Large Granular Lymphocytic Leukemia with Coexisting Malignancies

Introduction:

Large granular lymphocytic leukemia (LGLL) is a rare, indolent malignancy arising from cytotoxic T cells (~85%) or NK cells (~15%). The coexistence of LGLL with MDS, B cell neoplasms and plasma cell dyscrasias have been reported in case reports and two retrospective studies (Gayal et al 2018, Viny et al 2008). However, those studies were limited by sample size and a small number of coexisting malignancies. To our knowledge, there is no systematic study on coexisting solid tumors with LGLL in the literature.

In present study, we systemically analyzed the coexisting malignancies (CMs) including solid tumors and hematological malignancies (liquid tumors) in our LGLL patients (pts). We compared the clinical features, immunophenotype, treatment responses and estimated outcomes of CM-associated LGLL and non-CM associated LGLL pts.

Methods:

Total 303 consecutive LGLL pts evaluated at Moffitt Cancer Center (MCC) 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 persistent LGL clonal populations along with characteristic immunophenotype, matching clinical features and bone marrow involvement with clonal LGL population were considered to have LGLL.

Results:

Among the total 303 LGLL patients, 123 (40.6%) had at least 1 CM. The maximum number of coexisting malignancies were 3. 114 pts (37.6%) had invasive cancers (excluding non-melanomatous skin cancer), 83 (27.1%) had solid tumors and 60 (19.8%) had liquid tumors. The most frequent invasive solid tumors in our LGLL patients were: prostate cancer, breast cancer, lung cancer, colorectal cancer, and melanoma. The most frequent liquid tumors were: B-cell lymphoma, T-cell lymphoma, CLL/SLL, CML and MM (Table 1).

The number of patients with CM diagnosed before, at the same time and after LGLL diagnosis were 103, 7, and 28, respectively.

Table 2 compared the clinical features of LGLL pts with CM to those without CM separated into liquid and solid tumors. Compared to pts without CM, LGLL patients with liquid tumors were older, less likely to have B symptoms (49.2% vs 73.6%, respectively), less likely to be neutropenic (32.1% vs 53.4%), less likely to be pRBC transfusion dependent (7.3% vs 24.5%), and less likely to require treatment (33.3% vs 61.1%). LGLL pts with solid tumors compared to those without CM were older, more likely to be Caucasians, more likely to have NK-cell LGLL (13.3% vs 6.1%), and less likely to have B symptoms (59.5% vs 73.6%).

NGS testing was performed in 25 patients. Activating STAT3 mutations were detected in 9/21 patients without CM and 1/4 patients with CM (p=0.63 by Fisher’s exact test). None of tested pts demonstrated STAT5B mutation.

An extensive 10-color flow cytometry panel was employed after year 2015 in 34 pts. The immunophenotype was similar between CM-associated T-LGLLs and non-CM associated T-LGLLs except that CM-associated T-LGLL pts were more likely to be CD56+ (6/11 vs 3/23, p=0.03 by Fish’s exact test).

CM-associated LGLLs and non-CM associated LGLLs had comparable response rates to treatments (Table 3).

The median follow-up time for all patient population was 4.8 years. Pts with CM had worse overall survival (OS) (Figure 1). The survival difference was significant after adjusting for age, gender and race (data not shown).

Conclusions:

In the present study of 303 consecutive LGLL pts, 40.6% had CM, including 27.4% with solid tumors and 19.8% with liquid tumors. CM-associated LGLLs had similar clinical and pathologic features and treatment responses as non-CM associated LGLLs, except that LGLL pts with CM were less likely to be symptomatic and less likely required treatment. Earlier detection of LGLL in pts with liquid tumors may have contributed to those differences. Interestingly, T-cell LGLLs with CM were more likely to be CD56 positive and LGLLs associated with solid tumors were more likely to be derived from NK-cell lineage.

The high incidence of CMs in LGLL can be explained by the following hypotheses: 1) CM provided an initial antigenic stimulus and contributed to the pathogenesis of LGLL or alternatively, 2) host immunosuppression due to clonal LGLL proliferation promoted development of CM. The findings in this study that a majority of CMs manifested prior to LGLL diagnosis favor the first hypothesis. Our results provide additional insights into the pathogenesis of LGLL.