Prognostic Implication of Therapy-Related Acute Myeloid Leukemia after Comparative Analysis with De Novo acute Myeloid Leukemia

Abstract

Background: Therapy-related acute myeloid leukemia (t-AML) is regarded as a complication after cytotoxic chemotherapy and/or radiation therapy, and also considered to have a poor survival outcome compared to de novo AML. We still have a question whether t-AML itself indicates a poor prognosis or whether the inferior outcome results from the association with such an adverse characteristics including cytogenetic risk or age or underlying malignancies.Methods: In this single center retrospective study, 1825 patients (median 46 years old [range, 17-92]) with variable karyotypes were enrolled from 2002 to 2013. Fifty-four (3.0%) patients had previous malignancies or autoimmune diseases, and all of them were treated with radiation or toxic chemotherapy before diagnosis of t-AML with a median duration of 36.3 months (range, 2.9-280.5). We analyzed clinical outcomes compared to 1771 de novo AML patients who were not related with any toxic therapies before.Results: Among 54 t-AML patients, 42 (77.8%) was in remission of prior malignant disease and 8 were in stable disease and 4 were in relapsed disease. In t-AML subgroup, median age was older (50 vs. 46 years old, p =0.119), leukocyte and bone marrow blast counts were significantly lower than de novo AML subgroup. There were more female patients in t-AML subgroup (70.3% vs. 45.4%, P =0.003). Among 38 female t-AML patients, 13 (34.2%) patients had breast cancer, 10 patients had hematological malignancies (i.e. APL in 5, lymphoma in 3, multiple myeloma in 2), and 8 (21.1%) had gynecological malignancies (i.e. ovarian and cervical cancer etc.). One or more chromosomal abnormalities (82.6% vs. 68.3%, P =0.015) and more adverse-risk karyotypes (41.2% vs. 20.0%, p<.001) were in t-AML subgroup. Especially, t-AML had more 5 or 7 chromosomal abnormalities (7.8% vs. 2.0%, P =.004) and complex karyotypes (27.5% vs. 7.6%, p<.001) which also included abnormal 5 or 7 chromosomes. Smaller number of t-AML patients received induction chemotherapy (74.1% vs. 87.6%, P =0.006) and early death rate was higher in t-AML group (22.2% vs. 13.7%, P =.083). After median follow-up of 70 months (range: 5.6-165.0), t-AML showed inferior 5-year overall survival (OS) compared to de novo AML (23.8% vs. 39.0%, p <.001). The result was more significant in intermediate to poor-risk group (9.2% vs. 30.0%, p<.001), but it was similar in favorable-risk group (75.0% VS. 62.8%, P =.532). In treated cohort, however, remission rate (70.0% vs. 79.3%, p =.149) and relapse rate (28.8% vs. 35.9%, p =.544) was not different, and multivariate analysis showed t-AML did not affect OS (HR=1.25, P =.185), while age >50 years old (HR=1.48, p<.001), hematopoietic cell transplantation (HCT, HR=0.37, p<.001), favorable-risk karyotype (HR=0.48, p<.001), and post-induction remission status (HR=0.26, p<.001) did. Five-year OS of t-AML patients treated with HCT (n=16) was 50.0%, and for intermediate to poor-risk subgroup treated with HCT, 5-year OS was 33.3%.Conclusion: In this study, t-AML was related with a larger proportion of adverse-risk karyotype, and many patients could not start induction chemotherapy due to old age, and remained prior malignant disease, which might result in poor survival outcome. On the other hand, response to induction chemotherapy of t-AML was similar with de novo AML consistent with a recent report (Kayser et al. Blood 2011). Therefore, if previous malignancy is in remission or in stable disease, aggressive treatment strategy using HCT may overcome poor survival outcome of t-AML. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.