Genomic Analysis of Diffuse Large B Cell Lymphoma in a Japanese Cohort Using Targeted DNA Sequencing

Introduction: Recent studies based on next-generation sequencing revealed the genetic landscape of diffuse large B cell lymphoma (DLBCL). A new genetic classification system was proposed, including four groups; the BCL2 and SGK1 groups that were both associated with germinal center B-cell like (GCB) DLBCL, an MYD88 group that was associated with non-GCB DLBCL, and a NOTCH2 group that was not associated with a specific cell-of-origin (COO). Unfortunately, this classification method is somewhat complex and thus not fully accepted in clinical practice. Furthermore, part of all cases remained unclassified; as such, it is clear that the diversity of the genetic landscape of DLBCL is not yet fully understood. In this study, we performed a retrospective analysis of genetic abnormalities and their correlations with clinical characteristics in a Japanese cohort of patients diagnosed with DLBCL.

Materials and Methods: We collected clinical data from 224 of a total of 234 patients who were diagnosed with DLBCL between 2013 and 2018 at Chiba University Hospital. We also collected formalin-fixed paraffin-embedded (FFPE) specimens obtained from 204 patients via standard diagnostic procedures. The COO for each sample was assessed using the Hans algorithm with immunohistochemical staining; rearrangements of BCL2, BCL6, and MYC were evaluated by fluorescence in situ hybridization. Targeted DNA sequencing was performed using a 144-gene custom panel.

Results: Of the 224 DLBCL patients evaluated in our study, the median age at diagnosis was 69 years (range, 18 - 92 years). The median follow-up time was 34 months; three-year progression-free survival (PFS) was 63.2%, and three-year overall survival was 78.0%. COO was determined in 218 cases; 93 of these cases (41.5%) were classified as GCB type and 125 cases (55.8%) as non-GCB type. We successfully performed targeted DNA sequencing in 167 of the 204 FFPE samples. Compared to previous reports, our cohort included relatively higher frequencies of mutations in PIM1 (44.3%), KMT2D (40.7%), MYD88^L265P (37.7%), and CD79B (24.5%); by contrast, we detected comparatively lower mutation frequencies in B2M (6.6%), ETV6 (5.4%), CDKN2A (4.7%), and TNFAIP3 (4.1%). Multivariate adjustment of genes extracted by univariate analysis and factors identified by International Prognostic Index score revealed that genetic mutations of MYD88^L265P (Hazard ratio [HR] = 2.045), IL16 (HR = 4.848), and BCOR (HR = 6.295) were significant prognostic factors. Mutations in IL16 and BCOR are novel factors that will be used to predict poor prognosis. In the 167 sequenced cases, 110 were classified into the aforementioned four subtypes based on mutations detected in 23 genes together with rearrangements of BCL2 and BCL6; specifically, 47 cases were classified as members of the MYD88 group, 28 in the BCL2 group, 19 in the NOTCH2 group, and 16 in the SGK1 group. We identified a trend toward differential PFSs among the four groups (p = 0.18); three-year PFS was 51.2% among patients with MYD88 group, 59.6% in the BCL2 group, 63.2% in the NOTCH2 group, and 75.0% for the SGK1 group. When compared outcomes of patients in the BCL6 group, patients in the MYD88 group had significantly inferior PFS (p = 0.043) among the non-GCB DLBCL patients; those in the BCL2 group revealed a trend toward inferior PFS compared with those in the SGK1 group (p = 0.34) among the GCB DLBCL patients.

We next focused on 28 cases of CD5-positive DLBCL; this is a distinct DLBCL subtype with poor prognosis that is identified more frequently among Asians. We observed significantly higher frequencies of frameshift or nonsense mutations in CD58 and missense mutations in MYD88^L265P. Furthermore, evaluation of gene expression data for CD5-positive DLBCL reported in a published database (Miyazaki K et al. Int J Hematol. 2015) revealed diminished expression of CD58 (p = 0.01) and augmented expression of MYD88^L265P (p < 0.001). These results suggest that there may be a specific correlation between genetic mutations and gene expression. Intriguingly, these specific genes are key regulators of the immune response, suggesting that inflammation contribute to the unfavorable outcome of CD5-positive DLBCL.

Conclusion: Our study provides support for the genetic classification of DLBCL using targeted DNA sequencing to evaluate outcomes in our Japanese cohort. We also identified mutations in novel candidate genes that may be used to predict clinical outcomes.