Genomic Complexity in Chronic Lymphocytic Leukemia Defined By Array-Based Analysis: Definitions, Associations with Other Biomarkers and Clinical Impact

Background: The advent of efficacious but also very expensive novel treatment regimens for high-risk chronic lymphocytic leukemia (CLL), has highlighted the need for accurate risk-stratification. Published evidence indicates that classic chromosome banding analysis (CBA) offers valuable complementary information to the routinely performed FISH analysis in CLL, potentially improving risk assessment. Indeed, recent studies identified complex karyotype (CK; defined as ≥3 abberations) as an independent predictive marker for refractoriness to ibrutinib and venetoclax, although for CLL, a precise evidence-based definition of genomic complexity (GC) is yet to be made. Array-based (CGH/SNP) analysis is capable of globally assessing the CLL genome, does not require in vitro mitogens, offers increased detection depth, and allows for easier standardization.

Aims: 1. Determine a prognostic relevant definition of GC in CLL by array-based analysis. 2. Head-to-head comparison of the sensitivity of CBA versus array-based analysis.

Methods: 1911 CLL and monoclonal B-cell lymphocytosis (MBL) patients (CLL=1895, 99% and MBL=16, 1%, respectively) were analyzed with CGH or SNP arrays mostly performed within the first year from diagnosis and before treatment administration (66% and 84%, respectively). For a subgroup (N=206) both CBA and array results were available (<1 year in between analyses). Main features of the studied cohort: median age: 60.7 years, males: 1252 (68%), Binet: A/B/C:623/324/141 (57%/30%/13%), IGHV-mutated CLL (M-CLL): 536/1082 (50%), del(13q): 1020/1910 (53%), del(11q): 353/1911 (19%), TP53 aberrations (TP53 abs, deletions and/or mutations of the TP53 gene): 205/1911 (11%). Recurrent chromosomal aberrations i.e. del(13)(q14), del(11)(q22.3) and del(17)(p13.1) defined by array were included in all analyses irrespective of size, while the cut-off for other aberrations are indicated.

Results: First, analysis was performed using a size cut-off ≥5Mb according to Schoumans et al (Genes Chrom Cancer 2016). Applying the criteria ≥3 structural and/or numerical aberrations to define GC, 451/1911 cases (24%) displayed GC analyzed with CGH/SNP arrays. GC was significantly related (p<0.001) with advanced clinical stage (Binet B/C), IGHV-unmutated CLL (U-CLL), del(11q), del(13q) and TP53 abs. On univariate analysis, GC was associated with inferior OS (median: 10.0 vs 12.2 (p<0.001) years). GC cases were stratified into those with 3 ('low-GC', N=174, 39%), 4 ('intermediate-GC', N=103, 23%) and ≥5 ('high-GC', N=174, 39%) aberrations. High-GC cases differed significantly (p<0.05) from the other two subgroups, being enriched for U-CLL and TP53 abs (84% and 45%, respectively). The median OS was 5.8 years for the high-GC vs 10 years for the low- and 8.1 years for the intermediate-GC group (p<0.01). Only 'high-GC' remained significant (p<0.01) on multivariate analysis along with TP53 abs and IGHV status (see table).

Lowering the array cut-off to 1Mb (N=353) resulted in detection of a higher number of aberrations (3.04 vs 2.34) compared to 5Mb cut-off (95% CI paired differences 0.475-0.924). Although a lower array cut-off resulted in upstaging of high risk GC (22 vs 12%), the median OS in the 'high risk' group was increased as compared to the cut-off of 5Mb (8.4 vs 5.5 years).

Array-analysis (5Mb cut-off) showed a higher number of total detected aberrations as compared to CBA (2.35 vs 1.84, 95% CI paired differences 0.221-0.798). With array more patients were defined as GC with higher discriminatory power regarding survival (concordance index: 58 vs 55%) than with CBA. Prognostic impact of array vs CBA and of specific aberrations is currently being analyzed.

Conclusion: According to this largest performed array-based analyses, current definitions for CK defined by the presence of ≥3 structural or numerical aberrations based on classical cytogenetic approaches may be suboptimal when applied for cases analyzed by array techniques. Instead, ≥5 aberrations detected by arrays were linked to high-risk disease, independent of clinical stage, IGHV and TP53 status. Since a lower array cut-off resulted in the detection of additional aberrations, without however been accompanied by the same dismal outcome as the 5Mb cut-off, further studies are warranted in order to define the actual prognostic relevance of smaller aberrations.

View largeDownload slide

View largeDownload slide

Close modal