Clinical Characteristics and Outcomes of Patients with Large Granular Lymphocytic Leukemia Treated with Methotrexate, Cyclophosphamide or Cyclosporine. a Large Single Institutional Retrospective Analysis

Introduction

Large granular lymphocytic leukemia (LGLL) is a rare mature clonal lymphoproliferative disorder derived from T-cell or natural killer (NK)-cell lineages. Even though LGLL has usually an indolent course, a majority of patients (pts) require therapy due to cytopenias, recurrent infections, autoimmune disorders or symptomatic splenomegaly. There is no standard frontline therapy established for pts with LGLL based on randomized prospective studies but a single-agent immunosuppressive therapy has been used for almost four decades. In the present study, we analyzed clinical and pathological characteristics, responses to therapy and outcomes of patients diagnosed at Moffitt Cancer Center.

Methods

A total of 303 consecutive LGLL pts diagnosed between 2001 and 2017 were included in the study. Pts who did not meet the WHO diagnostic criteria of peripheral blood (PB) LGL count ≥2.0 k/ μL but presented with smaller, persistent LGL clonal populations along with characteristic immunophenotype and clinical features such as cytopenias, recurrent infections, autoimmune disorders, splenomegaly, and bone marrow involvement with identical clonal LGL population were considered to have LGLL. Responses to treatment were categorized as complete response (CR), partial response (PR) and no response. CR was defined as Hb>12g/dL, ANC≥1.5k/μL, PLT>150k/μL ALC<4 k/μL and LGL<0.5K/μL. PR was defined as Hb>8g/dL, ANC>0.5K/μL, PLT>50k/μL, and no transfusion requirement. Cox proportional hazard mode was used for multivariate survival analysis with all the covariates listed in table 3.

Results

The median age at diagnosis was 65 years (range: 21 – 90 years). 92.1% of pts were diagnosed with T-LGLL and 7.9% with NK-LGLL. At presentation, 50.4% of pts had anemia; 48.4% had neutropenia and 34.6% had thrombocytopenia. 21.7% of pts were transfusion dependent; 20.3% had severe neutropenia (ANC < 0.5K/μL) and 33% had splenomegaly. Preexisting diagnosis of autoimmune disorders was observed in 26.4% of pts (Table 1).

The Foundation One Heme NGS assay (Foundation Medicine Inc, Cambridge, Massachusetts, USA) was employed in 25 pts (www.foundationmedicine.com/genomic-testing/foundation-one-heme). Nine of 25 pts were positive for STAT 3 and none for STAT5 mutations.

54.8% of pts required treatment. Table 2 summarizes the first three lines of therapy. Methotrexate (MTX), cyclophosphamide, and cyclosporine A (CSA) were the most commonly used front-line therapies. Overall response (CR + PR) for all 3 agents ranged between 55%-67% with CR rate of 11%-25%. 25% of pts receiving Cytoxan achieved CR and 11.7% on MTX achieved CR however statistical testing was not significant due to the limitation of sample size. Rates of overall response (OS) were not different among the treatment regiments. The rate of 10-year overall survival (OS) was 56.8%. Median OS was 14 years [95% CI (9.7 years, .) (Figure 1).

In the multivariate Cox model, increasing age, neutropenia, thrombopenia and coexisting malignancies were independent factors associated with poor survival. Absolute number of LGLL cells in blood was not associated with OS. T-LGLL and NK-LGLL had comparable OS (Table 3).

Conclusions

This study is one of the largest retrospective single-institutional studies of LGLL. Consistent with prior studies, LGLL demonstrated an indolent course with a long median OS. ORR and CR rates did not significantly differ among cyclophosphamide, MTX and CSA. However, a significant percentage of patients did not respond to the frontline immunosuppressive therapy. Clinical trials with novel therapeutic agents are necessary to further improve outcomes of patients with LGLL.