In this issue of Blood, Wang et al describe that 2 groups of additional chromosomal abnormalities (ACAs) beyond the Philadelphia chromosome (Ph) impact prognosis in chronic myelogenous leukemia (CML) patients treated with tyrosine kinase inhibitors (TKIs).1
At diagnosis, 10% to 12% of patients with chronic phase (CP) CML have chromosomal changes besides the Ph. Among these changes, variant translocations have usually not been considered detrimental to the prognosis of the patients. The other ACAs that have been observed in a minority of patients (5%) have been subdivided into major and minor routes.2,3 The major route ACAs, such as trisomy 8, a second Ph, isochromosome 17q or trisomy 19 have been associated with a negative impact on survival. Minor route ACAs have not been investigated to the same extent as major ACAs. Six minor route changes, including five numerical abnormalities (−7, −17, +17, +21, and −Y) and also one structural aberration t(3;21)(q26;q22), were initially described by Mitelman.3 The classification proposed by Mitelman was based only on the frequency of ACAs. A number of studies have been published on the relationship between ACAs and outcome, and it is clear that the situation remains unclear.
In a large group of patients in the German CML Study IV, 3.6% of patients had a major route ACA at diagnosis.4 After a median observation period of 5.3 years, for patients with minor and major routes, the 5-year progression-free survival (PFS) was 96% and 50%, and the 5-year overall survival (OS) was 96% and 53%, respectively. In this trial exploring the value of different doses of imatinib and the combination of imatinib with interferon or cytarabine, the times to complete cytogenetic response (CCyR) and to major molecular response (MMR) were longer in patients with major route ACAs. A similar study was conducted by the Italian GIMEMA Working Party on CML.5 Based on 559 patients enrolled into 3 different trials exploring 2 imatinib dosages (400 mg and 800 mg for different subgroups of patients according to Sokal score), ACAs at diagnosis were associated with worse outcome. In patients with major route ACAs, rates of CCyR and MMR were significantly inferior to those with only Ph translocation, and time to achieve CCyR and MMR were significantly longer. In contrast to the results of the German study, the Italian group did not observe differences between ACA patients and those without, when PFS and OS were considered. However, the current recommendation by the European Leukemia Net is to consider a major route ACA as a warning signal in patients treated frontline with TKI.6 The detection of ACAs has been considered a feature of the accelerated phase (AP). But the World Health Organization includes ACAs a feature of the AP only if they are not present at diagnosis (ie, as evidence of clonal evolution).7
In the current study, Wang et al revisited the relationship between ACAs detected in Ph+ cells of patients with CML and their outcome in the TKI era. The issue with the previous studies was that the heterogeneous group of minor route ACAs was not fully investigated. For example, the 3q26 rearrangement, a minor route ACA, is actually associated with TKI resistance and poor prognosis. More complex trisomy 8, a major route ACA, is considered a good prognostic ACA when occurring alone, whereas it is detrimental when associated with other chromosomal changes at initial diagnosis. Thus, the main objective, which is quite new, was to propose a revised classification of ACAs. In this paper, a large group of 608 CML cases treated with TKI had bone marrow (BM) biopsy with conventional G banding cytogenetic analysis at baseline and at 3- to 12-month intervals thereafter. Overall, patients with emergence of 2 or more ACAs simultaneously had a worse survival. When single, the 6 more frequently detected ACAs were trisomy 8, −Y, extra Ph, i(17)(q10), −7/del(7q), and 3q26.2 rearrangements. Two groups of patients were identified: the first group included trisomy 8, –Y, and an extra Ph with a relatively good response and better survival; the second group included i(17)(q10), −7/del(7q), and 3q26.2 rearrangements with a relatively poor response and worse survival. Of note, patients with –Y showed no significant survival difference from patients who had no ACAs. The timing of ACA emergence was also evaluated. In group 1, trends to higher CCyR and MMR rates were observed in patients with ACAs at diagnosis compared with patients with ACAs arising during the CML course. In addition, trisomy 8 and an extra Ph (that are both major route) had no significant impact on survival when they developed during the CP (no other concurrent AP features). In this study, two minor route changes, not well characterized in prior studies, 3q26 rearrangement and −7/del(7q) are clearly identified as ACAs associated with a poorer prognosis. The prognostic significance of 3q26 abnormality has been already detailed.8 Ph− ACAs were not considered in this paper. Unlike Ph+ ACAs, Ph− ACAs do not affect outcomes except in rare cases of myelodysplastic syndrome-associated abnormalities, which are typically accompanied by cytopenia.
Thus, physicians caring for CML patients should understand that, although molecular testing is important, there is still a place for careful cytogenetic analysis (ie, the type, frequency, timing of emergence of ACAs, and the phase of the disease) of their CML patients. In cases with insufficient molecular response, ABL1 mutations should be considered and performed in parallel with BM cytogenetic analysis. A relationship between ACAs and the initial clinical scoring system (Sokal, Euro, Eutos, and ELTS) would be of interest. It would be extremely helpful if all CML working groups could collaborate in order to increase the value of subsequent studies, allowing a deeper analysis of all ACAs individually and also exploring their potential relationship. No doubt, a study based on thousands of CML cases treated with TKIs would be of immense value.
Conflict-of-interest disclosure: The author declares no competing financial interests.