Predictive Model May Determine Risk of Subsequent CNS Tumour in Survivors of Childhood Cancer

Genetic variants in genes associated with CNS tumours form the basis of the model

Although survivors of childhood cancer who were exposed to cranial radiation therapy at a young age are known to be at risk of developing central nervous system (CNS) tumours as adults1, a recent study appearing in the Journal of Clinical Oncology suggests that specific genetic components may further increase this risk for certain individuals.

Noting that there is significant interindividual variability in the risk of developing subsequent CNS tumours, first author Xuexia Wang, University of North Texas in Denton, USA and colleagues hypothesized that genetic variants in candidate genes could be used to create a prediction model for the determination of high or low risk of developing subsequent CNS tumours by survivors of childhood cancer.

Most of the candidate genes used in this model mediate pathways involved in DNA damage response and repair, telomere homeostasis, or drug metabolism 

The investigators curated candidate genetic variants observed in adult onset primary CNS tumours from previously published studies occurring in the Medline and EMBASE databases from January 2005 through April 2014, which uncovered 42 publications containing 129 representative single nucleotide polymorphisms (SNPs) located on 46 genes. Of these, 34 SNPs with minor allele frequency were excluded, leaving 95 SNPs for the replication analysis, which was carried out in 82 survivors of childhood cancer with and without subsequent CNS tumorus and 228 matched controls. 

Genomic DNA was isolated from peripheral blood or saliva obtained from the participants and genotyping was performed on the Illumina semicustom HumanExome+v1.1 array that was specifically designed for the Childhood Oncology Group (COG) ALTE03N1 study.

Successful validation of the model done with cases and matched controls

Prediction models were generated to identify survivors at high or low risk for subsequent CNS tumours, which were validated in an independent matched case-control study of 25 survivors and 54 controls. The final model included genetic variants, and also contained age at primary cancer, sex, and cranial radiation therapy dose.

The model was successfully validated, demonstrating sensitivity and specificity of predicting survivors of childhood cancer at highest or lowest risk of subsequent CNS tumours of 87.5% and 83.5%, respectively. The positive predictive value (PPV) of the model was 60.9% and the negative predictive value (NPV) was 95.8% for all subsequent CNS tumours. The model’s validation for subsequent glioma, showed sensitivity of 84.6%, specificity 90.6%, PPV of 68.8%, and NPV 90.6%. For subsequent meningioma, the prediction model had sensitivity of 82.8%, specificity 74.4%, PPV was 68.6%, and the NPV was 86.5%.

A link between specific SNPs and all subsequent CNS tumours was determined, which implicated several SNPs including rs15869 (BRCA2), rs1805389 (LIG4), rs25489 (XRCC1), rs1673041 (POLD1), rs8079544 (TP53), and rs11615 (ERCC1); the associated odds ratios (ORs) of the SNPs to subsequent CNS tumours ranged from 1.6 to 4.0.

Exploratory analyses were also conducted for glioma and meningioma

The SNP rs15869 (BRCA2) was found to associate with an increased risk of both glioma and meningioma, which the authors suggest may be due to exposure to DNA damaging agents delivered by radiation therapy in the presence of impaired DNA repair mechanism.

The risk of subsequent glioma associated with 3 SNPs: rs2909430 (TP53), rs1805389 (LIG4), and rs15869 (BRCA2), demonstrating an OR range from 3.7 to 19.7. The risk of subsequent meningioma associated with 4 SNPs: rs15869 (BRCA2), rs25489 (XRCC1), rs1801270 (CDKN1A), and rs1673041 (POLD1) and risk of subsequent meningioma; the ORs ranged from 2.0 to 9.6.


A model that could be used for determination of risk of subsequent CNS tumours in adult survivors of childhood cancer would be a useful tool to inform surveillance in this population, offering the potential for early detection of subsequent CNS tumours.

While this study demonstrates that it is possible to identify adult survivors from childhood cancers at risk for subsequent CNS tumours on the basis of genetic and clinical information, the low case numbers make this study a proof of principle study requiring confirmation in a larger study. 


This study was supported in part by National Cancer Institute Grants No. R01CA139633 U10CA98543 (Children’s Oncology Group [COG] Chair’s Grant), U10CA180886 (National Clinical Trials Network [NCTN] Operations Center Grant), U10CA098413 (COG Statistics and Data Center Grant), and U10CA180899 (NCTN Statistics and Data Center Grant); Leukemia and Lymphoma Society Award No. 6093-08 (S.B.), Mathew Larson Foundation Award No. MIL110389, and St Baldrick’s Foundation through an unrestricted grant to the COG.


  1. Bowers DC, Nathan PC, Constine L, et al. Lancet Oncol 2013; 14:e321-e328.


Wang X, Sun C-L, Hageman L, et al. Clinical and Genetic Risk Prediction of Subsequent CNS Tumors in Survivors of Childhood Cancer: A Report From the COG ALTE03N1 Study. JCO; Published on 4 October, 2017.