in cytopenias while immunosuppressive agents take effect. Growth factors may be useful in emergencies and in synergy with other agents.22, 31 Recent reports on the treatment of this disease have focused on CSA and, less frequently, MTX: Our current results confirmed the efficacy and tolerability of both. Although ORR and TTR ranges were better for MTX, CHRs were more frequent with CSA, particularly when it was used as first-line therapy. Normalization of blood count was not essential; and, clinically, PRs appeared to e as relevant as complete responses. Most responses to CSA and MTX occurred in the first 1 to 3 months, although exceptional patients took up to 16.5 months. Thus, therapeutic trials of these agents should span at least 3 months in the absence of side effects. Maximal responses may occur with MTX after prolonged follow-up,20 and it is possible that a proportion of patients who achieved PRs in our series may achieve CHRs.

Despite the initially reported efficacy of MTX (n = 10 patients),20 only 2 other small series12, 32 with overlapping patients have described the use of MTX for T-cell LGL leukemia with response rates from 50% to 100% (Table 5). Therefore, we have contributed to the limited data for MTX in this disorder. Similarly, although data have been reported on CSA, patient numbers were relatively small, follow-up was limited, and BM responses were not always reviewed. Sood et al.22 documented improvement in neutropenia in 5 patients who received CSA with responses that were dependent on the combination of CSA and granulocyte-macrophage–colony stimulating factor in 1 patient, and, similar to our series, continuing BM involvement. Battiwala et al.31 reported response rates of 67% (CHR rate, 41.7%; PR rate, 24%) in 12 patients who had pure T-cell LGL leukemia. Again, synergism with growth factors was documented. Single case reports19, 23, 24, 26 also support the efficacy of CSA and recognize persistence of the LGL clone despite treatment.

Table 5. Summary of Results for Methotrexate Therapy in Patients with T-Cell Large Granular Lymphocyte Leukemia
ReferenceTherapyNo. of patientsResponse criteria for CRResponse rate (No. of responses)Comments
  1. CR indicates complete response; MTX, methotrexate; LGL, large granular lymphocyte; CHR, complete hematologic response; PR, partial response; CR, complete response.

Loughran et al., 199420MTX with or without prednisolone10Normalization of full blood count and CD3-positive LGL count60% (5 CHR, 1 PR)Molecular CR in 3 patients (30%); 3 patients continued low-dose prednisolone
Hamidou et al., 200032MTX (7.5mg/week)4Normalization of white cell count with <0.4 × 109/L LGLs100% (3 CHR, 1 PR)Molecular CR in 50%
Dhodapkar et al., 199412MTX plus prednisolone2Normalization of blood counts including LGLs100%Molecular CR in 50%
Osuji et al., 2006 (current series)MTX with or without prednisolone8Normalization of full blood count78.7% (1 CHR, 5 PR)No continuing prednisolone; LGLs persist

Although Sood et al.22 suggested therapeutic levels of 200 ng/L for CSA, and many authors have adopted a higher induction dose tapered slowly to the lowest effective maintenance dose, we observed no correlation between CSA levels and neutrophil counts in responders and, thus, could not identify a fixed target level.

Current thinking suggests that CSA and MTX need to be given indefinitely to maintain response,19, 22 which constitutes a major potential disadvantage of these drugs. It is noteworthy that we were able to discontinue CSA treatment in 3 patients (Patients 4, 7, and 20), all women, without any recurrence of cytopenias.

Our results continue to support the utility of purine analogues. Response rates of 40% to 67% have been documented in small numbers of patients who received with pentostatin,30, 33, 34 fludarabine either alone21 or in combination,35 and cladribine.36 With comparable responses, rapid TTR, and independence from maintenance therapy, purine analogues should remain in the therapeutic armamentarium and may represent an attractive alternative, particularly for younger patients. Appropriate prophylaxis and clinical vigilance should be provided.

Although the response rate to cyclophosphamide was only 33% in a small number of patients, this response was the most durable encountered in the current review. Previous studies reported response rates of 44% (n = 16 patients) with CHR and molecular remission in 19% using low-dose oral cyclophosphamide and/or prednisolone.12

Apart from a report by Battiwala et al. suggesting the utility of human leukemic D-related antigen 4 (HLADR4) status in predicting response to CSA,22, 31 it remains unclear which factors predict response. Because HLADR4 is associated strongly with rheumatoid arthritis in this leukemia, it is possible that this association represents a surrogate phenomenon for coexisting rheumatoid arthritis. The size of our study subgroups allowed us to assess features that predicted response only among patients who received CSA; and, among the features studied (gender, splenomegaly, and rheumatoid factor status), only gender showed a significant correlation with response. Complete and durable remissions in our series were associated with BM clearance. Evaluation of the effects of CSA and MTX on BM not only confirmed the lack of a correlation between BM disease and hematologic parameters, but also confirmed the underlying resistance of the LGL clone to eradication19, 21, 23–26 and highlighted the importance of immunosuppression rather than cytotoxicity in its management.

Whether second malignancies developed solely because of underlying immune dysregulation, advanced age, and/or the intrinsic possibility of high-grade transformation of the LGL clone,37 or because of the combination of these features with long-term immunosuppressive therapy is not clear. The incidence of these malignancies highlights the need for clinical vigilance, particularly where combined, intensive immunosuppression is used.25 Our survival data were consistent with published reports showing an overall mortality rate of 10% to 28% over 4 years among patients with this leukemia12, 31, 38 and demonstrated that up to 25% of patients with T-cell LGL leukemia (Fig. 1) survive beyond 20 years.

We have shown the efficacy of 4 main agents (MTX, CSA, pentostatin, and cyclophosphamide) in patients with T-cell LGL leukemia with an absence of cross-resistance. An approach that uses CSA or MTX with cross-over for nonresponders, consolidated with cyclophosphamide or pentostatin in an attempt to circumvent the need for life-long therapy, may bear consideration for future prospective trials. The high intensity of CD52 expression on malignant LGL cells39 in conjunction with reports of its successful use in T-cell LGL leukemia29, 40, 41 supports the potential efficacy of alemtuzumab in this condition. More intense immunosuppression using reduced intensity allogeneic stem cell transplantation also may hold promise for suitable patients and/or for patients with refractory/recurrent disease.


  1. Top of page
  2. Abstract
  6. Acknowledgements

We are grateful to the many clinicians who have referred patients and/or samples over the years.


  1. Top of page
  2. Abstract
  6. Acknowledgements