Outcome and prognostic factors of CBF pediatric AML patients with t(8;21) differ from patients with inv(16)

In this study, it was found that patients with inv (16) were far common in White race than those with t(8; 21), which indicated that the prevalence of CBF-AML subtype were associated with race. Moreover, the FAB type of t(8; 21) was nearly M2 and mostly of inv (16) were M4, which was consistent with previous reports [5, 6]. Subsequently, much higher initial WBC were observed in childhood patients with inv(16) compared with those with t(8; 21), we suggested that inv(16) was associated with leukocytosis and extramedullary infiltrative manifestations [1]. Interestingly, the percentage of chloroma was lower in patients with inv(16) than those with t(8; 21), and this was never been reported before. In accordance with previous studies, when referred to relapse, BM relapse and CNS relapse were more frequent in patients with inv(16) than those with t(8; 21) [7,8,9].

Using the large TARGET database of pediatric cases of CBF-AML, we characterized the secondary cytogenetic abnormalities in patients with CBF-AML, defined by either inv(16) or t(8;21). Most previous studies showed that del(9q) and loss of a sex chromosome were more frequent in patients with t(8;21) [10,11,12]. Similar results were also observed in our study, del(9q), minus X and minus Y had a much higher percentage in t(8;21). Von Neuhoff et al. [4] showed that the 5-year EFS of children with t(8;21) combined with loss of a sex chromosome was significantly higher than that of children with t(8;21) (100% vs 71%, P = 0041). Other findings also supported the association of loss of a sex chromosome with a better prognosis in children with t(8;21) [3, 29, 30]. However, Duployez et al. [13] showed that loss of a sex chromosome had no impact on the prognosis of CBF-AML patients in a large mixed cohort study of 73 children and 125 adult patients. In contrast to a previous report by Duployez et al., we found loss of a sex chromosome was not a prognostic factor in our cohort. Klein et al. [2] showed that CBF-AML patients in the del(9q) group (n = 104) had a lower CR rate than those in the non-del(9q) group (n = 734) (P = 0.01), while another study [4] showed a good prognosis for children with the t(8;21) AML subgroup with del(9q). In the present study, patients with del(9q) did not show any significant differences in terms of survival and remission rates. This might be explained by the inconsistent sample size of these studies.

Children with CBF-AML respond well to chemotherapy, and other research centers have reported CR rates of up to 90% after chemotherapy [14,15,16]. In our cohort, the CR rates were 95.2% in the CBF-AML patients, and 94% and 95.2% in inv(16) and t(8;21), respectively [17, 18]. Although the previously reported mediocre OS and high relapse rate were confirmed in this cohort, the 10-year OS of 84.4% among the CBF-AML patients were relatively good, especially given the number of patients who were diagnosed many years ago. In 2015, AML–Berlin-Frankfurt-Münster (BFM)-98 Study showed the favorable outcome in the subgroups of patients with inv(16) and t(8;21), with an 5-years OS of 87 and 91%, and the 5-years CIR were 16% and 12%, respectively [4]. In the current study, we also demonstrated an excellent 10-years OS of 89.5 and 80.7% but with a relative higher 10-years CIR of 35.2% and 25.4% in the patients with inv(16) and t(8;21), respectively. Compared to our study, the CIR were much lower in the BFM-98 Study, we suggested that the reason was that sample of CBF-AML patients in the BFM-98 Study were smaller (only 99 cases) and the median follow-up were shorter than ours. In spite of these difference, we found patients with inv(16) had a significant higher survival rate and relapse rate than those with t(8;21) in our study. In terms of clinical characters and prognosis, we might concluded that the patients with inv(16) and those with t(8;21) were the two clinically distinct entities.

The c-kit mutation were the most common in CBF-AML children [19, 20]. Recent reports have demonstrated that the prevalence of c-kit mutation in children with CBF-AML was 10-54.5% [8, 19,20,21,22,23]. Chen et al. [21] showed that the incidence of c-kit mutation in children with t(8;21) ranged from 17 to 42%, and 21 to 55% in those with inv(16). In the present study, c-kit mutations were 25.4% among CBF-AML children, and the percentage of c-kit mutation in children with t(8;21) were 24.4%, and 26.9% in those with inv(16). The results of this study were generally consistent with previous findings. The c-kit mutations were widely reported in adults with CBF-AML, and most investigators believed that the mutations suggested a poor prognosis. Tokumasu et al. [23] showed that 46 pediatric patients with t(8;21) accompanying c-kit mutations had a significantly lower EFS than 61 cases without mutations (n = 61). Our multivariate analysis also addressed that c-kit mutations were the independent adverse factor that influenced CIR and OS. However, the studies on c-kit mutations in children with CBF-AML were still rare, and the relationship between mutations and prognosis remained controversial. An international, multicenter survey of 97 patients of CBF-AML showed that CBF-AML patients with FLT3-ITD had much lower 4-year relapse-free survival rate compared to the patients without FLT3-ITD (38% vs 80%, P = 0.02) [24]. In our study, FLT3-ITD positive demonstrate a poor outcome in terms of OS an CIR, and this was further confirmed in our multivariate analysis. Based on the second strike doctrine, we implied that c-kit mutation and FLT3-ITD mutations play an important role in the pathogenesis of CBF-AML.

The erythromycin plus cytarabine induction chemotherapy regimen and the high-dose eytarabine based consolidation chemotherapy regimen are the clinical standard first-line chemotherapy regimens for CBF-AML [25, 26]. A Cancer and Leukemia Group B Study showed that patients in the two consolidation groups, multicourse HDAC (n = 149) and single-course HDAC (n = 48), had significant difference on 10-years CIR (41% vs 64%, P = 0.009) [27]. CALGB 8461 study demonstrated that the CIR was significantly decreased in patients assigned to receive three to four cycles of HDAC(n = 28) compared with patients assigned to one course (n = 20) (5-year CIR, 43% v 70%, P = 0.03) [28]. In contrast to the two reported results, our study demonstrated that no significant difference were found between four (21.6 g/m2 cytarabine) and five (45.6 g/m2 cytarabine) chemotherapy courses in terms of CIR and OS. Most interestingly, subgroup analysis showed that CIR of patients with t(8;21) can be decrease by five chemotherapy courses, and this suggested that maybe only patients with t(8;21) could benefit from more cumulative cytarabine exposure.

The impact of GO treatment on outcome was subsequently evaluated in our study. A meta-analysis that included five randomized controlled trials showed that GO treatment improved the risk of relapse and 5-year OS in CBF-AML patients, with a definite survival advantage for CBF-AML patients with GO treatment compared to those without GO treatment(OR = 0.47, 95%CI:0.31–0.73, P < 0.001) [29,30,31,32,33]. Although our data did not show significant difference between GO and No-GO treatment, the subgroup analysis showed that the patients with inv(16) who did not receive GO had significant higher CIR and similar OS when compared to those with t(8;21). In contrast, inv(16) and t(8;21) receiving GO treatment had comparable outcomes as well as OS and CIR, and this suggested the GO added to conventional chemotherapy improved outcomes for only inv(16). In the current study, we confirmed that the outcome of patients with c-kit mutations could be improved by GO treatment. Thus, we implied that due to c-kit mutations were more common in inv(16), so the patients with inv(16) were also improved by GO treatment.

In summary, we concluded that patients with inv(16) and t(8;21) pediatric AML constitute two separate entities clinically, in that they differ with regard to clinical characteristics, prognosis and treatments. Notably, we showed the impact of GO treatment on patients with inv(16)) and cumulative cytarabine exposure on patients with t(8;21). Furthermore, due to our data, based on a prolonged follow-up, show that the rates of relapse are still disappointing for both patients with inv(16) AML and those with t(8;21) AML, it is important that future studies identify and target therapeutically the leukemogenic mechanisms accountable for molecular and clinical differences between the two cytogenetic groups of CBF AML.

留言 (0)

沒有登入
gif