Page 399 - 2014 Printable Abstract Book
P. 399
been selected based either on promising results from main effects or interaction analyses in Stage 1, as
well as candidate SNPs including those reported as having a genome-wide significant relationship with
cancer risk in previous GWAS of women with breast cancer, and those that fall in or near genes in
pathways of interest in radiation-associated breast cancer. The analyses evaluated the significance of
known GWAS risk alleles (from GWAS of breast cancer) in predicting risk of CBC, either in the presence of
radiation therapy or not. Single-SNP and genetic-risk score (GRS) analyses of approximately 60 of the
current "known hits" were conducted, to learn whether either the individual risk alleles or the GRS are
predictive of risk of CBC overall or in association with radiation (i.e. exhibit interactions with radiation). A
second set of analyses focused on the discovery of novel SNPs which predict the risk of CBC or which may
either interact with radiation in increasing risk of subsequent cancer. Both individual SNP analyses (main
effects and interactions) will be presented as well as analyses focused on whether appropriately
constructed GRSs act as modifiers of radiation effects. In both sets of analyses the goal of the GRS x
radiation interaction analysis is to address whether the most genetically sensitive women (e.g., who are
most likely to have a CBC after an initial cancer overall) are especially sensitive to the effects of radiation.
The GRS from the first set of analyses makes a (testable) assumption that predictors of initial cancer risk
also are predictors of CBC risk. The second set of analyses relaxes this assumption and defines the GRS in
terms of the SNPs alleles that appear most associated with CBC in the WECARE Study. Because of the
limited numbers of study subjects, the WECARE Study has relatively low power for the discovery (at a
genome-wide level of significance) of genetic variants that either uniquely predict CBC risk or which
interact with radiation in raising risk. However, the power of the study to define how known susceptibility
alleles (by themselves or in combination in a GRS) synergize with radiation to increase cancer risk is much
greater. As the basis for genetic susceptibility is further understood (e.g. with ever larger GWAS studies,
and subsequent meta-analyses revealing ever more associated alleles) the possible synergy between
predicted individual genetic susceptibility and observed individual sensitivity to exposures such as
radiation will become increasingly relevant in radiation protection and clinical practice as well as cancer
etiology. Accordingly, unique data such as what the WECARE Study provides will be at the vanguard of an
expansion of knowledge on the role genetic susceptibility plays in modifying the risk of radiation-induced
cancers.
(SNE04) Genetic susceptibility to subsequent neoplasms among childhood cancer survivors. Lindsay M.
2
1
1
1
1
Morton ; Joshua N. Sampson ; Zhaoming Wang ; Gregory T. Armstrong ; Margaret A. Tucker ; Leslie L.
2
3
1
Robison ; Stephen J. Chanock ; and Smita Bhatia, National Cancer Institute, National Institutes of Health,
2
1
3
DHHS, Bethesda, MD ; St. Jude Children's Research Hospital, Memphis, TN ; and City of Hope, Duarte, CA
Childhood cancer survivors face substantially elevated risks for developing subsequent
neoplasms, which are a major cause of morbidity and mortality. Primary treatments for childhood cancer
are important contributors to subsequent neoplasm risk (e.g., radiation-related breast, skin, thyroid,
central nervous system, and lung cancers). However, little is known about the role of genetic susceptibility
outside the context of rare inherited cancer predisposition syndromes. We therefore initiated a genome-
wide association study in the Childhood Cancer Survivor Study, a long-term follow-up study of children
who were diagnosed with cancer during 1970-1986 at ≤20 years of age and survived ≥5 years after
diagnosis. Genotyping was conducted for 5324 childhood cancer survivors of European ancestry using the
Illumina HumanOmni5Exome microarray, with over 4.1 million loci passing quality control thresholds
(based on locus and sample missing rates, sample heterozygosity, and sex discrepancies). Genotype
well as candidate SNPs including those reported as having a genome-wide significant relationship with
cancer risk in previous GWAS of women with breast cancer, and those that fall in or near genes in
pathways of interest in radiation-associated breast cancer. The analyses evaluated the significance of
known GWAS risk alleles (from GWAS of breast cancer) in predicting risk of CBC, either in the presence of
radiation therapy or not. Single-SNP and genetic-risk score (GRS) analyses of approximately 60 of the
current "known hits" were conducted, to learn whether either the individual risk alleles or the GRS are
predictive of risk of CBC overall or in association with radiation (i.e. exhibit interactions with radiation). A
second set of analyses focused on the discovery of novel SNPs which predict the risk of CBC or which may
either interact with radiation in increasing risk of subsequent cancer. Both individual SNP analyses (main
effects and interactions) will be presented as well as analyses focused on whether appropriately
constructed GRSs act as modifiers of radiation effects. In both sets of analyses the goal of the GRS x
radiation interaction analysis is to address whether the most genetically sensitive women (e.g., who are
most likely to have a CBC after an initial cancer overall) are especially sensitive to the effects of radiation.
The GRS from the first set of analyses makes a (testable) assumption that predictors of initial cancer risk
also are predictors of CBC risk. The second set of analyses relaxes this assumption and defines the GRS in
terms of the SNPs alleles that appear most associated with CBC in the WECARE Study. Because of the
limited numbers of study subjects, the WECARE Study has relatively low power for the discovery (at a
genome-wide level of significance) of genetic variants that either uniquely predict CBC risk or which
interact with radiation in raising risk. However, the power of the study to define how known susceptibility
alleles (by themselves or in combination in a GRS) synergize with radiation to increase cancer risk is much
greater. As the basis for genetic susceptibility is further understood (e.g. with ever larger GWAS studies,
and subsequent meta-analyses revealing ever more associated alleles) the possible synergy between
predicted individual genetic susceptibility and observed individual sensitivity to exposures such as
radiation will become increasingly relevant in radiation protection and clinical practice as well as cancer
etiology. Accordingly, unique data such as what the WECARE Study provides will be at the vanguard of an
expansion of knowledge on the role genetic susceptibility plays in modifying the risk of radiation-induced
cancers.
(SNE04) Genetic susceptibility to subsequent neoplasms among childhood cancer survivors. Lindsay M.
2
1
1
1
1
Morton ; Joshua N. Sampson ; Zhaoming Wang ; Gregory T. Armstrong ; Margaret A. Tucker ; Leslie L.
2
3
1
Robison ; Stephen J. Chanock ; and Smita Bhatia, National Cancer Institute, National Institutes of Health,
2
1
3
DHHS, Bethesda, MD ; St. Jude Children's Research Hospital, Memphis, TN ; and City of Hope, Duarte, CA
Childhood cancer survivors face substantially elevated risks for developing subsequent
neoplasms, which are a major cause of morbidity and mortality. Primary treatments for childhood cancer
are important contributors to subsequent neoplasm risk (e.g., radiation-related breast, skin, thyroid,
central nervous system, and lung cancers). However, little is known about the role of genetic susceptibility
outside the context of rare inherited cancer predisposition syndromes. We therefore initiated a genome-
wide association study in the Childhood Cancer Survivor Study, a long-term follow-up study of children
who were diagnosed with cancer during 1970-1986 at ≤20 years of age and survived ≥5 years after
diagnosis. Genotyping was conducted for 5324 childhood cancer survivors of European ancestry using the
Illumina HumanOmni5Exome microarray, with over 4.1 million loci passing quality control thresholds
(based on locus and sample missing rates, sample heterozygosity, and sex discrepancies). Genotype
