Relationship between expression levels of miR-155, ZAP-70, IGHV mutation status, and TFS. (A) In the training dataset, Kaplan-Meier curves depicting the TFS probability over time from diagnosis of patients who were segregated into 2 groups (miR-155-Lo or miR-155-Hi) according to the relative amounts of miR-155 expressed by the blood CLL cells of each patient. (B,C) Kaplan-Meier curves depicting the TFS probability over time from diagnosis of patients segregated by miR-155 and ZAP-70 status (B), or miR-155 and IGHV mutation status (C). (D) In the validation dataset, Kaplan-Meier curves depicting the TFS probability over time from diagnosis for patients who were segregated into 2 groups (miR-155-Lo or miR-155-Hi) based upon whether or not the blood CLL cells expressed miR-155 above the training set–defined threshold for high miR-155HG. (E,F) In the validation dataset, Kaplan-Meier curves depicting the TFS probability over time from diagnosis for patients segregated by miR-155 and ZAP70 status (E) or miR-155 and IGHV status (F). Statistical significance was determined by log-rank test (P < .05). P values for the comparisons between subgroups are indicated below the graph for panels B, C, E, and F.