Discussion
IKZF1 is composed of eight exons and encodes the transcription factor Ikaros [9]. IKZF1 alterations can occur frequently in germline cells and somatic cells, and it has been studied extensively. IKZF1 alterations in somatic cells are found in approximately 15% of pediatric patients with ALL[9] patients, but the alternations increase in prevalence among adult patients with ALL. The most common types of IKZF1 alterations are referred to as focal deletions, such as the whole gene and exons 4–7 deletions, resulting in loss-of-function in Ikaros, whereas IKZF1 mutations were less reported, including missense mutations, nonsense mutations, and frameshift mutations[2]. IKZF1 deletions are important molecular genetic makers in the development and relapse in ALL and attract significant enough attention, until Mullighan et al. first described the IKZF1 deletions in patients with BCR-ABL1 positive ALL[12] in 2008. Overall, the prevalence of IKZF1 deletions is 10% -30% in pediatric patients with B-ALL who are BCR-ABL1 negative [10], and can be as high as 56.8%-70.6% and 27%-68% in pediatric patient with B-ALL who are BCR-ABL1 positive or who have Ph-like ALL. These patients are often associated with high WBC counts, older age, poor chemotherapy responses, higher risks of relapse, and insensitivity to TKIs[13, 14]. In contrast, among ALL subtypes with good prognosis, the prevalence of IKZF1 deletions is low, accounting for 3%–6% and 15% of ETV6-RUNX1 and hyperdiploid cases[10, 15, 16]. Thus, for a long time, IKZF1 gene deletions have been established to be an independent adverse prognostic factor for patients with ALL [3, 17]. Therefore, screening for IKZF1 deletions is very important in recognizing high-risk patients and initiating risk-adapted chemotherapy for pediatric patients with ALL.
Most of the published data on IKZF1 deletions have been generated using MLPA analyses, SNP arrays, or multiplex PCR assays. Previous results have shown that RQ-PCR based on ALB as a control gene is a reliable and sensitive method for detecting IKZF1 deletions[5], and there is a good correlation between quantification of MRD and the level of IKZF1 deletions[5, 17, 18]. In this study, we conducted a quantitative analysis of IKZF1 deletions in pediatric patients with BCP-ALL to assess the impact of different levels of IKZF1 deletions on prognosis. The level of IKZF1 deletions in our patients ranged from 0.01 to 243.76%, with 60.8% of them being in the IKZF1 deletions of < 1% group. The rate of IKZF1 deletions in patients with ALL who were ETV6-RUNX1 positive or hyperdiploid were account for 10.8% and 14.9%, respectively, which is consistent with previous observations. Overall, 13.5% of the patients were BCR-ABL1 positive but patients with IKZF1 deletions of ≥ 1% had a significantly higher BCR-ABL1 positive rate (31.0%) than those with IKZF1 deletions of < 1% (2.2%). In contrast, all patients who were ETV6/RUNX1 positive had IKZF1 deletions of < 1%. Wu et al. [5]have reported that 90% of adult patients with IKZF1 deletions had levels of IKZF1 deletions of ≥ 1%, and the rate of IKZF1 deletions in patients who were BCR-ABL1 positive was 68.8%. These results differ from ours. One reason for this may be that patients who are in the high-risk group and BCR-ABL1 positive are more likely to had high levels of IKZF1 deletions; therefore, in our study fewer pediatric patients with ALL were in the high-risk group and were BCR-ABL1 positive than adult patients.
In our study, the percentages of patients whose age at onset was ≥10 years old, had WBC ≥ 50×109/L at initial diagnosis, and in high-risk group were significantly higher in the IKZF1 deletions of ≥1% group than in the IKZF1 deletions of <1% group. In contrast, there were no differences in other clinical and laboratory characteristics between the two groups. Various clinical trials have shown that IKZF1 deletions are associated with an older age at diagnosis, a higher WBC count, higher levels of MRD after induction, and a higher risk of relapse, which corresponded with our study[2, 13, 15, 20-22]. We found no difference in MRD at 15 and 33 days between the two groups of patients with different levels of IKZF1 deletions, whereas the rate of relapse/refractory in IKZF1 deletions of ≥1% group was significantly higher than in the IKZF1 deletions of <1% group, suggesting that there was more HR ALL in the IKZF1 deletions of ≥1% group.
Almost all clinical studies have demonstrated that IKZF1 deletions are an independent adverse prognostic factor in both pediatric patients with BCP-ALL who are either BCR-ABL1 positive or negative, especially in combination with an early MRD response. In children with BCR-ABL1 negative ALL who had been treated using the BFM-2000 protocol[15], the 5-year EFS was 69% for patients with IKZF1 alterations vs. 85% for patients without the IKZF1 alterations. Patients with IKZF1 alterations also had a higher incidence of relapse (21% vs. 10%). Similar poor outcomes in pediatric patients with IKZF1 deletions were observed in a Japanese pediatric Ph-negative B-ALL cohort[23], the AEIOP-BFM 2000 cohort[24], and the IC-BFM-2002 cohort[25]. Of note, the ETV6-RUNX1 subtype should be ruled out in patients with Ph-negative B-ALL who also have IKZF1 alterations because they have favorable outcomes despite the presence of IKZF1 alterations [26]. Therefore, the prognosis is still poor for BCR-ABL1–positive and Ph-like patients who have IKZF1 deletions, even if they are treated with TKIs. In AALL0622, a Children’s Oncology Group (COG) phase II trial, patients with ALL who were Ph+ and also had IKZF1 deletions were treated with dasatinib plus intensive chemotherapy and were found to have inferior 5-year EFS and OS compared to those patients with ALL who were Ph+ but had wild type IKZF1 (52% vs. 82%, P = 0.04 and 80% vs. 100%, P = 0.04, respectively) [27].Our results also showed that the 3-year OS and 3-year EFS in the IKZF1 deletions of ≥1% group were both significantly poorer than the 3-year OS (79.3% vs. 97.7%,P =0.022) and 3-year EFS (62.4% vs. 83.2%, P =0.019) in the IKZF1 deletions of <1% group. Similar results were found in BCR-ABL1 negative patients in corresponding different levels of IKZF1 deletions group. Therefore, the prognosis is worse in patients with higher percentage of IKZF1 deletions, regardless of BCR-ABL1 positive or negative. Because there was only one BCR-ABL1 positive patient in the IKZF1 deletions of <1% group, we could not compare the effect of different levels of IKZF1 deletions on the prognosis of BCR-ABL1 positive children. Multivariate analysis showed that the level of IKZF1 deletions of ≥ 1% and CNSL were independent risk factors of EFS. The level IKZF1 deletions of ≥1% group is likely to be a poor prognosis factor of OS, although no significant statistical difference was indicated due to a limited sample size. In the future, studies on the effect of different levels of IKZF1 deletions on prognosis with large cohorts of pediatric patients with BCP-ALL should be carried out. In addition, pediatric BCP-ALL patients with the level IKZF1 deletions of ≥1% may be stratified into high-risk groups due to poor prognosis and receive intensive chemotherapy.
Our study has several limitations. First, this was a single-center study with a small sample size. Second, due to the limitation of sample size, comparison between subgroups could not be performed.
In conclusion, pediatric BCP-ALL patients with IKZF1 Δ2-8/ALB deletions of <1% had a good prognosis, whereas patients with IKZF1 Δ2-8/ALB deletions of ≥1% had a poor outcomes, and these patients always had other risks such as older age, higher WBC count, with positive for the BCR/ABL1 fusion gene, and a higher relapse rate.
Conflicts of interest : The authors have no conflicts of interest to declare.