Outcomes in acute lymphoblastic leukemia (ALL), the most common childhood malignancy, have improved significantly due to risk-based multi-agent chemotherapy strategies derived from diagnostic advances in cytogenetics and molecular biology. Characterization of recurrent genetic lesions has enabled further prognostic stratification on recent clinical trials. Abnormalities involving the short arm of chromosome 9 are well-documented in pediatric ALL. Three critical tumor suppressors: CDKN2a (p16INK4a), CDKN2b (p15INK4b) and p14ARF have been localized to band 9p21. While 9p21 abnormalities have been reported in 10% of childhood ALL, their prognostic significance remains uncertain and may depend on which of these genes are lost/inactivated. Moreover, 9p deletions may be cryptic and/or not readily identified by standard G-banding techniques. We investigated the incidence and prognostic significance of 9p21 abnormalities identified retrospectively in pediatric ALL patients.
A total of 76 children were diagnosed with ALL at the IWK Health Centre in Halifax, Nova Scotia from 2000–2005. Cell pellets were available for analysis on 48 patients. Chart review was conducted retrospectively. Approval was obtained by the IWK Research Ethics Board. 9p21 abnormalities were assessed by two modalities: 1) fluorescence in-situ hybridization (FISH) using a commercial p16 probe; and 2) methylation specific multiplex ligation-dependent probe amplification (MS-MLPA), using SALSA MLPA kit ME024-A1 9p21 CDKN2A/2B region (MRC Holland).
Of the 48 samples studied by FISH, 18 (38%) had a 9p21 deletion, only 4 of which were previously identified by G-banding. MLPA was performed on 40 samples and identified a total of 18 (45%) patients with a 9p21 deletion, 5 of which were not identified by FISH. Aberrant methylation at CDKN2b (p15INK4b) was found by MLPA in 19 patients (46%). Overall, 32 of 48 (67%) patients were determined to have an abnormal 9p21, representing a substantially greater frequency than previously reported. The detection rate for deletions at 9p21 between FISH and MLPA, was not statistically different (p=0.31). Interestingly, all patients with T-cell disease (n=5) had an abnormal 9p21, suggesting an association between this immunophenotype and 9p21 abnormalities (p=0.02). Overall, 9p21 abnormalities were associated with National Cancer Institute (NCI) high risk criteria (p=0.04), specifically white blood cell count >50,000/μL at diagnosis (p=0.019) and a higher percentage of leukemic blasts in the peripheral blood (p=0.016). Only four B-cell lineage patients in our cohort relapsed, three of which had 9p21 abnormalities and one of these ultimately died of disease.
Using FISH and MLPA, our study has revealed a much higher incidence of 9p21 abnormalities in childhood ALL than has previously been described. Moreover, our findings demonstrate an association with high risk features and possibly inferior prognosis. It is possible that the lack of consistent reports linking 9p21 abnormalities to prognosis are related to suboptimal ascertainment. A small sample size in our study prevents the conclusive determination of an independent prognostic impact. However, our results justify the need for a larger multi-center investigation to evaluate how 9p21 abnormalities might affect ALL outcome and possibly influence future risk stratification for determining treatment.
No relevant conflicts of interest to declare.