Risk stratification of chromosomal abnormalities in chronic myelogenous leukemia in the era of tyrosine kinase inhibitor therapy

Abstract

Clonal cytogenetic evolution with additional chromosomal abnormalities (ACAs) in chronic myelogenous leukemia (CML) is generally associated with decreased response to tyrosine kinase inhibitor (TKI) therapy and adverse survival. Although ACAs are considered as a sign of disease progression and have been used as one of the criteria for accelerated phase, the differential prognostic impact of individual ACAs in CML is unknown, and a classification system to reflect such prognostic impact is lacking. In this study, we aimed to address these questions using a large cohort of CML patients treated in the era of TKIs. We focused on cases with single chromosomal changes at the time of ACA emergence and stratified the 6 most common ACAs into 2 groups: group 1 with a relatively good prognosis including trisomy 8, -Y, and an extra copy of Philadelphia chromosome; and group 2 with a relatively poor prognosis including i(17)(q10), -7/del7q, and 3q26.2 rearrangements. Patients in group 1 showed much better treatment response and survival than patients in group 2. When compared with cases with no ACAs, ACAs in group 2 conferred a worse survival irrelevant to the emergence phase and time. In contrast, ACAs in group 1 had no adverse impact on survival when they emerged from chronic phase or at the time of CML diagnosis. The concurrent presence of 2 or more ACAs conferred an inferior survival and can be categorized into the poor prognostic group.

Figure 1

The distribution of ACAs in CML and their impact on survival. Cases with ACAs were stratified based on the numbers of ACAs at the time of initial ACA emergence (A). The impact of ACAs on survival was analyzed. Date of ACAs emergence was used as the starting time point in panel B (OS after ACAs emergence), and date of CML diagnosis was used as the starting time point in panel C (OS after CML diagnosis). Of note, survival analysis in this figure included all stages of disease (CP, AP, and BP).

Figure 2

The stratification and survival impact of ACAs in CML. In cases with single ACAs, the types and frequencies of common ACAs were listed (A), and survival analysis following ACAs emergence (B) and CML diagnosis (C) was performed. In panel C, patients with no ACAs were included as a control. (D-E) Survival comparison between patients with 2 or more ACAs and patients with single ACAs. In patients with single ACAs, cases within each group were combined for survival analysis. Group 1 included patients with −Y, trisomy 8, and an extra Ph. Group 2 included patients with i(17), −7/Del(7q), and 3q26 rearrangements. Of note, survival analysis in this figure included all stages of disease (CP, AP, and BP).

Figure 3

Survival comparison between 2 groups with single ACAs based on ACA emerging phase. (A-B) Survival comparison in patients with ACAs emerging from CP. (C-D) Survival comparison in patients with ACAs emerging from AP and BP. For panels A and C, OS was calculated from the time of ACA emergence. For panels B and D, OS was calculated from the time of CML diagnosis.

Figure 4

Survival comparison between 2 groups with single ACAs based on ACA emerging time. (A) Survival comparison in patients with ACAs detected at CML diagnosis. (B-C) Survival comparison in patients with ACAs emerging during CML disease. For panels (A) and (C), OS was calculated from the time of CML diagnosis. For panel (B), OS was calculated from the time of ACA emergence.