Virtual Library
Start Your Search
M.R. Spitz
Moderator of
-
+
O18 - Cancer Control and Epidemiology II (ID 133)
- Event: WCLC 2013
- Type: Oral Abstract Session
- Track: Prevention & Epidemiology
- Presentations: 8
- Moderators:M.R. Spitz, L. Irving
- Coordinates: 10/29/2013, 10:30 - 12:00, Bayside 103, Level 1
-
+
O18.01 - Multi-'omic analysis of an arsenic-associated lung squamous cell carcinoma reveals specific DNA level signatures (ID 283)
10:30 - 12:00 | Author(s): V.D. Martinez Zamora, K.L. Thu, E.A. Vucic, R. Hubaux, M.I. Adonis, L. Gil, C.E. Macaulay, S. Lam, W.L. Lam
- Abstract
- Presentation
Background
Chronic low-level exposure to arsenic is an emerging cancer risk factor in many parts of the world, including North America. The lung is one anatomical site prominently affected by the carcinogenic effects of arsenic, evident by the striking incidence of lung cancer in never smokers with chronic exposure. Histologically, arsenic related lung tumors are indistinguishable from those induced by other lung carcinogens, and molecularly, arsenic specific DNA copy-number, methylation and expression changes have been identified. Arsenic mediated carcinogenesis occurs through a combination of molecular mechanisms; however, high resolution, multi-'omic analyses of arsenic related tumors have been difficult due to the lack of fresh frozen samples required to obtain high quality DNA and RNA. In this study, we sought to characterize global changes in DNA sequence and methylation levels and their impacts on gene expression in a lung tumor from a patient with chronic arsenic exposure (As-LUSC).Methods
Tumor and non-malignant lung tissues were obtained from a never smoker with lung squamous cell carcinoma (LUSC) who had no family history of lung cancer and 50 years of chronic exposure to high levels of arsenic-contaminated drinking water. Whole genome sequencing was performed and the tumor's mutational signature was compared to those observed in 194 previously characterized NSCLC tumors from the cancer genome atlas (TCGA). DNA methylation was measured using high density methylation arrays and gene expression by RNA sequencing.Results
The As-LUSC exhibited alterations typical of LUSC, such as copy number gains at 3q26 (SOX2 locus) and expression of squamous markers including up-regulation of KRT6B, DSG3, MMP12, KRT5, and down-regulation of PDK4, which are consistent with LUSC histology. However, the As-LUSC harbored a low number of point mutations (only 49 non-synonymous mutations affecting coding DNA sequences) and had a remarkably high fraction of T>G/A>C and low fraction of C>A/G>T transversions, which are features uncharacteristic of LUSCs that suggest arsenic is associated with a distinct mutational spectrum. Furthermore, at the gene level, we identified a G>C mutation in TP53 which is rare in lung tumors (<0.2%) but has been observed in other arsenic-related malignancies. Clustering analysis using ~450,000 methylation probes revealed that the As-LUSC methylation profile was completely distinct from never smoker LUSCs from the TCGA. Of interest, the As-LUSC exhibited lower levels of methylation at CpG islands sores that are not associated with genes, although have been described to exhibit cell type specific methylation patterns.Conclusion
By applying whole genome sequencing, methylation and expression profiling of a LUSC from a never-smoker patient chronically exposed to arsenic, we identified a distinct mutational spectrum and methylation pattern in the As-LUSC. Our results support the concept that arsenic induces lung cancers through mechanisms different from tobacco smoke and other carcinogens. Further study of the mutational profiles of additional arsenic-related cancers is warranted and may yield valuable insight into arsenic associated tumourigenesis, leading to the development of novel diagnostic and therapeutic targets for environmental monitoring and treatment.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
- Abstract
- Presentation
Background
EGFR and ALK are important driver mutations in never smokers. While we reported the significant association of increased environmental tobacco smoke (ETS) with EGFR mutations in Japanese cohort (Kawaguchi, Clin Cancer Res, 2011), it has not been fully understood in other ethnicities and also the correlation of ETS with ALK has not been reported yet. In this study, we evaluated the association of ETS with the prevalence of EGFR mutations and ALK translocations in various ethnicities including East-Asia (Japan, Korea, China, and Singapore) and the USA.Methods
ETS exposure on never smokers with non-small cell lung cancer (NSCLC) was evaluated using the standardized questionnaire including exposure period, place, and duration. Cumulative dose of ETS (CETS) was defined as a sum of the number of the exposure years in childhood/ adulthood and at home/ workplace, and was treated as a continuous variable or quintile. EGFR mutations and ALK rearrangements were tested by PCR-based detection and fluorescence in situ hybridization, respectively. Multivariate analyses were done using the generalized linear mixed model (GLIMMIX procedure, SAS v9.3).Results
From March 2008 to December 2012, 498 never smokers with NSCLC were registered with the following patient characteristics: ethnicity (nationality) of Asian/ Caucasian/ others, 425 (Japanese 250, Korean 102, Chinese 46, others 2)/ 48/ 25; male/ female, 114/ 384; age <65/ >=65, 286/ 212; histology of adenocarcinoma/ BAC/ squamous cell carcinoma/ adenosquamous cell carcinoma/ other NSCLC, 459/ 12/ 13/ 5/ 9; frequency (%) of CETS < median CETS (40 years) in Japanese/ Korean/ Chinese/ Caucasian, 32.8/ 44.1/ 71.7/ 83.3. EGFR status was wild type 43.6%, exon 19 deletion 25.3%, L858R 21.5% and other mutations 9.6%. ALK status was wild type 52.0%, rearranged 10.6% and unknown 37.3%. Average CETS (years) of NS with EGFR (+), ALK (+) and wild type tumors were 45.4, 26.9 and 37.7, respectively. In multivariate generalized linear mixed model, incidence of activating EGFR mutations, not ALK rearrangements, was significantly associated with the increment of CETS in female, not in male gender. Odds ratios (OR) for EGFR mutations in female (n=384) were 1.084 (95% CI, 1.003-1.171; p=0.0422) for each increment of 10 years in CETS while OR in male (n=114) were not significant (OR 0.890; 95% CI, 0.725-1.093; p=0.2627). OR for ALK rearrangements in female (n=238) and those in male gender (n=74) were 0.930 (0.791-1.094; p=0.3814) and 0.854 (0.620-1.178; p=0.3319).Conclusion
Increased ETS exposure was closely associated with EGFR mutations in never smokers with female gender and NSCLC in the expanded multinational cohort. However, the association of ETS and ALK rearrangements in never smokers with NSCLC was not significant.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O18.03 - The BioCAST / IFCT-1002 study: a comprehensive overview of demographic, risk exposure and somatic mutations of non-small cell lung cancer occurring among French never smokers (ID 3293)
10:30 - 12:00 | Author(s): S. Couraud, P. Souquet, C. Paris, R. Gervais, H. Doubre, E. Pichon, A. Dixmier, I. Monnet, B. Etienne-Mastroianni, M. Vincent, J. Tredaniel, M. Perrichon, P. Foucher, B. Coudert, D. Moro-Sibilot, E. Dansin, S. Labonne, P. Missy, G. Zalcman
- Abstract
- Presentation
Background
Lung cancer occurring in never-smoker (LCINS) is a particular entity. Although the definition is strict (less than 100 cigarette in lifetime) never-smokers are frequently misclassified and no study gives a comprehensive analysis of this group, particularly in a European setting.Methods
All consecutive never-smoker patients diagnosed with a non-small cell lung cancer in one of the 75 participating centers throughout France, between November 2011 and January 2013, were included in this prospective survey. All patients underwent a detailed questionnaire supported by a trained staff during a phone interview. Somatic mutations and cancer clinical and histological data were also recorded from medical charts.Results
Overall, 384 never-smokers were included and 336 interviews were completed. Most of them were women (n=319, 83.1%). The mean age at diagnosis was 69.8 ± 12.02 and 10.9% were under 55 years-old. None reported alternative smoking (pipe, cigar, water-pipe, gum, or cannabis). Most of them originated from Western and Southern Europe (90.5%). Overall, 219 (65.6%) reported a passive smoking exposure in a domestic setting (n=198; 59.3%), and/or at workplace (n=60; 18.0%). Patients had a personal history of pneumonia in 6.2%, tuberculosis in 8.3%, COPD in 13.0%, and a cancer at another site in 16.6%. Eighty patients reported at least two relatives with lung cancer (24.0%). Definite occupational exposure was observed in 12.0% (n=44) for diesel, 7,1% (n=26) for asbestos, 3.3% (n=12) for poly-aromatic hydrocarbons, 2.4% (n=9) for silica, 0.8% (n=3) for chrome, and 0.5% (n=2) for painting. Exposure to cooking oil was noted in 123 patients (36.8%) with a mean of 49.4 ± 356.7 cooking-dish year. Moreover, 79.7% (n=259) patients were ever exposed to solid fuel fumes for cooking or heating (21.2% during more than 50% of their lifetime). Among women, 91.7% already reached menopause (mean age 49.3 ± 5.6 years-old), 115 (41.7%) were ever-exposed to oral contraceptive (mostly oestrogen-containing drugs), and 25.5% to post-menopause hormone replacement therapy (oral or transdermal). Most of lung cancers were adenocarcinoma (n=327, 85.2%) followed by squamous cell carcinoma (n=29, 7.6%) and large cell carcinoma (n=17; 4.4%). Among adenocarcinoma, 71% were invasive, 4% in-situ, 2% minimally-invasive, 2% variant of invasive, and 20.0% were NOS. Cancer stage was I in 9.2%, II in 5.8%, III in 11.8% and IV in 73.2%. At least one biomarker was tested in 359 patients (93.5%). We found 148 patients with EGFR mutations (43.5% out of the EGFR-tested patients), 20 with KRAS mutations (6.8%), 24 with ALK translocation (12.5%), 10 with BRAF mutation (4.5%), 8 with HER2 mutation (4.0%) and 4 with PIK3CA (2.1%). Overall, 27.0% samples remain wild type, 2.1% with multiple mutations, 71.0% with a single mutation, and 20.6% with missing data.Conclusion
We provide here the largest cohort of LCINS in a European setting with reliable data on tobacco intoxication, occupational exposure, and hormonal treatments, since collected by a trained staff through phone interview. In this perfectly clinically characterized cohort, molecular analyses showed that 72% of tumors exhibited oncogenic targetable mutations.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O18.04 - Impact of Passive Smoking on molecular pattern in Never Smokers with Non-Small Cell Lung Cancer: Findings from the BioCAST / IFCT-1002 Study (ID 3305)
10:30 - 12:00 | Author(s): S. Couraud, P. Dumont, L. Moreau, D. Debieuvre, J. Margery, E. Quoix, B. Duvert, L. Cellerin, N. Baize, B. Taviot, M. Coudurier, J. Cadranel, P. Missy, G. Zalcman, P. Souquet
- Abstract
- Presentation
Background
EGFR and HER2 mutations are usually associated with never-smokers while KRAS and BRAF mutations are thought to be link with smoking behavior in Non-Small Cell Lung Cancer (NSCLC). Passive smoking exposure is a well-known risk factor for lung cancer. Only EGFR and KRAS mutations were investigated in association with passive smoking and showed conflicting results. We aimed to investigate mutation rate of EGFR, HER2, KRAS, BRAF and ALK in a cohort of never smokers regarding their passive smoke exposure.Methods
The BioCAST / IFCT-1002 study is a prospective cohort of NSCLC patients diagnosed in French never-smokers patients (less than 100 cigarettes in lifetime) between November 2011 and January 2013, Passive smoking exposure was evaluated through standardized questionnaire. We obtained biomarkers mutation results through routine testing. We used Fisher, Chi-square, median test and Mann-Whitney U test for comparisons as appropriate. We used logistic regression to calculate adjusted odds ratio for risk of each mutations.Results
Out of the 384 patients included in the BioCAST database, 334 (87.0%) had available data on passive smoking exposure. Among them, 219 patients (65.6%) were ever exposed to passive smoking in their lifetime. 198 (59.3%) reported a domestic exposure (122 during childhood at least) and 60 (18.0%) a workplace exposure. Result of at least one biomarker mutation was available in 313 patients (93.7%). including 128 EGFR mutations in 297 patients, 8/174 HER2 mutations, 18/256 KRAS mutations, 10/196 BRAF mutations, and 20/171 ALK gene rearrangements. We found no difference in mutation rate according to passive smoke exposure (cf. Table 1). There was no difference when comparing cumulative year of exposure, smoker-year or passive-pack year (as continuous variable) to the mutation rate, for any biomarker. When considered as categorical variable – after division in quartiles – we found also no difference. Results were similar when focusing on domestic (childhood versus adulthood included) and workplace exposure only. Finally, we found no significant increased risk for mutation for any biomarker in logistic regression adjusted for most of other lung cancer risk factors.EGFR Mt (n=297) HER2 Mt (n=171) KRAS Mt (n=256) BRAF Mt (n=196) ALK Fusion (n=171) % % % % % Overall exposure Never 46.5 3.5 6.7 5.9 13.0 Ever 41.3 5.1 7.2 4.7 11.1 Domestic exposure Never 45.8 2.9 7.7 6.4 11.1 Ever 41.3 5.7 6.6 4.2 12.0 Exposure at workplace Never 43.3 5.5 7.0 5.4 12.2 Ever 42.3 0 7.0 3.6 8.3 Total 43.1 4.6 7.0 5.1 11.7 Exposure in childhood Ever 40.5 3.0 6.5 2.7 14.7 Only in adulthood 42.6 10.3 6.8 6.7 7.5 Conclusion
Although we report the largest and more comprehensive study focusing on this topic, we found no significant difference in the biomarker mutation profile of NSCLC occurring in French never-smokers regarding their exposure to passive smoking as compared with the pattern of mutations described never-smoker patients with any passive smoking.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O18.05 - DISCUSSANT (ID 3998)
10:30 - 12:00 | Author(s): P.P. Massion
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O18.06 - Vietnamese non-small cell lung cancer patients in California: molecular profiles and clinical characteristics (ID 1079)
10:30 - 12:00 | Author(s): K.H. Nguyen, M. Das, K. Ramchandran, J. Shrager, R.E. Merritt, C. Hoang, B. Burt, A. Tisch, J. Pagtama, J. Zehnder, G. Berry, H.A. Wakelee, A. Nguyen, J.W. Neal
- Abstract
- Presentation
Background
Lung cancer is the leading cause of cancer-related deaths worldwide with 1.3 million deaths per year. Discoveries of oncogenic mutations in non-small cell lung cancer (NSCLC) over the past decade have led to targeted therapies against epidermal growth factor receptor (EGFR) activating mutations, anaplastic lymphoma kinase (ALK) gene rearrangement, and repressor of silencing 1 (ROS1) gene rearrangement. The frequencies of these mutations and gene rearrangements have been elucidated in the Western and East Asian populations. However, the frequencies of these oncogenic alterations remain unknown in Vietnam, where lung cancer is one of the leading causes of cancer mortalities but molecular testing is not routinely performed due to limited resources. In this project, we aimed to analyze the Vietnamese patients treated at Stanford, California, with a future plan to compare with another cohort inside Vietnam.Methods
We collected molecular and clinical variables of NSCLC patients of Vietnamese origin, based on patients' self-reported ethnicity, language, or country of origin, treated at Stanford from 2009 to 2012. Comparison of the molecular and clinical characteristics of never smokers versus smokers was performed with Pearson's chi-squared test for nominal variables and Student's t test for continuous variables. Survival analyses were done using the Kaplan-Meier method and Cox proportional hazards modeling.Results
Forty-six patients of Vietnamese origin were seen at the Stanford thoracic oncology clinic from 2009 to 2012, including 22 men and 24 women with a mean age of 58 years. Twenty-seven (58.7%) were never-smokers. Forty-two (91.3%) of the tumors were adenocarcinoma. Ten patients (21.7%) presented at stage I, none at stage II, 8 patients (17.4%) at stage III, 28 patients (60.9%) at stage IV. Fifteen patients out of 28 tested for EGFR (53.6%) had an activating mutation; 14 of these 15 patients were never-smokers. Five patients out of 16 tested for ALK (31.3%) had ALK gene rearrangement. No ROS1 gene rearrangement out of 3 patients tested was found. Only one patient, a former smoker, out of 23 tested (4.4%) was found to have a KRAS mutation. Eighteen out of 27 never-smokers (66.7%) and 3 out of 19 smokers (15.8%) had a targetable driver mutation (EGFR, ALK, or ROS1). For all stages, the median overall survival (OS) for never-smokers was 22.3 months (95% confidence interval (CI); 11.9 months, 24.3 months) compared to 12.9 months (95% CI; 5.8 months, 20.0 months) for smokers. For only stage IV, the median OS for never-smokers was 21.2 months (95% CI; 13.0 months, 24.3 months) compared to 11.6 months (95% CI; 1.4 months, 30.9 months) for smokers.Conclusion
Approximately two-thirds of never-smoker patients of Vietnamese origin had NSCLC with a targetable driver mutation. OS differ markedly by smoking status. The high percentage of Vietnamese patients in California with driver mutations warrants further studies to evaluate the frequencies of NSCLC driver mutations inside Vietnam, strongly suggesting that nationwide implementation of routine molecular testing will have a positive impact on clinical management of Vietnamese patients with NSCLC.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
- Abstract
Background
Nickel is an essential trace metal used in the occupational setting and is naturally found in the general environment, resulting in both occupational and nonoccupational exposures to individuals at varying levels. Exposure to nickel has been associated with several toxicites and the International Agency for Research on Cancer has concluded that there is sufficient evidence in humans associating exposure to nickel or nickel compounds with risk of lung cancer. We evaluated overall and cause-specific mortality among Chinese workers involved in nickel production or utilization in order to examine the long-term health effects of occupational exposure to nickel compounds.Methods
The study design was a retrospective cohort mortality study including 432,526 workers who were involved with nickel mining or smelt between 2001 and 2010. We calculated standardized mortality ratios (SMR) using the death rates of Gansu Province in China, and estimated by the exact probabilities of the Poisson distribution.Results
Overall, the all-cause mortality was decreased in all workers compared to the general population of Gansu province (SMR= 0.53, 95%CI: 0.51-0.55). Analyses examining cause-specific mortality revealed an increase in the mortality from bronchogenic carcinoma and lung cancer (SMR = 2.05, 95% CI = 1.84-2.29), cor pulmonale (SMR =4.08, 95% CI = 3.25-5.01), and silicosis (SMR = 13.59, 95%CI =11.90-15.52) in the workers exposed to nickel.Conclusion
This study confirmed a significant excess of mortality from diseases of the lung including silicosis , lung cancer, and cor pulmonale among workers involved in nickel mining or smelt in China. -
+
O18.08 - DISCUSSANT (ID 3999)
10:30 - 12:00 | Author(s): C. Amos
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
Author of
-
+
MS07 - Epidemiology and Prevention (ID 24)
- Event: WCLC 2013
- Type: Mini Symposia
- Track: Prevention & Epidemiology
- Presentations: 1
- Moderators:K. Park, N. Van Zandwijk
- Coordinates: 10/28/2013, 14:00 - 15:30, Bayside Auditorium B, Level 1
-
+
MS07.4 - Risk Prediction Models (ID 489)
14:00 - 15:30 | Author(s): M.R. Spitz
- Abstract
- Presentation
Abstract
Results from the National Lung Screening Trial (NLST) showing a 20% reduction in lung cancer mortality in the screened arm have heightened awareness of the need for reliable risk prediction tools for estimating the probability of lung cancer. A key issue of uncertainty is which smokers should be targeted for low-dose computed tomography (LDCT) screening. The NLST used 55 - 74 years, ≥30 pack-years of smoking and up to 15 years since quitting as selection criteria. 7 million U.S. adults meet these entry criteria, and an estimated 94 million U.S. adults are current or former smokers. Validated risk prediction models could improve the outcomes of screening efforts. Such models have substantial public health implications and value in clinical decision making as well. Further, risk prediction tools could be incorporated into the design of smaller, more powerful, and “smarter” prevention trials. The first lung cancer risk prediction model was developed by Bach et al. using data from the Carotene and Retinol Efficacy Trial (CARET) of 14,000 heavy smokers and over 4,000 asbestos-exposed men. Variables included in the final model were age, gender, asbestos exposure, smoking history, cigarettes per day, duration of smoking and duration of cessation. Cronin et al. externally validated the Bach model in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study control arm (c statistic of 0.69). Spitz et al. expanded on this model adding epidemiologic and clinical data derived from an ongoing lung cancer case-control study. Their model included environmental tobacco smoke (for never and former smokers), family cancer history, asbestos and dust exposures, prior respiratory disease, history of hay fever, and smoking history variables. These variables have strong biologically plausibility and are relatively easy to ascertain through patient interview. However, the validated area under the curve (AUC) statistics for former and current smoker models were modest (0.63 and 0.58, respectively), although consistent with those from other risk prediction models. The LLP model based on data from the Liverpool Lung Project included age, sex and smoking, as well as family history of lung cancer, exposure to asbestos, prior diagnosis of pneumonia and of a malignancy other than lung cancer. Prior diagnoses of emphysema and lung cancer lost significance in the multivariate model. Young et al. developed a risk model using a 20-single nucleotide polymorphism (SNP) panel including cell-cycle control, oxidant response, apoptosis, and inflammation genes, as well as 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 linearly related to risk with a bimodal distribution. These data have not been well replicated. In 2011,Tammemagi published a carefully constructed risk prediction model based on data from 70,962 control subjects in the Prostate, Lung, Colorectal, Ovarian cancer screening trial (PLCO). Model 1 included age, education, body mass index (BMI), family history of lung cancer, chronic obstructive pulmonary disease (COPD), recent chest x-ray, smoking status (never, former, current), pack-years smoked, and smoking duration. Model 2 also included time in years since ever-smokers permanently quit smoking. In external validation, performed with 44,223 PLCO intervention arm participants, Models 1 and 2 had area under the curves of 0.84 and 0.78, respectively. Tammemagi and colleagues also showed that their risk prediction model for lung cancer incidence was a more sensitive indicator of pre-screening risk of developing lung cancer than were NLST eligibility criteria. Kovalchik et al. subsequently showed that 88% of LDCT-prevented lung cancer deaths occurred among the 60% of NLST participants with highest pre-screening risk, while just 1% occurred among the 20% at lowest risk. This finding reinforces the role for risk-based screening. Maisonneuve et al. incorporated lung nodule characteristics and CT diagnosed emphysema into the Bach model. Presence of nonsolid nodules (RR = 10.1), nodule size > 8 mm (RR = 9.89), and emphysema (RR = 2.36) at baseline CT were all significant predictors of subsequent lung cancers. Incorporation of these variables into the Bach model increased the predictive value of the model (c-index = 0.759). Hoggart et al used prospective data from the European EPIC cohort. Using smoking information alone gave good predictive accuracy: the AUC and 95% CI in ever smokers was 0.843 (0.810-0.875). Adding other risk factors (10 occupational/environmental exposures previously implicated with lung cancer, and SNPs at two loci identified by GWAS of lung cancer) had a negligible effect on the AUC. An extended model was constructed incorporating two markers of DNA repair capacity that have been shown in case-control analyses to be associated with increased lung cancer risk. Addition of the biomarker assays improved the sensitivity of the models over epidemiologic and clinical data alone. These in vitro lymphocyte culture assays, however, are time-consuming and require technical expertise, and are not applicable for widespread population-based implementation. Spitz et al. added 3 SNPS that were most significant in their GWAS data – rs1051730 from 15q25 and two SNPs from the 5p15.33 locus (rs2736100 and rs401681 that were not in strong LD) to the baseline model. The AUC for the baseline epidemiologic/clinical model including 1016 cases and 1111 controls (all ever smokers) was 0.59. There was evidence of a gene dosage effect with an odds ratio over threefold elevated in the highest genetic risk score (GRS) stratum. With addition of the GRS to the model, the AUC showed modest improvement, to 0.61, although this was significantly improved over the baseline model, (P< 0.001). Current lung cancer risk prediction models are hampered by a restricted number of potential predictors, generally low overall predictive performance, and methodological limitations. To date, one can argue that the Tammemagi 2013 model exhibits the highest AUC among all the prediction models. It is important to conduct additional external validations of all models in diverse populations.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
