Clonal fraction of 2q37 COPS7B/COPS8 loss increases only during IMiD-based therapies. (A) Schematic showing sequential sample numbers analyzed at each state of LEN or POM exposure/refractoriness. (B) Summary table of incidence of patients acquiring 2q37 loss (filtered to include loss of COPS7B and/or COPS8 containing-regions only) during IMiD-based therapy vs no IMiD exposure. (C) “Fishplot” diagrams demonstrating behavior of 2q37 loss-containing subclones (green colored) over sequential samples, in relation to other CNA-defined subclones and drug exposure. Total tumor burden over time (gray area) is derived from serial serum M-protein/restricted free light chain measurements (taken at white vertical line timepoints). Only IMiD-containing drug regimes are marked. Subclone emergence points have been inferred by linear growth assumptions. Note behavior of subclones during intervening months/years between WGS sampling points (taken at black vertical line timepoints) is inferred, for example in plot (v), where precise max CCF of 2q37-containing subclone reached, and timing of its expiration is unknown. This plot shows 1 likely scenario. (D) For 2 of the cases in (C), CRBN immunohistochemistry (IHC) of bone marrow biopsies from same timepoints as the WGS data are shown, plus κ (κ) and Λ (λ) light chain (LC) in situ hybridization (ISH) to indicate tumor burden. Graphs show corresponding CRBN protein quantification (percentage of cell surface stained) across disease stage, each point representing 1 cell. Significance as shown determined by Mann-Whitney (case 2, 2 timepoints) and Kruskal-Wallis (case 3, 3 timepoints) tests for nonparametric data.