Longitudinal investigation of CH mutations in 47 patients. (A) Overview of samples for longitudinal investigation. (B) Evolution of CH mutation at the 4 time points. Black horizontal lines represent medians; boxes, interquartile ranges (IQRs); and vertical lines, 1.5 × IQRs. Percentages at the top indicate the relative change in VAF size of the CH mutations in paired samples. P values were calculated by Wilcoxon rank sum test for paired samples. (C) Evolution of DNA repair mutations (red) and non-DNA repair mutations (turquoise) at the 4 time points. Black horizontal lines indicate medians; boxes, IQRs; and vertical lines, 1.5 × IQRs. P values indicate the differences in VAF change between the 2 groups (for paired samples only) by Mann-Whitney U test. (D) Evolution of DNA repair (red) and non-DNA repair (turquoise) mutations during chemotherapy-free follow-up. P values indicate the comparison of the coefficients of the 2 linear regression models for the DNA repair and non-DNA repair VAFs, respectively. (E) Illustrates the presence of posttreatment CH mutations at the time point before any chemotherapy was administered. Closed circles represent mutations detected by next-generation sequencing (NGS); Xs, mutations detected by droplet digital polymerase chain reaction (ddPCR); and the open circle, the 1 undetected mutation. Gray area represents the range of detection limits for the ddPCR assays; red, DNA repair mutations; turquoise, non-DNA repair mutations. BEAC, carmustine, etoposide, cytarabine, and cyclophosphamide; BEAM, carmustine, etoposide, cytarabine, and melphalan; FU, follow-up; R hd AraC, rituximab plus high-dose cytarabine; R-maxi-CHOP, rituximab, cyclophosphamide (1200 mg/m2), doxorubicin (75 mg/m²), vincristine (2 mg), prednisolone (100 mg) day 1-5.