Childhood Acute B-Lineage Lymphoblastic Leukemia With CDKN2A/B Deletion Is a Distinct Entity With Adverse Genetic Features and Poor Clinical Outcomes

To further emphasize the clinical–genetic features and prognosis of CDKN2A/B deletions in childhood acute lymphoblastic leukemia (ALL), we retrospectively analyzed 819 consecutive B-ALL patients treated with the Chinese Children’s Cancer Group ALL-2015 (CCCG-ALL-2015) protocol, and fluorescence in situ hybridization (FISH) analysis on CDKN2A/B deletion was available for 599 patients. The prevalence of CDKN2A/B gene deletions was 20.2% (121/599) of B-ALL. CDKN2A/B deletions were significantly associated with older age, higher leukocyte counts, a higher percentage of hepatosplenomegaly, and a higher frequency of BCR-ABL (p < 0.05). Those patients achieved similar minimal residual disease (MRD) clearance and complete remission compared to patients without CDKN2A/B deletion. The CDKN2A/B deletions were correlated with inferior outcomes, including a 3-year event-free survival (EFS) rate (69.8 ± 4.6 vs. 89.2 ± 1.6%, p = 0.000) and a 3-year overall survival (OS) rate (89.4% ± 2.9% vs. 94.7% ± 1.1%, p = 0.037). In multivariable analysis, CDKN2A/B deletion was still an independent prognostic factor for EFS in total cohorts (p < 0.05). We also detected a multiplicative interaction between CDKN2A/B deletions and TP53 deletion on dismal prognosis (p-interaction < 0.05). In conclusion, CDKN2A/B deletion is associated with distinct characteristics and serves as a poor prognostic factor in pediatric ALL, especially in TP53 deletion carriers.


INTRODUCTION
Pediatric acute lymphoblastic leukemia (ALL) is one of the most curable malignancies, with 5-year event-free survival rates exceeding 80% in many developed countries and even exceeding 90% in high-income countries (1). With the increased understanding of genetic alterations in ALL, many molecular markers have been identified and applied to risk stratification and treatment protocols in leukemia. CDKN2A/B is one of the most frequent abnormal genes in ALL and can be detected by fluorescence in situ hybridization (FISH), multiplex ligation-dependent probe amplification (MLPA), array-based comparative genomic hybridization (aCGH) analysis, and single-nucleotide polymorphism array (SNPA) (2)(3)(4)(5)(6). As a secondary genetic event in the development of leukemia, the CDKN2A/B deletions were found in approximately 20%-25% of B-cell precursor acute lymphoblastic leukemia (BCP-ALL) and 38.5%-50% of T-ALL patients (2,5,7,8). Despite the high frequency of CDKN2A/B deletions in pediatric ALL, the prognostic importance of the deletions is still inconclusive. Most investigators have concluded that the deletions were associated with the recurrence of pediatric ALL (7,9,10), and some researchers found that the inactivation of CDKN2A/B did not influence the outcome of childhood B-lineage ALL (4,11). As previous inconsistent conclusions were drawn from small cohorts, large sample-sized studies are needed to clarify the prognostic impact of CDKN2A/B deletion in pediatric ALL. In this study, we assessed the clinical and biological characteristics and prognostic factors of CDKN2A/B deletions in 662 pediatric ALL patients.

Patients and Treatment Protocols
The cohort included 902 patients with newly diagnosed pediatric ALL who were treated according to the CCCG-ALL 2015 protocol (ClinicalTrials.gov identifier: ChiCTRIPR-14005706) (12) registered at the Blood Disease Hospital of CAMS & PUMC between May 2015 and December 2019. As the FISH test has been performed since September 2016, FISH data from 662 patients were collected in this cohort finally ( Figure 1). The protocol described in this study was approved by the Ethics Committee, Institute of Hematology and Blood Disease Hospital, Diseases Hospital, Chinese Academy of Medical Sciences (CAMS) and Peking Union Medical College (PUMC) (No. IIT2015010-EC-1). All patients or their legal guardians signed written informed consent before treatment.

Risk Group Assignment
Patients enrolled in the CCCG-ALL-2015 study were assigned to d i ff e r e n t r i s k g r o u p s b a s e d o n m o r p h o l o g y a n d immunophenotypic and genetic features of leukemia cells. Patients were assigned to the low-risk group if they had B-cell ALL and were aged between 1 year and 10 years; had a leukocyte count of less than 50 × 10 9 /L, a chromosome number of more than 50, or the ETV6-RUNX1 fusion gene; and did not have CNS 3 status or testicular leukemia, and had a minimal residual disease of less than 1% on day 19 of induction and less than 0.01% on day 46 of induction. Patients with a minimal residual disease of 1% or more (or ≥5% blasts morphologically without suitable markers for minimal residual disease) in bone marrow on day 46 of induction and infants younger than 6 months with KMT2A rearrangement and a leukocyte count of 300 × 10 9 /L or more were classified as high-risk ALL. The remaining participants were assigned to the intermediate-risk group (12). All eligible patients received minimal residual disease-directed, risk-stratified treatment modified from the St Jude Children's Research Hospital Total Therapy 15 and 16 studies (13,14) and the Shanghai Children's Medical Center ALL-2005 trial (15), with IR/HR patients receiving more intensive treatment.

FISH and Probe
Pretreatment bone marrow aspirates were taken at diagnosis and at least 1-2 ml of bone marrow aspirates was analyzed by FISH for cytogenetic abnormalities (including CDKN2A/B, KMT2A rearrangement, TP53 deletion, and BCR-ABL1), polymerase chain reaction (PCR) for fusion gene (including ETV6-RUNX1, BCR-ABL, and TCF3-PBX1) (16), and karyotyping. We analyzed interphase cells according to the instructions of the probe manufacturer (America Abbott). The CDKN2A/B probe spanned approximately 222 kilobases (kb) and contained many genes, including methylthioadenosine phosphorylase, CDKN2A (INK4A and ARF), and CDKN2B (INK4B) in the 9p21 chromosome region. The cutoff level for positive results was calculated to be 5%, and at least five hundred cells were analyzed (5,17). Some cases with two different deleted populations (one biallelic and one monoallelic) were classified as having a biallelic deletion.
Diagnosis of complete remission (CR) was based solely on bone marrow morphology with a cutoff value of 5% of leukemic blasts on treatment day 46. Bone marrow MRD that is not lower than 10 -5 was defined as positive. A poor response to prednisone was considered if the peripheral blast count on treatment day 5 was not lower than 1×10 9 /L. The primary endpoint event-free survival was calculated from the date of diagnosis until the following events: induction treatment failure, any relapse after CR, second malignancy, and death due to any cause. Overall survival was considered as the time from diagnosis to death due to any cause.

Statistical Analysis
Chi-squared test and Mann-Whitney U tests were used to compare categorical and continuous variables, respectively. Event-free and overall survival were calculated and compared using Kaplan-Meier analysis and log-rank tests. Cox proportional hazard regression analyses were performed using univariable and multivariable regression approaches. Variables that were found statistically significant and a p-value of approximately 0.1 were included in multivariable Cox regression analysis. We performed subgroup analyses for prespecified baseline factors with rates of inferior events by factor interaction with the use of Cox models. Statistical significance was defined as a p-value of less than 0.05. All analyses were performed using SPSS v. 24. and SAS v. 9.4 software.

Comparison of Demographic and Clinical Characteristics of Childhood ALL With and Without CDKN2A/B Deletions
In total, 599 patients were included in the analysis. The final follow-up was in December 2020 and the median follow-up time was 34 months (range: 0 to 58 months). The median age was 5 years (range, 0 to 14), and the male/female ratio was 1.4 (350/ 249). The prevalence of CDKN2A/B gene deletions in B-ALL was 20.2% (121/599). Among patients with CDKN2A/B deletions, 57 (45.3%) harbored CDKN2A/B biallelic deletions, whereas 82 (54.7%) harbored monoallelic deletions, and 11 (7.3%) patients harbored both biallelic and monoallelic deletions.
Compared to patients with wild-type CDKN2A/B (Table 1), patients with CDKN2A/B deletions were significantly associated with older age (age >10 years; 30.6% vs. 15.2%, p < 0.001), a higher leukocyte count (median: 24.7 vs. 8.9×10 9 /L, p < 0.001), a lower platelet count (median: 51 vs. 64 g/L, p = 0.049), and a higher percentage of hepatosplenomegaly (64.5% vs. 43.9%, p < 0.001). A higher rate of central nervous system status 2(CNS 2)/ traumatic lumbar puncture was found among patients with CDKN2A/B deletion, but no statistical significance was found in the two groups. Patients with CDKN2A/B deletions belonged more to the intermediate-risk groups (58.7% vs. 32.6%, p < 0.01) than patients without CDKN2A/B deletions based on CCCG-ALL 2015 risk stratification. There was no significant difference between patients with biallelic deletions and monoallelic deletions. The comparison of characteristics for the two groups is shown in Table S1.

Correlation of CDKN2A/B Deletion With Other Cytogenetic Alterations
Karyotype analysis, 43 fusion genes, and FISH studies of pretreatment bone marrow were available for 599 patients.  (Figures 2A, B).

DISCUSSION
In this study, we present a large retrospective study of pediatric ALL patients with CDKN2A/B deletions treated in a single center  and demonstrated the adverse effect of CDKN2A/B deletions on clinical outcomes. The prevalence of CDKN2A/B deletions in our study was 20.2%, which was higher than Agarwal's research (19.8%) but lower than Sulong's research (22.0%) (2,5). Previous investigators have reported the characteristics and clinical impact of CDKN2A/ B deletions in pediatric ALL, and the prognostic importance of CDKN2A/B deletions in pediatric ALL is still controversial (2,4,5,10,19). We conducted a comprehensive analysis of CDKN2A/B deletions in 599 pediatric B-ALL patients. Data from our cohort showed that CDKN2A/B deletions were associated with older age at diagnosis, higher white blood cell counts, and prominent hepatosplenomegaly, which were consistent with the previous studies (5).
Sulong et al. reported that pediatric ALL with CDKN2A/B deletions had recurrent cytogenetic abnormalities including high frequencies of TCF/PBX1, BCR/ABL, hyperdiploidy, and KMT2A, and low frequencies of ETV6/RUNX1 compared to patients without CDKN2A/B deletions (5,11). We conclude a similar result in which CDKN2A/B deletion carriers had a higher prevalence of BCR/ABL and a low prevalence of ETV6/RUNX1. However, we did not find ph+ patients had inferior outcomes than ph-patients in CDKN2A/B deletion groups; this is most likely due to the utilization of tyrosine kinase inhibitors (TKIs) in ph+ patients for the entire duration of ALL therapy (22,23). Pfeifer et al. demonstrated that CDKN2A/B deletions were adverse despite allogeneic stem cell transplantation in adult Philadelphia chromosome-positive (Ph+) ALL (24,25); further studies in a larger cohort of pediatric Ph+ ALL with CDKN2A/B deletions are needed.
In our cohort, patients with CDKN2A/B deletions were more frequently steroid-resistant, whereas they had a better MRD clearance on day 19 of induction therapy compared with patients without CDKN2A/B deletions. These results differed from a study by Braun who found higher MRD levels on day 15 of induction therapy in patients with CDKN2A/B deletions (26). The differences in MRD clearance may be related to the patients' race, chemotherapy protocols, and examination methods. An earlier study demonstrated that CDKN2A/B deletions were associated with unfavorable outcomes independent of MRD level in adult patients (27). However, both MRD positive on day 46 of induction therapy and CDKN2A/B deletion were still independent poor indicators for childhood ALL patients in our study.
Our study also indicated that patients with CDKN2A/B deletion had the worst prognosis in the TP53 deletion subgroup. A similar outcome was found by Delfau-Larue for mantle cell lymphoma; i.e., patients with both CDKN2A/B and TP53 deletions had the worst prognosis (28). The CDKN2A/B gene controls the cell cycle through the P53-MDM pathway; thus, co-occurrence of CDKN2A/B deletions and TP53 deletions might enhance the aggressiveness of disease by strongly increasing the self-renewal capacity of leukemia cells (29)(30)(31). The hypothesis needs to be proved in future research.
Several limitations exist in this study. First, because of the limited sample size and short follow-up duration in relapse patients with CDKN2A/B deletions, the role of allogeneic transplant in the treatment needs to be interpreted carefully.
Second, for the same reason, we did not conclude whether there had been significant interactions between two TP53 deletions and CDKN2A/B deletions in pediatric ALL. CDKN2A gene can control cell cycle through the P53-MDM pathway; it might be attributed to the fact that CDKN2A deletion cooperates with the TP53 deletion to enhance the aggressiveness of the disease by strongly increasing the self-renewal capacity of leukemia cells, but there is no theory to support this hypothesis.
In conclusion, a significant proportion of pediatric ALL patients still experience relapse, particularly patients with both CDKN2A/B and TP53 deletions despite the high survival rate of childhood ALL. MDM2-P53 targeted agents are still in experimental research. In addition, CDK4/6 inhibitors (e.g., palbociclib) combined with chemotherapy are in clinical trials for the management of pediatric patients with relapsed/refractory ALL (32,33). In the future, a new therapeutic strategy (e.g., target drug) based on genetic events might be applied in some subtypes of pediatric ALL.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The protocols described in this study were approved by the Ethics Committee, Institute of Hematology and Blood Disease Hospital, CAMS & PUMC. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

AUTHOR CONTRIBUTIONS
All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.

FUNDING
The work is supported by the National Nature Science Foundation of China grants 81770175 (YCZ), 81870131 (XZ), and 81670112 (XZ).

ACKNOWLEDGMENTS
We would like to thank the patients for participating in this study.