Low Peripheral Blood NK Cell Count Is Associated with Worse Clinical Outcome in Patients with Follicular Lymphoma (FL) and Diffuse Large B-Cell Lymphoma (DLBCL) Treated with Immunochemotherapy: Results from the Frontline Phase 3 GALLIUM and GOYA Trials

Introduction: Natural killer (NK) cells are key elements of the innate immune system. Obinutuzumab (GA101; G) is an anti-CD20 monoclonal antibody with enhanced direct cell death activity and antibody-dependent cellular cytotoxicity (ADCC) vs rituximab (R). Upon binding CD20-bound G/R, NK cells are activated and attack the target cell. Therefore, low pre-treatment NK cell count (NKCC) may be associated with worse outcomes in G/R-treated patients (pts). This exploratory post hoc analysis evaluated the prognostic impact of baseline (BL) NKCC in FL pts treated with G/R plus chemo in the Phase 3 GALLIUM trial (NCT01332968) and DLBCL pts treated with G/R plus CHOP in the Phase 3 GOYA trial (NCT01287741).

Methods: BL peripheral blood (PB) NKCCs, assessed by flow cytometry (CD3-CD56+CD16+ cells; normal range [central lab]: 95-640 cells/µL), were available in 1064/1202 (88.6%) FL pts and 1287/1418 (90.8%) DLBCL pts. Cell-of-origin (COO) was determined in 933/1418 DLBCL pts using the Nanostring Research Use Only Lymphoma Subtyping Test (LST). COO and BL NKCC were available in 857/1418 pts. Whole transcriptome gene expression was analyzed using TruSeq RNA sequencing in tumor tissue from 552/1418 DLBCL pts. A 57-gene signature designed to reflect NK cell tumor infiltration was applied to RNA sequencing data from 552 pts; median score was used to define high/low subgroups. Kaplan-Meier methodology was used to estimate progression-free (PFS) and overall survival (OS), and a Cox regression univariate model was used to estimate corresponding HR and CI. The relative importance of BL variables (NKCC, gender, geographic region, treatment arm, chemo backbone/no. of planned chemo cycles, FLIPI/IPI, extranodal/bone marrow involvement, sum of products of diameter, Ann Arbor stage) was evaluated using multivariate (MV) Cox regression models with a stepwise approach.

Results: Median (range) BL NKCCs were 220 cells/μL (0-3300) in FL and 200 cells/μL (0-1900) in DLBCL pts. Overall, 108/1064 (10.2%) FL pts and 255/1287 (19.8%) DLBCL pts had low BL NKCC (<100 cells/μL). By COO subtype, 83/485 (17.1%) germinal center B-cell-like (GCB), 37/140 (26.4%) unclassified, and 54/232 (23.3%) activated B-cell-like (ABC) DLBCL pts had low BL NKCC. BL disease characteristics of pts with low vs normal NKCCs are shown in Table 1. Low BL NKCC was associated with advanced disease.

On univariate analysis low BL NKCC was associated with shorter PFS in FL (HR 1.57, 95% CI 1.10-2.25, p=0.01; 3-yr PFS 71.6% vs 80.1%) and DLBCL (HR 1.36, 95% CI 1.07-1.72, p=0.01; 3-yr PFS 62.8% vs 70.0%; Figure 1), and shorter OS in FL (HR 2.58, 95% CI 1.51-4.42, p=0.0003; 3-yr OS 87.6% vs 94.3%). DLBCL pts with low NKCC had a trend to worse OS vs pts with normal NKCC (HR 1.35, 95% CI 1.00-1.82, p=0.052; 3-yr OS 77.6% vs 82.3%). On MV analysis, low BL NKCC was independently associated with PFS in FL (HR 1.48, 95% CI 1.02-2.14, p=0.04) and DLBCL (HR 1.36, 95% CI 1.01-1.83, p=0.04). The DLBCL result appeared to be driven by COO subtype, with the highest estimated HR in GCB (HR 1.58, 95% CI 1.0-2.5, p=0.05), and no effect in unclassified and ABC. Interestingly, in line with the enhanced ADCC of G, the impact of low BL NKCC on PFS was stronger in G-treated FL pts (HR 2.06, 95% CI 1.24-3.41, p<0.01) vs R-treated pts (HR 1.19, 95% CI 0.71-1.99, p=0.5), while less pronounced in the GCB DLBCL subgroup (G-treated pts: HR 1.25, 95% CI 0.91-0.77, p=0.182; R-treated pts: HR 1.47, 95% CI 1.06-2.06, p=0.021; Figure 2).

Although there was no correlation between PB NKCC and tumor NK cell gene expression among biomarker-evaluable pts in GOYA, low tumor NK cell gene expression was associated with shorter PFS in G-treated DLBCL pts (all COO subtypes; HR 1.95, 95% CI 1.2-3.1, p<0.01), with a trend for low CD56 mRNA expression alone (as continuous variable) correlating with shorter PFS (p=0.043).

Conclusions: In this largest prospective collection of PB NK cells to date in FL and DLBCL, substantial numbers of pts had reduced NKCCs at BL, which was associated with advanced disease. Univariate and MV analyses may suggest that low PB NKCC is independently associated with shorter PFS in FL and DLBCL and shorter OS in FL. Likewise, low NK cell gene expression in tumor tissue was associated with shorter PFS in G-treated DLBCL pts. Collectively, our results from this exploratory analysis indicate that the number of NK cells in PB and tumor tissue may impact clinical outcome of non-Hodgkin lymphoma pts treated with anti-CD20 antibodies.

View largeDownload slide

View largeDownload slide

Close modal