Prospective, Multicentric, Phase II Randomized Controlled Trial Comparing the Efficacy and Toxicity of Methotrexate to Cyclophosphamide in Large Granular Lymphocyte Leukemia: A French National Study (NCT01976182)

Introduction

Large granular lymphocyte (LGL) leukemia is characterized by a clonal expansion of CD3+ cytotoxic T or CD3- NK cells. Clinical features include neutropenia, anemia and autoimmune-associated diseases such as rheumatoid arthritis (RA). So far, low dose methotrexate (MTX), oral cyclophosphamide (CP), or ciclosporin are the 3 main options for initial therapy. In 2013, we launched a prospective trial comparing CP to MTX in untreated patients in need of treatment.

Patients and methods

This was a multicentric, national (43 centers), open-label, randomized trial on two parallel groups comparing CP to MTX. Previously-untreated patients with LGL leukemia were included if they had at least one of the following indications of treatment: isolated severe neutropenia (ANC <0.5x10⁹/L) or neutropenia (ANC <1.5x10⁹/L) with infections, anemia requiring transfusions or symptomatic anemia, associated complications such as systemic diseases or auto-immune diseases resistant to steroids and/or immunomodulating agents (colchicin, disulone, hydroxychloroquine). Patients received either MTX (10 mg/m²/w) or CP (100 mg/d) for 4 months. Responders (CR+PR) at M4 continued the same treatment until M12 (CP was delivered at 50 mg/d by M8). Non-responders at M4 were randomly assigned to receive either ciclosporin (3 mg/kg/d) or the drug which had not been administered at the first randomization (MTX or CP). Response was assessed using previously published criteria (Lamy T, Blood 2011). To stop the trial as soon as sufficient information was collected, a sequential analysis was planned every 20 evaluable patients using the triangular test. The primary (binary) endpoint was the CR rate at M4.

Results

From Nov 2013 to Oct 2023, 166 patients met inclusion criteria among which 160 were randomized and 150 were evaluable at M4 (MTX: n=76; CP: n=74). The trial, which had to be continued after the 7^th sequential analysis, was stopped by the sponsor on the recommendations of both DSMB and Scientific Committee based on the differences observed between the two groups on CR and ORR at M4. Among the 150 evaluable patients, LGL subtypes were: T-LGL: n=134; NK-LGL: n=15; biphenotypic T/NK-LGL: n=1. The 2 arms were well balanced in terms of clinical and biological parameters. Baseline characteristics of patients were (% or median, range): gender: 70 male (47%), 80 female (53%); age: 64 years (23-86); Hb: 11.6 g/dl (5.4-16.0) with 27 (18%) transfusion-dependent patients; neutrophils: 1.3 G/L (0.1-10.7) with 99 (66%) neutropenic patients of whom 60 (40%) with recurrent infections; LGL count: 2.8 G/L (0-27.2); 52 (35%) patients had splenomegaly and 11 (7%) had autoimmune disease (RA: n=3; vasculitis: n=2; others: n=2; autoimmune cytopenia: n=4). STAT3, STAT5B, TNFAIP3, TET2 and DNMT3A mutations were analyzed on 140 patients and observed in 95 (68%), 5 (4%), 16 (11%), 14 (10%) and 9 (6%) patients, respectively. At M4, CR rate was 24% (18/74) in CP-arm vs. 12% (9/76) in MTX-arm (p=0.047) and ORR was 73% (54/74) in CP-arm vs. 32% (24/76) in MTX-arm (p<0.001). ORR were 58/95 (61%) and 14/44 (32%) in STAT3 mutated and wt patients, respectively.

Among the 72 refractory patients at M4, 63 had a second randomization: 15, in the CP-arm, received ciclosporin (n=9) or MTX (n=6), and 48, in the MTX-arm, received ciclosporin (n=23) or CP (n=25). Among them, 33 reached CR or PR at M12. ORR of CP after MTX failure was 64% (16/25) and ORR of MTX after CP failure was 33% (2/6). Fifteen out of the 32 (47%) patients receiving ciclosporin reached CR or PR at M12, 5/9 (56%) in the CP-arm and 10/23 (43%) in the MTX-arm. The ORR rate at M12 was 73% (CR: n=51; PR: n=48) among the 135 patients evaluable so far. With a median follow-up of 51 months, considering responders at M4 and M12 (patients who did not switch to other treatment between M4 and M12), 18% (8/44) in the CP-arm and 25% (4/16) in the MTX-arm required a new treatment from M12 with median times of 14.5 (2-35) and 22.5 (2-88) months, respectively. There were no differences in terms of toxicity between the two arms (grade 3-5 infections, secondary cancers). There were 30 deaths (with 4 deaths before M12) occurring at a median time of 33 months (3-100).

Conclusions

This prospective randomized controlled trial in previously-untreated patients with LGL leukemia shows that CP induces a significantly better response than MTX in terms of CR and ORR after 4 months of treatment without any difference in terms of toxicity or secondary malignancies.