Search

Perspective on this Article from An Expanded Risk Prediction Model for Lung Cancer

Demographic and Clinical Characteristics of Glioma Patients in the Parent Epidemiological Study (N=1247).

Supplemental Figures 1-2 including: (1) Supplementary Figure 1. Genome-wide genetic association and lung cancer risk in discovery stage; (2) Supplementary Figure 2. Genome-wide gene-asbestos exposure interactions and lung cancer risk in discovery stage.

Background: Occupational asbestos exposure has been found to increase lung cancer risk in epidemiologic studies.Methods: We conducted an asbestos exposure–gene interaction analyses among several Caucasian populations who were current or ex-smokers. The discovery phase included 833 Caucasian cases and 739 Caucasian controls, and used a genome-wide association study (GWAS) to identify single-nucleotide polymorphisms (SNP) with gene–asbestos interaction effects. The top ranked SNPs from the discovery phase were replicated within the International Lung and Cancer Consortium (ILCCO). First, in silico replication was conducted in those groups that had GWAS and asbestos exposure data, including 1,548 cases and 1,527 controls. This step was followed by de novo genotyping to replicate the results from the in silico replication, and included 1,539 cases and 1,761 controls. Multiple logistic regression was used to assess the SNP–asbestos exposure interaction effects on lung cancer risk.Results: We observed significantly increased lung cancer risk among MIRLET7BHG (MIRLET7B host gene located at 22q13.31) polymorphisms rs13053856, rs11090910, rs11703832, and rs12170325 heterozygous and homozygous variant allele(s) carriers (P < 5 × 10−7 by likelihood ratio test; df = 1). Among the heterozygous and homozygous variant allele(s) carriers of polymorphisms rs13053856, rs11090910, rs11703832, and rs12170325, each unit increase in the natural log-transformed asbestos exposure score was associated with age-, sex-, smoking status, and center-adjusted ORs of 1.34 [95% confidence interval (CI), 1.18–1.51], 1.24 (95% CI, 1.14–1.35), 1.28 (95% CI, 1.17–1.40), and 1.26 (95% CI, 1.15–1.38), respectively, for lung cancer risk.Conclusion: Our findings suggest that MIRLET7BHG polymorphisms may be important predictive markers for asbestos exposure–related lung cancer.Impact: To our knowledge, our study is the first report using a systematic genome-wide analysis in combination with detailed asbestos exposure data and replication to evaluate asbestos-associated lung cancer risk. Cancer Epidemiol Biomarkers Prev; 24(10); 1564–73. ©2015 AACR.

Background: Tobacco-induced lung cancer is characterized by a deregulated inflammatory microenvironment. Variants in multiple genes in inflammation pathways may contribute to risk of lung cancer.Methods: We therefore conducted a three-stage comprehensive pathway analysis (discovery, replication, and meta-analysis) of inflammation gene variants in ever-smoking lung cancer cases and controls. A discovery set (1,096 cases and 727 controls) and an independent and nonoverlapping internal replication set (1,154 cases and 1,137 controls) were derived from an ongoing case–control study. For discovery, we used an iSelect BeadChip to interrogate a comprehensive panel of 11,737 inflammation pathway single-nucleotide polymorphisms (SNP) and selected nominally significant (P < 0.05) SNPs for internal replication.Results: There were six SNPs that achieved statistical significance (P < 0.05) in the internal replication data set with concordant risk estimates for former smokers and five concordant and replicated SNPs in current smokers. Replicated hits were further tested in a subsequent meta-analysis using external data derived from two published genome-wide association studies (GWAS) and a case–control study. Two of these variants (a BCL2L14 SNP in former smokers and an SNP in IL2RB in current smokers) were further validated. In risk score analyses, there was a 26% increase in risk with each additional adverse allele when we combined the genotyped SNP and the most significant imputed SNP in IL2RB in current smokers and a 36% similar increase in risk for former smokers associated with genotyped and imputed BCL2L14 SNPs.Conclusions/Impact: Before they can be applied for risk prediction efforts, these SNPs should be subject to further external replication and more extensive fine mapping studies. Cancer Epidemiol Biomarkers Prev; 21(7); 1213–21. ©2012 AACR.

PDF file - 54K, Summary of Inflammation subpathways, genes and SNP's in Illumina chip

Background: Established psychosocial risk factors increase the risk for experimentation among Mexican origin youth. Now, we comprehensively investigate the added contribution of select polymorphisms in candidate genetic pathways associated with sensation seeking, risk taking, and smoking phenotypes to predict experimentation.Methods: Participants (N = 1,118 Mexican origin youth) recruited from a large population-based cohort study in Houston, TX, provided prospective data on cigarette experimentation over 3 years. Psychosocial data were elicited twice—baseline and final follow-up. Participants were genotyped for 672 functional and tagging variants in the dopamine, serotonin, and opioid pathways.Results: After adjusting for gender and age, with a Bayesian False Discovery Probability set at 0.8 and prior probability of 0.05, six gene variants were significantly associated with risk of experimentation. After controlling for established risk factors, multivariable analyses revealed that participants with six or more risk alleles were 2.25 [95% confidence interval (CI), 1.62–3.13] times more likely to have experimented since baseline than participants with five or fewer. Among committed never-smokers (N = 872), three genes (OPRM1, SNAP25, HTR1B) were associated with experimentation as were all psychosocial factors. Among susceptible youth (N = 246), older age at baseline, living with a smoker, and three different genes (HTR2A, DRD2, SLC6A3) predicted experimentation.Conclusions: Our findings, which have implications for development of culturally specific interventions, need to be validated in other ethnic groups.Impact: These results suggest that variations in select genes interact with a cognitive predisposition toward smoking. In susceptible adolescents, the impact of the genetic variants appears to be larger than committed never-smokers. Cancer Epidemiol Biomarkers Prev; 21(1); 228–38. ©2011 AACR.

XLS - 104K

PDF file, 63KB, Minor allele frequencies (MAF), genotype counts, and results in the 13-cRA treatment arm.

Purpose: Accumulating evidence supports the contention that genetic variation is associated with neurocognitive function in healthy individuals and increased risk for neurocognitive decline in a variety of patient populations, including cancer patients. However, this has rarely been studied in glioma patients.Experimental Design: To identify the effect of genetic variants on neurocognitive function, we examined the relationship between the genotype frequencies of 10,967 single-nucleotide polymorphisms in 580 genes related to five pathways (inflammation, DNA repair, metabolism, cognitive, and telomerase) and neurocognitive function in 233 newly diagnosed glioma patients before surgical resection. Four neuropsychologic tests that measured memory (Hopkins Verbal Learning Test—Revised), processing speed (Trail Making Test A), and executive function (Trail Making Test B, Controlled Oral Word Association) were examined.Results: Eighteen polymorphisms were associated with processing speed and 12 polymorphisms with executive function. For processing speed, the strongest signals were in IRS1 rs6725330 in the inflammation pathway (P = 2.5 × 10−10), ERCC4 rs1573638 in the DNA repair pathway (P = 3.4 × 10−7), and ABCC1 rs8187858 in metabolism pathway (P = 6.6 × 10−7). For executive function, the strongest associations were in NOS1 rs11611788 (P = 1.8 × 10−8) and IL16 rs1912124 (P = 6.0 × 10−7) in the inflammation pathway, and POLE rs5744761 (P = 6.0 × 10−7) in the DNA repair pathway. Joint effect analysis found significant gene polymorphism-dosage effects for processing speed (Ptrend = 9.4 × 10−16) and executive function (Ptrend = 6.6 × 10−15).Conclusions: Polymorphisms in inflammation, DNA repair, and metabolism pathways are associated with neurocognitive function in glioma patients and may affect clinical outcomes. Clin Cancer Res; 21(14); 3340–6. ©2015 AACR.

Purpose: We performed a study to determine if a fluorescence in situ hybridization (FISH)–based assay using isolated peripheral blood mononuclear cells (PBMCs) with DNA probes targeting specific sites on chromosomes known to have abnormalities in non–small cell lung cancer (NSCLC) cases could detect circulating genetically abnormal cells (CACs).Experimental Design: We evaluated 59 NSCLC cases with stage I through IV disease and 24 controls. PBMCs and matched tumors were hybridized with 2 two-color [3p22.1/CEP3 and 10q22.3 (SP-A)/CEP10) and 2 four-color [CEP3, CEP7, CEP17, and 9p21.3 (URO); and EGFR, c-MYC, 6p11-q11, and 5p15.2 (LAV)] FISH probes. Percentages of cytogenetically abnormal cells (CACs) in peripheral blood and in matched tumor specimens were quantified by using an automated fluorescent scanner. Numbers of CACs were calculated based on the percentage of CACs (defined as PBMCs with genetic abnormalities) per milliliter of blood and expressed per microliter of blood.Results: Patients with NSCLC had significantly higher numbers of CACs than controls. Mean number of CACs ranged from 7.23 ± 1.32/μL for deletions of 10q22.3/CEP10 to 45.52 ± 7.49/μL for deletions of 3p22.1/CEP3. Numbers of CACs with deletions of 3p22.1, 10q22.3, and 9p21.3, and gains of URO, increased significantly from early to advanced stage of disease.Conclusions: We have developed a sensitive and quantitative antigen-independent FISH-based test for detecting CACs in peripheral blood of patients with NSCLC, which showed a significant correlation with the presence of cancer. If this pilot study can be validated in a larger study, CACs may have a role in the management of patients with NSCLC. Clin Cancer Res; 16(15); 3976–87. ©2010 AACR.

Purpose: The development of second primary tumors (SPT) or recurrence alters prognosis for curatively treated head and neck squamous cell carcinoma (HNSCC) patients. The 13-cis-Retinoic acid (13-cRA) has been tested as a chemoprevention agent in clinical trials with mixed results. Therefore, we investigated whether genetic variants in the PI3K/PTEN/AKT/mTOR pathway could serve as biomarkers to identify which patients are at high risk of an SPT/recurrence, while also predicting response to 13-cRA chemoprevention.Experimental Design: A total of 137 pathway single-nucleotide polymorphisms were genotyped in 440 patients from the Retinoid Head and Neck Second Primary Trial and assessed for SPT/recurrence risk and response to 13-cRA. Risk models were created based on epidemiology, clinical, and genetic data.Results: Twenty-two genetic loci were associated with increased SPT/recurrence risk, with six also being associated with a significant benefit following chemoprevention. Combined analysis of these high-risk/high-benefit loci identified a significant (P = 1.54 × 10−4) dose–response relationship for SPT/recurrence risk, with patients carrying four to five high-risk genotypes having a 3.76-fold [95% Confidence Interval (CI), 1.87–7.57] increase in risk in the placebo group (n = 215). Patients carrying four to five high-risk loci showed the most benefit from 13-cRA chemoprevention, with a 73% reduction in SPT/recurrence (95% CI, 0.13–0.58) compared with those with the same number of high-risk genotypes who were randomized to receive placebo. Incorporation of these loci into a risk model significantly improved the discriminatory ability over models with epidemiology, clinical, and previously identified genetic variables.Conclusions: These results show that loci within this important pathway could identify individuals with a high-risk/high-benefit profile and are a step toward personalized chemoprevention for HNSCC patients. Clin Cancer Res; 18(13); 3705–13. ©2012 AACR.

Purpose: The development of second primary tumors (SPT) or recurrence alters prognosis for curatively treated head and neck squamous cell carcinoma (HNSCC) patients. The 13-cis-Retinoic acid (13-cRA) has been tested as a chemoprevention agent in clinical trials with mixed results. Therefore, we investigated whether genetic variants in the PI3K/PTEN/AKT/mTOR pathway could serve as biomarkers to identify which patients are at high risk of an SPT/recurrence, while also predicting response to 13-cRA chemoprevention. Experimental Design: A total of 137 pathway single-nucleotide polymorphisms were genotyped in 440 patients from the Retinoid Head and Neck Second Primary Trial and assessed for SPT/recurrence risk and response to 13-cRA. Risk models were created based on epidemiology, clinical, and genetic data. Results: Twenty-two genetic loci were associated with increased SPT/recurrence risk, with six also being associated with a significant benefit following chemoprevention. Combined analysis of these high-risk/high-benefit loci identified a significant (P = 1.54 × 10−4) dose–response relationship for SPT/recurrence risk, with patients carrying four to five high-risk genotypes having a 3.76-fold [95% Confidence Interval (CI), 1.87–7.57] increase in risk in the placebo group (n = 215). Patients carrying four to five high-risk loci showed the most benefit from 13-cRA chemoprevention, with a 73% reduction in SPT/recurrence (95% CI, 0.13–0.58) compared with those with the same number of high-risk genotypes who were randomized to receive placebo. Incorporation of these loci into a risk model significantly improved the discriminatory ability over models with epidemiology, clinical, and previously identified genetic variables. Conclusions: These results show that loci within this important pathway could identify individuals with a high-risk/high-benefit profile and are a step toward personalized chemoprevention for HNSCC patients. Clin Cancer Res; 18(13); 3705–13. ©2012 AACR.

Background: Tobacco-induced lung cancer is characterized by a deregulated inflammatory microenvironment. Variants in multiple genes in inflammation pathways may contribute to risk of lung cancer. Methods: We therefore conducted a three-stage comprehensive pathway analysis (discovery, replication, and meta-analysis) of inflammation gene variants in ever-smoking lung cancer cases and controls. A discovery set (1,096 cases and 727 controls) and an independent and nonoverlapping internal replication set (1,154 cases and 1,137 controls) were derived from an ongoing case–control study. For discovery, we used an iSelect BeadChip to interrogate a comprehensive panel of 11,737 inflammation pathway single-nucleotide polymorphisms (SNP) and selected nominally significant (P < 0.05) SNPs for internal replication. Results: There were six SNPs that achieved statistical significance (P < 0.05) in the internal replication data set with concordant risk estimates for former smokers and five concordant and replicated SNPs in current smokers. Replicated hits were further tested in a subsequent meta-analysis using external data derived from two published genome-wide association studies (GWAS) and a case–control study. Two of these variants (a BCL2L14 SNP in former smokers and an SNP in IL2RB in current smokers) were further validated. In risk score analyses, there was a 26% increase in risk with each additional adverse allele when we combined the genotyped SNP and the most significant imputed SNP in IL2RB in current smokers and a 36% similar increase in risk for former smokers associated with genotyped and imputed BCL2L14 SNPs. Conclusions/Impact: Before they can be applied for risk prediction efforts, these SNPs should be subject to further external replication and more extensive fine mapping studies. Cancer Epidemiol Biomarkers Prev; 21(7); 1213–21. ©2012 AACR.

Background: Established psychosocial risk factors increase the risk for experimentation among Mexican origin youth. Now, we comprehensively investigate the added contribution of select polymorphisms in candidate genetic pathways associated with sensation seeking, risk taking, and smoking phenotypes to predict experimentation. Methods: Participants (N = 1,118 Mexican origin youth) recruited from a large population-based cohort study in Houston, TX, provided prospective data on cigarette experimentation over 3 years. Psychosocial data were elicited twice—baseline and final follow-up. Participants were genotyped for 672 functional and tagging variants in the dopamine, serotonin, and opioid pathways. Results: After adjusting for gender and age, with a Bayesian False Discovery Probability set at 0.8 and prior probability of 0.05, six gene variants were significantly associated with risk of experimentation. After controlling for established risk factors, multivariable analyses revealed that participants with six or more risk alleles were 2.25 [95% confidence interval (CI), 1.62–3.13] times more likely to have experimented since baseline than participants with five or fewer. Among committed never-smokers (N = 872), three genes (OPRM1, SNAP25, HTR1B) were associated with experimentation as were all psychosocial factors. Among susceptible youth (N = 246), older age at baseline, living with a smoker, and three different genes (HTR2A, DRD2, SLC6A3) predicted experimentation. Conclusions: Our findings, which have implications for development of culturally specific interventions, need to be validated in other ethnic groups. Impact: These results suggest that variations in select genes interact with a cognitive predisposition toward smoking. In susceptible adolescents, the impact of the genetic variants appears to be larger than committed never-smokers. Cancer Epidemiol Biomarkers Prev; 21(1); 228–38. ©2011 AACR.

Background: Breast cancer is associated with treatment-related infertility and has been demonstrated to be a major concern for premenopausal survivors. Detection of a BRCA deleterious mutation may also affect attitudes regarding future childbearing. Preimplantation genetic diagnosis (PGD) allows women to use in vitro fertilization (IVF) to implant only those embryos without a BRCA mutation. The ability to test the fetus for BRCA mutations is also available through amniocentesis and chorionic villus sampling (CVS). The objective of this study was to evaluate attitudes about childbearing and fertility in women being evaluated for a BRCA mutation. Methods: Women with childbearing potential who were referred to the Clinical Cancer Genetics Clinic to be evaluated for a BRCA mutation were invited to participate in this survey. The questionnaire was administered prior to genetic counseling. A follow-up was administered after the BRCA results were disclosed. The survey queried participants regarding their attitudes on fertility, pregnancy as it may relate to cancer and the potential of a BRCA mutation. Other questions detailed attitudes regarding IVF, PGD, and CVS in these instances. Descriptive statistics were used. Results: One hundred and twenty-eight women completed pre-questionnaires and to date 76 have completed post results disclosure questionnaires. The mean age was 33 (range 21–44) with 69.5% with a diagnosis of breast cancer, 39.8% received chemotherapy and 60.9% already had at least 1 biological child. A future child was desired by 45.3% although 53.1% worried that their children would have an increased risk of cancer. Regarding PGD, although only 30.9% (38/123) said that they would use PGD, 80.2% felt that the testing should be available to families with inherited cancers. Regarding fetal testing via amniocentesis or CVS, 29.7% would have the fetus tested and 7% would consider termination if a genetic mutation was identified. Additionally, 69.5% felt it was important to receive fertility counseling and treatment at the same place where they receive their cancer care. To date 8 women have been diagnosed with a BRCA1 mutation and 4 with a BRCA2 mutation. When asked similar questions after their genetic results were disclosed, 2 women who had previously stated they would not use PGD changed their mind. Conclusions: Future pregnancies are important to many breast cancer survivors. BRCA mutation carriers have the option to have children without passing on their genetic risk for cancer. Although few would use these interventions, a large majority felt it was important to have information about these choices and to have options for fertility preservation options addressed at the center where cancer care is delivered. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-13-05.

Because existing risk prediction models for lung cancer were developed in white populations, they may not be appropriate for predicting risk among African-Americans. Therefore, a need exists to construct and validate a risk prediction model for lung cancer that is specific to African-Americans. We analyzed data from 491 African-Americans with lung cancer and 497 matched African-American controls to identify specific risks and incorporate them into a multivariable risk model for lung cancer and estimate the 5-year absolute risk of lung cancer. We performed internal and external validations of the risk model using data on additional cases and controls from the same ongoing multiracial/ethnic lung cancer case-control study from which the model-building data were obtained as well as data from two different lung cancer studies in metropolitan Detroit, respectively. We also compared our African-American model with our previously developed risk prediction model for whites. The final risk model included smoking-related variables [smoking status, pack-years smoked, age at smoking cessation (former smokers), and number of years since smoking cessation (former smokers)], self-reported physician diagnoses of chronic obstructive pulmonary disease or hay fever, and exposures to asbestos or wood dusts. Our risk prediction model for African-Americans exhibited good discrimination [75% (95% confidence interval, 0.67–0.82)] for our internal data and moderate discrimination [63% (95% confidence interval, 0.57–0.69)] for the external data group, which is an improvement over the Spitz model for white subjects. Existing lung cancer prediction models may not be appropriate for predicting risk for African-Americans because (a) they were developed using white populations, (b) level of risk is different for risk factors that African-American share with whites, and (c) unique group-specific risk factors exist for African-Americans. This study developed and validated a risk prediction model for lung cancer that is specific to African-Americans and thus more precise in predicting their risks. These findings highlight the importance of conducting further ethnic-specific analyses of disease risk.

The multi-endpoint cytokinesis-blocked micronucleus assay is used for assessing chromosome aberrations. We have recently reported that this assay is extremely sensitive to genetic damage caused by the tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and that the binucleated cells with micronuclei, nucleoplasmic bridges, and nuclear buds in lymphocytes (chromosome damage endpoints measured by the assay) are strong predictors of lung cancer risk. In the current study, we refined our analysis to include toxicity endpoints (micronuclei in mononucleated cells, apoptosis, necrosis, and nuclear division index) to investigate the benefit of including these variables on improving the predictive value of the assay. Baseline and NNK-induced micronuclei in mononucleated cells were significantly higher in patients (n = 139) than controls (n = 130; P < 0.001). Baseline apoptosis was higher among cases; however, the controls showed a significant higher fold increase in NNK-induced apoptosis compared with baseline (P < 0.001). Principal components analysis was used to derive a summary measure for all endpoints and calculate the positive predictive value (PPV) and negative predictive value (NPV) for disease status. First principal component for NNK-induced chromosome damage endpoints (binucleated cells with micronuclei, nucleoplasmic bridges, and nuclear buds) had an area under the curve = 97.9 (95% confidence interval, 95.9-99.0), PPV = 94.8, and NPV = 92.6. The discriminatory power improved when micronuclei in mononucleated cells were included: area under the curve = 99.1 (95% confidence interval, 97.9-100.0), PPV = 98.7 and NPV = 95.6. The simplicity, rapidity, and sensitivity of the assay together with potential for automation make it a valuable tool for screening and prioritizing potential cases for intensive screening. (Cancer Epidemiol Biomarkers Prev 2008;17(5):1111–9)

In this case-control study, we modified the cytokinesis-block micronucleus (CBMN) assay, an established biomarker for genomic instability, to evaluate susceptibility to the nicotine-derived nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) by measuring the frequency of NNK-induced chromosomal damage endpoints (micronuclei, nucleoplasmic bridges, and nuclear buds) per 1,000 binucleated lymphocytes. Spontaneous and NNK-induced chromosomal damage were significantly higher in lung cancer patients compared with controls. Forty-seven percent of cases (versus 12% of controls) had ≥4 spontaneous micronuclei, 66% of cases (and no controls) had ≥4 spontaneous nucleoplasmic bridges, and 25% of cases (versus 5% of controls) had ≥1 spontaneous nuclear bud (P < 0.001). Similarly, 40% of cases (versus 6% of the controls) had ≥5 NNK-induced micronuclei, 89% of cases (and no controls) had ≥6 induced nucleoplasmic bridges, and 23% of cases (versus 2% of controls) had ≥2 induced nuclear buds (P < 0.001). When analyzed on a continuous scale, spontaneous micronuclei, nucleoplasmic bridges, and nuclear buds were associated with 2-, 29-, and 6-fold increases in cancer risk, respectively. Similarly, NNK-induced risks were 2.3-, 45.5-, and 10-fold, respectively. We evaluated the use of CBMN assay to predict cancer risk based on the numbers of micronuclei, nucleoplasmic bridges, and nuclear buds defined by percentile cut points in controls. Probabilities of being a cancer patient were 96%, 98%, and 100% when using the 95th percentiles of spontaneous and NNK-induced micronuclei, nucleoplasmic bridges, and nuclear buds, respectively. Our study indicates that the CBMN assay is extremely sensitive to NNK-induced genetic damage and may serve as a strong predictor of lung cancer risk. (Cancer Res 2006; 66(12): 6449-56)

Introduction: In the past decade, advances in genetics have led to the discovery of numerous lung cancer susceptibility variants. The majority of these variants have been found to influence lung cancer susceptibility via tobacco exposure or nicotine addiction; however, recently, researchers have observed lung cancer susceptibility variants that mediate through Chronic Obstructive Pulmonary Disease (COPD). Studies involving potential genetic factors related to other lung-related conditions, such as pneumonia and/or hay fever have been limited. The genetic variants related to underlying mechanism of hay fever on the development of lung cancer are interesting to pursue as hay fever has been showed to be protective against lung cancer. Methods: Cases included 1154 histological-confirmed Caucasian lung cases from MD Anderson Cancer Center in Houston, Texas, and controls included 1137 individuals recruited through the Kelsey-Seybold Clinics in Houston, Texas. These cases and controls were a subset of participants from a lung cancer case-control study conducted that has available Genome-wide association study (GWAS) data (317,498 SNPs). We first conducted an association analysis in PLINK to determine the association of each SNP with both lung cancer and hay fever and find SNPs that showed a joint significance for both hay fever and lung cancer in opposite directions (p-value < 0.05 for both). We then performed mediation analyses with the SNPs that showed significance for both hay fever and lung cancer. Finally, we inferred SNPs in regions that contained a collection of tag SNPs that mediate lung cancer risk through hay fever. Results: Two hundred and forty six SNPs were found to be statistically significant (p-value < 0.05) for both hay fever (protection) and lung cancer (risk), and 76 of these SNPs maintained their significance level after mediation analysis. SNP rs7159751 on chromosome 14 had the highest mediation effect of 19.83%. A collection of SNPs on the Neuregulin 3 gene (NRG3) that mediate lung cancer risk through hay fever was found on Chromosome 10 which has been shown to activate the JAK-STAT signal transduction pathway and stimulate lung epithelial cell proliferation. Conclusion: This is the first study to have investigated the mediation effects of hay fever on lung cancer risk. Our data supports the mediation role of certain SNPs in lung cancer through hay fever and points to specific transduction pathway. Future studies are needed to validate these results in external populations. Citation Format: Anthony D'Amelio, Chi H. Nguyen, Randa El-Zein, Margaret R. Spitz, Xifeng Wu, Carol J. Etzel. Evidence for genetic mediation of lung cancer through hay fever. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 1345. doi:10.1158/1538-7445.AM2013-1345

Introduction: Activation of the DNA damage response (DDR) is an important cellular mechanism for maintaining genomic integrity in the face of genotoxic stress. While the cellular response to genotoxic stress has been extensively studied in cancer models, less is known about the cellular response to oncogenic stress in the premalignant context. Methods: DNA damage (γ-H2AX), cell proliferation (ki-67) and apoptosis (cleaved caspase-3) were assessed by immunohistochemistry in breast tissue samples from 3 cohorts of women with: i) histologically normal breast tissue (n=50), ii) histological changes of increased risk (LCIS, ADH, ALH) (n=54) but without a personal history of breast cancer and iii) ductal carcinoma in situ (n=46). DNA damage parameters were also assessed in a panel of cell lines derived from the MCF-10A line that represents the multi-step process of breast tumorigenesis (AT1, DCIS, CA1d). Lastly, we generated stable cell clones over expressing cyclin E in the normal mammary epithelial cell line, MCF10A. DNA damage was measured by immuno fluorescence of γ-H2AX in the cell lines models. Apoptosis and autophagy were measured by staining cells with acridine orange and annexin V respectively followed by cell sorting with flow cytometry. Protein expression of AKt-mTOR pathway was assessed by western blotting. Results: Breast tissues samples from women at different risk levels for invasive breast cancer revealed that DNA damage is inversely correlated with breast cancer risk with the highest levels of damage seen in histologically normal tissues and lowest levels in DCIS (p 10 foci= 19%) and lowest levels (> 10 foci= 0%) in the DCIS line and in the invasive cancer cell line CA1d (p Citation Format: Anjana Bhardwaj, Mei Liu, Yan Liu, Nivetha Ganesan, Daniel Rosen, Carol Etzel, Constance T. Albaracin, Qiang Hao, Ashley Gullett, Isabelle Bedrosian. Suppression of Akt-mTOR pathway: a novel component of oncogene induced DNA damage response (DDR) barrier in breast. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 634. doi:10.1158/1538-7445.AM2013-634

Models predicting liver cancer are only available for the high risk population with hepatitis carriers but not for the general population. we developed risk prediction models for liver cancer using data collected from a large health screening program in Taiwan with long term follow up. The cohort included 444,023 subjects at baseline. During a median follow up time of 7.4 years, 1,668 subjects developed liver cancer. Epidemiologic data, medical history and routine blood panel including serum transaminases data were collected with optional additional testing for HBV or HCV status. Liver cancer cases were ascertained by computerized record linkage with both the National Cancer Registry and the National Death Certification profiles in Taiwan. Stepwise Cox regression analysis was performed to identify significant predictors in the multivariate models. Individualized risk of developing liver cancer in 10 years was calculated from baseline probability and relative risk profile estimated from the Cox regression model. Models were developed separately to provide risk prediction for subjects who chose to have HCV tested (130,533 subjects at baseline and 416 liver cancer cases) and who chose otherwise (313,490 subjects and 1,252 liver cancer cases). Since the results for both sub-cohorts are comparable and we only reported results for the sub-cohort with HCV tested. In the sub-cohort with HCV tested, the model with only questionnaire data identified significant main effects for male gender, older age, prior history of diabetes, pack year of smoking, alcohol use, and physical inactivity. This epidemiologic model had an AUC of 0.798 (95% CI=0.772-0.815) for 10-year risk prediction. With only data from measures of Transaminases, the model achieved an AUC of 0.927 (95% CI=0.911-0.938), a significant increase as compared to the model with only epidemiologic variables. The addition of HBV improved the AUC to 0.936 (95% CI=0.913-0.958). Similar results were obtained for sub-cohort without HCV testing. The addition of HCV status in combination with HBV status, the AUC improve to 0.941 (95% CI=0.918-0.967). Among models using key history and blood panels, the use of serum Transaminases only, available in routine health check-ups, was able to provide most information needed in predicting liver cancer in the general public. Additional testing of HBV and /or HCV further increased the prediction power. This simple transaminase-based model for the general public can be valuable in clinical practice for identifying high risk individuals. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 5509. doi:1538-7445.AM2012-5509

Introduction: To further understand and elucidate the genetic risk factors for Lung Cancer (LC) will substantially improve cancer prevention, screening programs, and treatment options for this insidious disease. However, our previous attempts to increase the discriminatory power in an existing risk model with the inclusion of SNPs identified via GWAS have not been very successful. In this analysis, we reassess the increase in discriminatory power by including haplotypes based on existing GWAS-identified SNPs and additional single SNPs. Methods: Cases included 1154 histological-confirmed Caucasian lung cases from MD Anderson Cancer Center in Houston, Texas, and controls included 1137 individuals recruited through the Kelsey-Seybold Clinics in Houston, Texas. Genetic information from both populations was selected for 157 SNPs that exist in the Texas GWAS and were part of the top 200 SNPs as determined by a previously published meta-analysis of 10 different lung cancer GWAS datasets (Landi et al. 2009). Plink and Haploview were used to determine the top SNPs and top haplotype blocks respectively. Then, univariate logistic regression was used to determine the best genetic model for the top SNPs. Joint logistic risk model regression was used to determine haplotype risk for each individual haplotype block. Finally, multivariate logistic regression was used to develop the final extended model that included the original set of risk factors from the Spitz model plus additional haplotypes and SNPs. We evaluated increase in discriminatory power by evaluated change in the area under the receiver-operating characteristic curve and the net reclassification index. Results: The discriminatory power for the original Spitz model was 66.1% (95% CI = 0.638-0.683). With the inclusion of only single SNPs in to the model, the discriminatory power increased by 4.5% to 70.6% (95% CI = 0.684-0.727, p-value = Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr LB-326. doi:1538-7445.AM2012-LB-326

One critical aspect of cancer prevention is the identification of high risk individuals who may benefit from active surveillance, screening trials or targeted chemoprevention. In the development, validation and application of risk models to quantify high risk, one of many questions arises, “Can one model be sufficient to estimate risk for a cancer across different populations?” In 2003, Gonzalez Burchard et al. posted to the Sounding Board section of the New England Journal of Medicine (March 20, 2003) about the importance of race and ethnic background in biomedical research. They advise that an individual's race and ethnic information is crucial for studying differences in prevalence of disease and also for the identification of different disease risk-factor profiles - an outcome of risk prediction models. In response to this posting, Swallen further pointed out that just because race is not a genetic variable, this does not mean that it is not an important one to consider in biomedical research. In contrast, Cooper et al. note that the topic of race is contentious and guards against its use as defined by genetic variants. In our lung cancer experience, we have observed that 1) prevalence of cancer varies among racial/ethnic groups, 2) different racial/ethnic groups may share environmental/occupational risk factors, but the prevalence of the risk factors and the levels of risk on the cancer of interest may be different, 3) a group that is geographically distinct may have risk factors which are unique, 4) there may exist biological differences among groups which result in varying genetic disease etiologies, 5) interactions among the varying genetic and environmental risk factors also result in varying levels of risk among groups. These observations prompted us to develop and validate group-specific risk models for lung cancer - to date, one for Caucasians, one for African Americans. The differences in risks and hence the need for group-specific risk models are not necessarily differences in socioeconomic status (SES, such as education level, income, and lack of insurance coverage) that have been implicated as the ‘sole cause’ of racial disparities in terms of cancer risk differences, but differences in key risk factors that are unrelated to SES or other economic-related factors, such as host-susceptibility, occupational and environmental exposures, genetic factors or interactions among these. While some may cite ethical dilemmas, especially in studies investigating genetic differences, and high costs associated with developing population-based and minority-specific cancer registries that could be obstacles in the development and application of group-specific risk models, we must think beyond these obstacles so that we can develop and implement the most valid and applicable models possible. In doing so, we can start to move toward personalized risk prediction, which has the promise to benefit those at the highest risks and not just those who are more likely to be randomly selected for a study. In this presentation, I will demonstrate that it is not proper to just assume that risk models constructed using data from Caucasians populations convey the same level of discriminatory power in diverse populations, but instead, it is necessary to construct group-specific risk models for application to diverse multi-racial, multi-ethnic populations. These demonstrations will be highlighted in existing risk models for lung, liver and breast cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr SY09-03. doi:1538-7445.AM2012-SY09-03

Background: There are an estimated 45 million current smokers in the US and 49 million former smokers. The challenge is to construct reliable risk prediction models to identify that fraction of smokers most likely to get lung cancer. High-risk individuals could undergo a program of screening surveillance that might not be appropriate for a lower risk population and receive the most intensive smoking cessation interventions. Models with improved discriminatory ability will also have clinical benefits. In the US, 150,000 patients are diagnosed with focal pulmonary lesions annually, with the prevalence of malignancy ranging from 10 to 70%. Current guidelines are to employ best clinical judgment and/or validated models. Finally, risk prediction tools could be incorporated into the design of smaller, more powerful, and “smarter” prevention trials. Cardiovascular and type 2 diabetes risk profiles using epidemiologic data have been effectively used for nearly two decades. In the cancer arena, risk prediction models for breast cancer have the longest history, although models have also been generated for prostate, lung, melanoma, ovary, colorectal, and bladder cancers. The advent of genome-wide association studies to identify low-penetrance common susceptibility alleles now heralds the possibility of incorporating panels of gene variants into existing models and to assess improvement in model performance. However, to date, the expanded models for many cancer sites have shown only modest incremental improvements in discrimination. Lung Risk Models: We and others have explored various approaches for lung cancer prediction. Such tools hold promise, but their interpretation is complex. Epidemiologic/clinical: Bach published the original risk prediction model based on data from the Carotene and Retinol Efficacy Trial (CARET) of 14,000 heavy smokers and >4,000 asbestos-exposed men, mostly white (1). The variables included age, gender, asbestos exposure, smoking history, cigarettes per day, duration of smoking and of cessation. The Bach model has been validated in an independent data from the placebo arm of the Alpha-Tocopherol, Beta-Carotene Cancer Prevention study. Our baseline model was based on data derived from an ongoing lung cancer case control study (2). Variables included were environmental tobacco smoke (for never and former smokers only), family cancer history, asbestos and dust exposures, prior respiratory disease, history of hay fever, and smoking history variables. All variables have strong biologically plausible etiologic roles, and are relatively easy to ascertain through patient interview. The validated area under the curve (AUC) statistics for the former and current smoker models were modest (0.63, and 0.58, respectively), although consistent with those from other risk prediction models. If age and smoking status (case-control matching variables) had been included in the models, the models would likely have performed far better. The Liverpool Lung Project (LLP) variables (3) include smoking duration, prior diagnosis of pneumonia, occupational exposure to asbestos, prior cancer diagnosis and family history of lung cancer. Most recently a carefully constructed model based on data from 70,962 control subjects in the Prostate, Lung, Colorectal, Ovarian cancer screening trial (PLCO) was published (4). This model includes age, SES (education), BMI, family history of lung cancer, COPD, recent chest x-ray, smoking status (never, former, or current), pack-years smoked, and smoking duration. A second model also included smoking quit-time. External validation was performed with 44,223 PLCO intervention arm participants. In the external validation sample, models 1 and 2 had area under the curves of 0.841 and 0.784, respectively. These models had high discrimination in women, men, whites, and nonwhites. Extended Models: Functional Data: We have estimated the improvement in model performance by incorporating two measures of DNA repair capacity that have been shown in case-control analyses to be associated with increased lung cancer risk (5). Addition of the biomarker assays does improve the sensitivity of the models over epidemiologic and clinical data alone. These in vitro lymphocyte culture assays, however, are time-consuming and require some level of technical expertise. Therefore while feasible in a controlled academic setting, they are not applicable for widespread population-based implementation. Genetic Data: Genetic data are stable, inexpensive to ascertain, accurate and amenable to high-throughput analysis. However, it remains uncertain whether enriched multimarker models give better discrimination. We have added three SNPS that were most significant in GWAS of lung cancer to the baseline lung epidemiologic model. These include a replicated SNP in the 15q25 chromosomal region that encompasses the nicotinic acetylcholine receptor subunit genes, CHRNA3 and CHRNA5, that have a defined role in nicotine dependence, and a hypothesized direct role in downstream signaling pathways that promote carcinogenesis. We also included two SNPs from the 5p15.33 locus (rs2736100 and rs401681) that contains two known genes: the TERT (human telomerase reverse transcriptase) gene and the CLPTM1L. The AUC showed only modest improvement. Such small increases in discriminative accuracy are unlikely to be of diagnostic or predictive utility. Young (6) developed a risk model using a 20-SNP panel including metabolizing, inflammation, DNA repair, anti-oxidant, apoptosis and addiction genes. The final model included the 20 SNPs, age, history of COPD, family history of lung cancer and gender. When numeric scores were assigned to both the SNP and demographic data, and sequentially combined by a simple algorithm in a risk model, the composite score was found to be linearly related to lung cancer risk with a bimodal distribution. Spira et al (7) have advanced the notion that upper airway gene expression in smokers may serve as a relatively noninvasive surrogate marker of the physiologic response of the lung to tobacco smoke and could be used in large-scale screening and chemoprevention studies for lung cancer. Gene expression profiles in cytologically normal large airway epithelium were obtained via bronchoscopic brushings and were predictive of cancer status in a combined clinicogenomic model (P < 0.005). There was a significant improvement in performance of the clinicogenomic relative to the clinical model (P < 0.05). Use of the clinicogenomic model may reduce invasive diagnostic procedures for individuals without lung cancer. The ROC curve may not be sensitive to differences in probabilities between models, and therefore insufficient to assess the impact of adding a new predictor. A substantial gain in performance may not yield a substantial increase in AUC, and only a very large independent association of the new marker with risk will yield a meaningful larger AUC. As an example, the Framingham Risk Score that is widely applied has an AUC of about 0.80. New metrics need to be developed to compare nested models. In summary, the ability to accurately predict risk of lung cancer among former and current smokers has public health, clinical and financial implications for primary prevention, surveillance programs, screening programs and chemoprevention trials.

Background: Although cigarette smoking is the primary risk factor for lung cancer, other risk factors have been shown to be associated with lung cancer including prior history of respiratory diseases. The purpose of this analysis was to assess the association between prior history of respiratory disease (i.e., asthma, chronic obstructive pulmonary disease (COPD), hay fever, and pneumonia) and lung cancer among Mexican Americans using data from a multi-racial/ethnic lung cancer case-control study. Methods: The data were collected between 1991 and 2010. Cases (n = 217) were patients who presented with newly diagnosed, histopathologically confirmed and previously untreated lung cancer. Healthy control participants (n = 339) were recruited from a large multispecialty physician group practice. Demographics, cigarette use, and history of respiratory disease were collected by personal interview using a validated questionnaire. Multivariable logistic regression models were used to estimate relative risk. All analyses were restricted to self-report Mexican Americans. Results: Prior history of COPD (OR = 2.33; 95% CI = 1.14–4.79) and pneumonia (OR = 2.88; 95% CI = 1.47–5.63) were associated with a statistically significant increased risk of lung cancer. Conversely, self-reported history of asthma (OR = 0.71, 95% CI = 0.20–2.45) or hay fever (OR = 0.59, 95% CI = 0.29–1.21) were inversely associated with lung cancer risk but the odds ratios were not statistically significant. Conclusions: These findings illustrate that COPD and pneumonia are associated with an increased for lung cancer among Mexican Americans. Identifying risk factors such as C. pneumonia infections and COPD may induce targeted treatments to help attenuate progression to lung cancer. At present this study is one of largest case-control analyses assessing respiratory disease and lung cancer risk among Mexican-Americans. Citation Information: Cancer Prev Res 2011;4(10 Suppl):B91.

Dietary magnesium intake is significantly less in African Americans than Caucasians. Deficiency of magnesium influences immune-inflammatory response and induces oxidative stress that leads to increased cell proliferation and DNA damage. Low magnesium level also increases the physiologic effects of calcium. Therefore, we hypothesize that low dietary intake of magnesium and high calcium intake will increase the risk of developing prostate cancer among African American men. To test this hypothesis, we conducted a pilot case-control study to determine whether deficiency of dietary magnesium and a high calcium intake are associated with an increased risk of prostate cancer. This study included 209 cases with histologically confirmed adenocarcinoma of the prostate (110 African American and 99 Caucasian men). 199 population-based controls were frequency matched by age (±5 years) and race. Risk factor questionnaire and nutrition data were collected via telephone interview. Dietary Assessment was performed using Modified Block-NCI Food Frequency Questionnaire with 129 food items, questions about restaurant intake and vitamin mineral supplementation. Dietary intake was calculated using DIETSYS + Plus version 5.9 dietary analysis program. Nutrition data for 166 cases and 155 controls were available for this analysis. Mean magnesium and calcium intake did not vary significantly between cases and controls. The odds ratios for magnesium was 0.79 (95% CI 0.39–1.61) and for calcium 1.35 (95% CI 0.87–2.09). However, among African Americans, the odds ratio for magnesium was 0.28 (95% CI 0.08–1.05) and for calcium 1.9 (95% CI 1.05–3.40). These preliminary data suggest that high dietary magnesium intake is inversely associated with prostate cancer, while higher calcium intake is associated with increased risk among African American men. These data may have significant public health implications since the identification of African American men deficient of magnesium can be targeted for preventive interventions. Results of this study needs to be confirmed in future studies with larger sample size. Citation Information: Cancer Prev Res 2011;4(10 Suppl):A114.

Genetic factors play a role in determining individual susceptibility to cancer. Accumulation of multiple genetic alterations leads to cancer development and is cytogenetically detectable. Our group has applied the Cytonkinesis-Blocked Micronucleus assay (CBMN) as a sensitive predictor of individual susceptibility to lung cancer (LC). The objective of this study was to investigate whether self-reported exposure to ≥1 environmental agents (asbestos, fibers, paints, dust, exhaust, bleach, solvents) is associated with increased genetic instability and/or elevated risk of LC. We recruited 490 LC cases from an ongoing study at the University of Texas MD Anderson Cancer Center. Controls (n=500) were recruited from a large multispecialty physician group practice and frequency matched to cases by age (±5 years) and sex. Peripheral blood lymphocytes were collected and cytogenetic cultures prepared. One thousand binucleated cells per participant were scored and the frequencies of micronucleus (BN-MN), nucleoplasmic bridges (NPBs) and nuclear buds (NBUDs) were recorded. Linear regression analyses were applied to estimate the adjusted means of the CBMN. Odds ratios and 95% confidence intervals were calculated using logistic regression and to control for confounding. We did not detect elevated odds of LC when exposure was assessed by self-report to ≥1 of the seven environmental agents. When stratified by smoking status, we found heavy smokers with ≥1 exposure had a 58% increased risk of LC compared to never smokers without exposure; yet heavy smokers with no exposure had a 79% increased risk compared to the referent group. Even though LC cases showed significantly higher frequencies than controls for all three CBMN endpoints, there were no significant differences in cases with or without exposure (p=0.15). However in controls, the differences were indirectly related to exposure. Mean BN-MN for controls without exposure (mean +SE = 2.68±0.07) was higher than the mean BN-MN for controls with ≥1 exposure (mean +SE = 2.44±0.07). We also found an indirect relationship between mean BN-MN levels and smoking among controls: never smokers (mean +SE = 2.81±0.09), light (mean +SE=2.76±0.11, moderate (mean +SE=2.51±0.11) and heavy smokers (mean +SE=2.26±0.10; p for trend= These results confirm our previous studies that found the CBMN assay to be a valuable tool for predicting LC risk; however, the risk confirmed by this biomarker is not related to exposure to the seven agents under study. Our new findings that CBMN frequencies were higher in the controls that were never or light smokers support the hypothesis that heavy smokers may develop a more robust DNA repair machinery due to the constant exposure to smoking-related carcinogens. These findings warrant the need for further studies to identify the underlying mechanisms involved in genetic susceptibility to lung cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 1878. doi:10.1158/1538-7445.AM2011-1878

Background: Risk modeling has been shown to be important in detection and prevention of disease, and this is especially true for cancer research. One prominent lung cancer risk model is the Spitz Lung Cancer Risk Model, which was constructed from case-control data and incorporates clinical and epidemiologic variables (smoking status, family history and select co-morbidities). Although the Spitz model is validated using case-control data, the model has not been calibrated. Methods: We used data from 109 lung cancer patients and 4622 controls that were a subset of a large longitudinal cohort study, the Dutch-Belgian randomized lung cancer trial (NELSON). We estimated absolute lung cancer risks from the Spitz risk model to evaluate the model's calibration, discriminatory power and clinical utility. Results: With the 10 year absolute risk calculations, the Spitz model was well calibrated with NELSON data, with slightly weaker calibration for five year absolute risk calculations. The overall discriminatory power was 0.69 (95% CI = 0.64-0.74), with current smokers having higher discriminatory power (0.73, 95% CI = 0.67-0.79) compared to former smokers (0.61, 95% CI = 0.51-0.70). When examining clinical utility, it peaked at four lung cancer patients correctly identified by the model for every control incorrectly identified as a lung cancer patient. Conclusions: This is the first time the Spitz model has been evaluated for calibration, with very good results except for the lowest and highest risk individuals. The calibration results also confirmed the notion by Dr. Peter Bach's group that 10 year absolute risk models are better indicators of cancer risk than five year models. The clinical utility results are very similar to those calculated by the Spitz model with Harvard data that was published in the British Journal of Cancer recently. The Spitz model showed moderate discriminatory power, especially among current smokers. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 1899. doi:10.1158/1538-7445.AM2011-1899

Mexican American youth have the highest rates of smoking experimentation. We have previously confirmed several psychosocial factors including cognitive susceptibility to smoking, outcome expectations, risk taking tendencies, family influence and demographic characteristics, as risk factors for experimentation in these youth. In this study, we evaluated the incremental value of adding select variants of two candidate genes in the dopamine pathway implicated in the smoking phenotype. Participants were Mexican origin youth recruited from a large population-based cohort study in Houston, TX, who provided prospective data on cigarette experimentation yearly for five years. Psychosocial data on family influence, risk taking tendencies and outcome expectations were elicited at two time points-baseline and final follow-up. At baseline, all participants provided buccal samples, from which DNA was extracted and genotyped for 27 genetic variants in the dopamine receptor DRD2, and Dopamine Transporter (DAT) genes. Of the 1002 youth in the current analysis, 52% were girls, 66% were born in the US, average age at baseline was 11.8 (SD=0.8) years, 78% reported being highly acculturated, and 22% reported being new experimenters (i.e. first experimented with cigarettes only subsequent to the baseline survey). Two SNPs were associated with experimentation: rs10052016 (p Citation Information: Cancer Prev Res 2010;3(1 Suppl):A86.

Background: Since lung cancer only occurs in a small fraction of long-term smokers, there is a need to develop risk prediction models to identify high-risk subgroups. Three lung cancer models, constructed using clinical and epidemiological variables, predicted absolute risk of lung cancer: one based on a cohort of patients recruited for the CARET study, and two constructed from case-control studies conducted in Houston, Texas and Liverpool, England. Given their potential application to primary chemo-prevention strategies and screening trials, it is important to compare the accuracy of these three models in an independent population. Methods: We used data for 3197 lung cancer patients and 1703 cancer-free controls recruited to an ongoing case-control study of lung cancer at Harvard School of Public Health and Massachusetts General Hospital (Boston, MA). We estimated 5-year lung cancer risk for each risk model and compared the discriminatory power, as measured by the area under the the receiver-operator characteristic curve, accuracy, as measured by the positive predictive value and negative predictive value, and clinical utility of these models, as measured with scaled rectangles. Results: Overall, the discriminatory power for the Liverpool Lung Project (LLP) (AUC = 0.69, 95% CI = 0.67–0.71) and Spitz models (AUC = 0.69, 95%CI = 0.66–0.71) were comparable, while the Bach model had significantly lower power (AUC =0.66, 95% CI = 0.64–0.69; P=0.02). Positive predictive values were highest with the Spitz model (0.882) compared to 0.809 for the Bach model and 0.759 for the LLP model. In contrast, the negative predictive values were highest for the LLP model (0.560) compared to 0.450 for the Spitz model and 0.447 for the Bach. The Spitz and Bach models had lower sensitivity but higher specificity compared to the LLP model. For instance, 26.6% of all lung cancer cases have a five-year absolute risk of lung cancer ≥ 2.5% for the Spitz model compared to 66.7% of all cases for the LLP model. However, only 5.6% of all healthy controls have a five-year absolute risk of lung cancer ≥ 2.5% with the Spitz model compared to 33.4% of all controls for the LLP model. Conclusion: We observed modest differences in discriminatory among the three lung cancer risk models. The level of the discriminatory powers of these three lung cancer risk models was moderate at best, which highlights the difficulty in developing effective risk models. There is considerable room for improvement in model performance by incorporating additional risk factors, such as genetic risk factors, to increase discriminatory power and accuracy, while still maintaining clinical utility. Citation Information: Cancer Prev Res 2010;3(1 Suppl):A127.

Background: With increasing amounts of genes available for study, more opportunities are available to discover key associations between genetic factors and disease. In a previous study on Hodgkin disease, we observed powerful positive associations between genes on chromosomes 3 and 19. In this study, we expanded the previous analyses to include 62 single nucleotide polymorphisms (SNPs) in DNA repair and inflammation pathway genes located throughout the genome and evaluated possible associations for Hodgkin disease by inferring haplotypes at the individual gene level. Methods: Genetic and epidemiological data were obtained from a case-control study conducted at M.D. Anderson Cancer Center between 1987 and 1992 including 200 cases and 220 controls. Haplotypes were initially inferred with Haploview using linkage disequilibrium plots, and differences in haplotypes between cases and controls were determined with PHASE. Linear modeling with haplotypes, through both joint and separate effects haplotype modeling, with sex, race, smoking status, and age as confounders, were conducted with Haplo.stats. A joint-effects linear model allows each haplotype to be compared to the most frequent haplotype. A separate-effects linear model compares a haplotype to all of the other haplotypes possible for study. Results: With the joint effects haplotype model, some significant haplotype associations with Hodgkin disease were observed with DNA repair gene XPC on chromosome 3 and inflammation pathway gene ILR4 on Chromosome 16. For DNA repair gene XPC, susceptible associations with Hodgkin disease were discovered with the use of an additive genetic model with haplotype CT from SNPs rs2228001 (XPC 499 A>C) and rs2228000 (XPC 939 C>T) (OR = 1.49, 95% CI = 1.03–2.16) and a dominant genetic model with the same haplotype (OR = 1.75, 95% CI = 1.15–2.13). For inflammation pathway gene ILR4, protective associations with Hodgkin disease were discovered with the use of an additive genetic model with haplotype TC from SNPs rs1805012 (XPC 499 T>C) and rs1805015 (XPC 939 T>C) (OR = 0.63, 95% CI = 0.33–0.99) and a dominant genetic model with the same haplotype (OR = 0.48, 95% CI = 0.26–0.88). The separate effects haplotype model also highlighted associations between haplotypes and Hodgkin disease on these same genes. Other significant associations between haplotypes and Hodgkin disease were observed on chromosomes 5 with inflammation pathway gene IL4 and chromosome 19 with DNA repair gene. Conclusions: The basis for haplotype analysis to analyze the effects of linked SNPs associated with Hodgkin disease has been demonstrated. Furthermore, the interactions between specific SNPs in both the inflammation and the DNA repair pathway have been shown to play an important role in possible Hodgkin disease development. By targeting specific aspects of the genome that show either an increase or decrease in Hodgkin risk, identification of high risk individuals could be achieved. Citation Information: Cancer Prev Res 2010;3(1 Suppl):A7.