Polynomic curve interpolation of MRD results per patient and per relapse mutation to model relapse dynamics. Full cohort of 74 patients (median, 3 trackable mutations per patient; range, 1-7), mutations in DTA and non-DTA genes were used as MRD markers. Relapse mutations (n = 213) were, if possible, repetitively monitored by NGS-MRD (total number of prerelapse NGS analyses n = 296; median, 1; range, 1-6). MRD status of each marker or each patient was assumed negative until the first sample was measured positive. MRD positivity was assumed for all time intervals following an MRD-positive result until relapse. By calculating the number of positive patients or mutations in a given interval as a fraction of the total number of patients or mutations, we were able to compare relapse kinetics patterns for patients () and mutations (), respectively. , .
