Predictive and Prognostic Significance of Comprehensive Genomic Profiling in Patients with Diffuse Large B-Cell Lymphoma

Introductions:

We assessed the benefit of adding comprehensive genomic profiling (CGP) to improve clinical outcomes in both first-line and relapse/refractory settings in adult patients with diffuse large B-cell lymphoma (DLBCL). Current standard first-line therapy affords long term remission in only 55-60% of patients, and less than half of these cases can be treated successfully following relapse. Thus, there is a clear unmet clinical need to improve the standard of care for all patients with DLBCL. We interrogated specimens from 125 patients with DLBCL using a novel next generation sequencing (NGS)-based CGP assay (FoundationOneŽ Heme) to identify potential clinically relevant therapeutic targets and improve risk stratification.

Methods:

Genomic profiles from 125 DLBCL specimens received in a CLIA-certified, CAP-accredited, NYS-approved laboratory were successfully characterized by FoundationOneŽ Heme. Sequencing libraries targeting 405 cancer-related genes by DNA-seq and 265 frequently-rearranged genes by RNA-seq.  All captured libraries were sequenced to high depth (Illumina HiSeq), averaging 498x for DNA and ~7M on-target unique pairs for RNA, to enable the sensitive and specific detection of substitutions, indels, copy number alterations and gene rearrangements. Additionally, relative mRNA expression of MYC, BCL2 and BCL6 was quantified by normalizing with the median ppm of each gene in the cohort. The statistical significance of expression between the rearrangement positive cohort and rearrangement negative cohort was determined by one-sided Wilcoxon rank-sum test.

Results:

Somatic driver alterations were identified in all 125 cases (including 28 relapse/refractory patients, 28 transformed from low grade lymphoma and remainder are being assessed) with clinically relevant genomic alterations observed in 97 (78%) patients. The cohort includes 79 male and 46 female patients with a median age of 63.  The mean mutation burden averaged 6.5 alterations per patient.  Genomic alterations previously described in DLBCL were confirmed in this study, including changes in MLLT3 (39.2%), TP53 (31.2%), CDKN2A (30.4%), BCL2 (28.8%), BCL6 (20%), CREBBP (20%) and MYC (12%). Importantly, alterations in several genes with known therapeutic associations were also detected, including EZH2 (11.2%), TET2 (8.8%), CD79 (7.2%) and PTEN (4.8%) (Fig 1). Therapeutically targetable kinase mutations classically associated with solid tumors (e.g. RET, ALK, BRAF, HRAS, NRAS, FGFR2 and ERBB2 amp) were also identified, though at lower frequency (<3%).   

Strong correlation was observed between genomic rearrangements and over-expression in MYC (p=4.64e-06) and BCL2 (p=1.39e-05) but not in BCL6 (p=6.64e-02) from the integrated NGS-based platform. In 10 samples, focal CNA or point mutation leads to overexpression of BCL,2 and point mutations occur in 18/125 (14%) of patients and co-occur with BCL2 rearrangement significantly (p < 1e-04) (Fig 2).

Conclusions:

Comprehensive genomic profiling identifies a complex genomic landscape in DLBCL that may impact diagnosis and prognosis, deepen our understanding of underlying biologic pathways, and guide targeted combination therapy in relapsed patients with limited treatment options.  Additionally, our data suggests that focal copy number amplification, point mutations and rearrangements can serve as distinct/interactive mechanism to potentially activate MYC and BCL2 and may explain the challenges currently encountered in subclassification and prognostication (e.g. double-hit) of DLBCL. Further functional validation of these MYC, BCL2 and BCL6 alterations through comparing orthogonal clinical test results and clinical outcomes could help further refine the subclassification in DLBCL.