19p loss is significantly enriched in older age neuroblastoma patients and correlates with poor prognosis

Abstract

Genomic aberrations of neuroblastoma occurring in late childhood and adolescence are still understudied. Publicly available DNA copy number profiles of 556 tumors (discovery set) and of 208 tumors obtained by array-CGH assay (validation set) were used to test if 19p loss is significantly over-represented in children and adolescents with neuroblastoma. The 19p loss occurrence was separately tested within different age groups in the discovery and validation set and the resulting P values were combined by meta-analysis and corrected by Bonferroni's method. In both sets, 19p loss was associated with older age at diagnosis. Particularly, the lowest age group significantly associated with 19p loss (discovery set: 20%; validation set: 35%) was 6 years. The 19p loss correlated with inferior overall survival in patients over 6 years of age. Relevant tumor suppressor genes (KEAP1, DNM2, SMARCA4, SLC44A2 and CDKN2D) and microRNAs (miR-181c, miR-27a, and mirR-199a-1) are located in the genomic region involved in 19p loss. Downregulation of DNM2, SLC44A2 and CDKN2D was associated with poor patient outcome and older age. Among the recurrent NB chromosomal aberrations, only 1q gain was enriched in patients older than 6, and its presence was mutually exclusive with respect to 19p loss. Our data demonstrate that 19p loss is a genomic biomarker of NB diagnosed in older children that can predict clinical outcome.

Keywords: Cancer genomics; Prognostic markers.

Fig. 1. Prevalence of 19p loss in different age groups.

The bar plot shows the prevalence of a 19p loss at different ages at diagnosis and of b diverse SCA at 6 years age group in two independent datasets. The statistically significant difference of 19p loss frequency is also shown.

Fig. 2. Deleted regions of 19p in patients older than 6 years.

Genomic view of chromosome 19p (hg19 genome assembly). Gray tracks report the deleted region found in patients older than 6 years. The minimally deleted region (track 4284), is indicated by the red box.

Fig. 3. Survival analysis.

a OS and EFS using Kaplan–Meier analysis show difference in the survival between NBs with and without 19p loss. b Forest plots represent the multivariate models of OS and EFS including potential prognostic factors as MYCN status (MYCN amplified versus MYCN non-amplified), stage (4 versus 1, 2, 3, 4S), and 19p loss (normal versus affected). The hazard ratio and the associated p values were calculated using Cox proportional hazard analysis.