The landscape of copy number variations in classical Hodgkin lymphoma: a joint KU Leuven and LYSA study on cell-free DNA

Abstract

The low abundance of Hodgkin/Reed-Sternberg (HRS) cells in lymph node biopsies in classical Hodgkin lymphoma (cHL) complicates the analysis of somatic genetic alterations in HRS cells. As circulating cell-free DNA (cfDNA) contains circulating tumor DNA (ctDNA) from HRS cells, we prospectively collected cfDNA from 177 patients with newly diagnosed, mostly early-stage cHL in a monocentric study at Leuven, Belgium (n = 59) and the multicentric BREACH study by Lymphoma Study Association (n = 118). To catalog the patterns and frequencies of genomic copy number aberrations (CNAs), cfDNA was sequenced at low coverage (0.26×), and data were analyzed with ichorCNA to yield read depth-based copy number profiles and estimated clonal fractions in cfDNA. At diagnosis, the cfDNA concentration, estimated clonal fraction, and ctDNA concentration were significantly higher in cHL cases than controls. More than 90% of patients exhibited CNAs in cfDNA. The most frequent gains encompassed 2p16 (69%), 5p14 (50%), 12q13 (50%), 9p24 (50%), 5q (44%), 17q (43%), 2q (41%). Losses mostly affected 13q (57%), 6q25-q27 (55%), 4q35 (50%), 11q23 (44%), 8p21 (43%). In addition, we identified loss of 3p13-p26 and of 12q21-q24 and gain of 15q21-q26 as novel recurrent CNAs in cHL. At diagnosis, ctDNA concentration was associated with advanced disease, male sex, extensive nodal disease, elevated erythrocyte sedimentation rate, metabolic tumor volume, and HRS cell burden. CNAs and ctDNA rapidly diminished upon treatment initiation, and persistence of CNAs was associated with increased probability of relapse. This study endorses the development of ctDNA as gateway to the HRS genome and substrate for early disease response evaluation.

Graphical abstract

Figure 1.

Detection and FISH validation of CNAs in cfDNA. (A) Illustrative ichorCNA output for 2 cases with a high (left) vs low (right) estimated clonal fraction. Genome-wide log2 ratios are plotted and colored according to CN status (green = loss; blue = copy neutral; brown = gain; red = amplification). (B) Validation of CNAs in the case of HL19 by interphase FISH on fixed cytogenetic specimens. At the left, the log2 plot for the chromosome of interest is shown, with an arrow indicating the genomic location of the FISH probe used. Probes are shown with their respective colors (green [G]: spectrum green; red [R]: spectrum orange; blue [A]: spectrum aqua; red/green [DC]: dual color break apart probe) and the FISH patterns of the large HRS cells are indicated. (C-D) Validation of novel recurrent CNAs in cfDNA by interphase FISH: loss of 3p13-p26 (C) and gain of 15q21-q26 (D). CEP, centromeric probe; LSI, locus specific identifier.

Figure 2.

Recurrent CNAs in cHL. (A) Recurrent CNAs in abnormal cases (n = 164) were evaluated using GISTIC. The y-axis shows the G score as a function of chromosomal location (x-axis) for gains (red) and losses (blue). G scores exceeding the green line represent significantly recurrent aberrations. The latter are shown with their frequency within this cohort, their genomic location, and their size. CancerMine was browsed for tumor suppressor and oncogenes involved in “hematologic cancers,” “lymphoma,” “B-cell lymphoma,” “Hodgkin lymphoma,” “non-Hodgkin lymphoma,” and “diffuse large B-cell lymphoma.” GO gene lists were extracted for the JAK-STAT pathway (GO:0007259), NF-kB pathway (GO:0038061 and GO:0007249), Notch signaling (GO:0007219), TNFR signaling (GO:0033209), cytokine-mediated pathway (GO:0019221), negative and positive regulation of apoptotic signaling pathway (GO:2001234 and GO:2001235), and B-cell receptor signaling (GO:0050853). A selection of genes that are located within the wider boundaries as determined by GISTIC is shown. Genes within the minimally common region (see “Materials and methods”) are printed bold. (B) Broad CNAs are defined as comprising at least 90% of a chromosome arm. Frequencies (%) are plotted for significantly gained (red) or lost (blue) p or q arms.

Figure 3.

ctDNA concentrations at diagnosis. (A) Box and whisker plot of baseline ctDNA concentrations for controls vs cHL patients in general and grouped by disease stage category. (B) ROC analysis for ctDNA concentrations at diagnosis determines the optimal ctDNA cutoff (Youden index) between cHL cases from controls (blue). Corresponding sensitivity and specificity are shown in blue. (C-F) Box and whisker plots of baseline ctDNA concentrations in cHL patients with treatment response vs failure (C), and for early-stage disease: absence vs presence of B symptoms (D), normal vs elevated ESR levels (E), and limited vs extensive nodal involvement (F). Outliers are defined as values 1.5 × IQR below the first or above the third quartile. *P < .05; **P < .01; ***P < .001. AUC, area under the curve; CR, complete remission.

Figure 4.

ctDNA dynamics after treatment initiation. Line plot showing the individual evolution of ctDNA concentration under treatment. Superimposed are box and whisker plots of ctDNA concentrations at each time point to highlight the general tendency.