Revisiting late relapses in acute myeloid leukemia

While patients with acute myeloid leukemia (AML) who are in remission for more than 5 years have a very low chance of disease recurrence, it does occur, and so it is useful to ask what the most appropriate approach to the disease should be. In this month's issue of Leukemia and Lymphoma, Verma et al. [1] from the M. D. Anderson Cancer Center (MDACC) report on their experience with such patients who had initial treatment between 1980 and 2003. Out of 2347 patients who received induction style therapy, 1366 achieved remission. Relapse occurred in 942 cases, with 11 of those being after 5 years. The recurrent disease was biologically different from the presentation in two ways: first, the cytogenetics of the relapsed disease was different from presentation, and second, the remission rate was more akin to that seen in secondary leukemia. Despite the low complete response (CR) rate, remissions of >30 months were obtained.

The current report, although quite different from another publication on topic by Medeiros et al. [2] earlier in this same journal, helps to define the clinical landscape of patients with this rare entity. Both series define late relapse as AML recurring at least 5 years after attainment of first remission. An earlier series by Kantarjian et al. [3] used 18 months as the definition of late relapse. Together, these series indicate the risk of AML relapse after 18 months as 17% and after 5 years as between 1.16 and 3% [2,3]. Several important differences, however, exist between the MDACC and Princess Margaret Hospital (PMH), Toronto, series. Verma et al. found the CR2 rate with chemotherapy to be 36%, in contrast to 86% as reported by Medeiros et al. These reports demonstrate that further intensive chemotherapy can result in CR2 durations of at least 12 months in 36–47% of patients with late AML relapse. Interestingly, only 2/11 patients with a prolonged CR2 duration in both series combined underwent allogeneic stem cell transplant, indicating that meaningful clinical outcomes may be achieved with salvage chemotherapy alone. Another striking difference was that Verma et al. identified new chromosomal abnormalities (2q+, −7, trisomies, 5q−, or −8) in 5/8 patients with paired diagnostic/relapse cytogenetic samples, compared to no confirmed cases of cytogenetic progression in the Medeiros series. Possible explanations for these differences include the older median age of the MDACC patients (66 years; range 37–79) as well as the extensive exposure to post-remission maintenance chemotherapy, with only one patient receiving two consolidations and the remaining patients receiving a mean of 19 post-therapy treatments. In contrast, the median age of the PMH series was 48 years (range 13–77), with only two consolidation courses delivered. The poorer response to therapy in the MDACC series is in keeping with the idea of the late relapses in fact being secondary leukemias.

While these articles support diametrically opposed views on the biology and response of relapsed AML, they in fact confirm our current understanding of this complex and heterogeneous collection of diseases. That is, recurrence of the same disease after one or many years of remission behaves similarly to the original leukemia, whereas secondary leukemias, whether they arise de novo or as a result of treatment for breast cancer, lymphoma, or, as in this case, AML, have a similar and almost invariably poor response to therapy.

Based on the above, the optimal therapy should follow current approaches for primary or secondary leukemias. In order to properly inform the patient, and choose the currently most appropriate treatment, the clinician should await the result of classic cytogenetics. Regardless of whether the relapse is of the original disease or is a secondary form of AML, allogeneic stem cell transplant should be offered to biologically appropriate patients, as in the former the propensity of the disease to recur has been demonstrated, and in the latter the almost uniform chance of relapse with current induction therapies is well known [4].

Taken together these two publications demonstrate that there is value to therapy of late relapsed AML. However, the studies from the MDACC group suggest that the disease occurring at a late time point may not be the same as the presentation disease. With modern molecular techniques, including array comparative genomic hybridization (aCGH), sequence analysis, and changes in methylation, it should be possible in the future to determine whether one is dealing with the initial disease or a de novo secondary leukemia. This knowledge will help inform treatment choices for the patient.