Changes in CD8⁺TRB repertoires are associated with durable remission post-DLI. (A) Comparison of CD8⁺TRB diversity in pre-DLI samples (filled circles) and d+14 post-DLI samples (filled squares) in patients without relapse (noR) post-DLI during the study follow-up (blue; n = 9) and patients with relapse (R) post-DLI (red; n = 9). (B) Change of CD8⁺TRB diversity in patients without (blue) and with relapse (red) in percent is displayed. (C) Number of unique clonotypes (y-axis) is displayed individually for each quartile of the CD8⁺ repertoire, comparing patients without relapse post-DLI (blue; n = 9) with patients developing relapse post-DLI (red; n = 9). Two time points are given for each group, with closed circles displaying the pre-DLI time point and closed squares d+14 post-DLI. (D) Difference in number of clones (y-axis) between pre-DLI and d+14 time point is displayed individually for each quartile of the CD8⁺ repertoire, comparing patients without relapse post-DLI (blue; n = 9) with patients developing relapse post-DLI (red; n = 9). (E) CD8⁺TRB repertoire proportions are displayed for the top 20, 21 to 50, and 51 to N CD8⁺ T-cell clones, comparing patients without relapse post-DLI (blue; n = 9) with patients developing relapse post-DLI (red; n = 9). Two time points are given for each group, with closed circles displaying the pre-DLI time point and closed squares d+14 post-DLI. (F) Difference in CD8⁺TRB repertoire (y-axis) occupied by top 20, 21 to 50, and 51 to N clones between pre-DLI and d+14 time point. (G) Top 20, 21 to N, and total overlapping CD8⁺ T-cell clones as percent of total CD8⁺TRB repertoire (y-axis) is shown for patients without development of relapse post-DLI (blue; n = 9) and patients developing relapse post-DLI (red; n = 9). Two time points are given for each group, with closed circles displaying the pre-DLI time point and closed squares d+14 post-DLI. (H) Change of overlapping CD8⁺TRB repertoire proportions is displayed for the top 20, 21 to N, and total overlapping CD8⁺ T-cell clones, comparing patients without relapse post-DLI (blue) with patients developing relapse post-DLI (red). Statistical analysis was performed by Wilcoxon matched-pairs signed ranked test (two-tailed) (A, C, E, G) and Mann-Whitney test (two-tailed) (B, D, F, H). Black lines represent median; error bars show the interquartile range. *P < .05; **P < .01; ns, not significant.