A Novel Pathway-Based Approach Improves Lung Cancer Risk Prediction Using Germline Genetic Variations

The HUNT lung-SNP model: genetic variants plus clinical variables improve lung cancer risk assessment over clinical models

PURPOSE

The HUNT Lung Cancer Model (HUNT LCM) predicts individualized 6-year lung cancer (LC) risk among individuals who ever smoked cigarettes with high precision based on eight clinical variables. Can the performance be improved by adding genetic information?

METHODS

A polygenic model was developed in the prospective Norwegian HUNT2 study with clinical and genotype data of individuals who ever smoked cigarettes (n = 30749, median follow up 15.26 years) where 160 LC were diagnosed within six years. It included the variables of the original HUNT LCM plus 22 single nucleotide polymorphisms (SNPs) highly associated with LC. External validation was performed in the prospective Norwegian Tromsø Study (n = 2663).

RESULTS

The novel HUNT Lung-SNP model significantly improved risk ranking of individuals over the HUNT LCM in both HUNT2 (p < 0.001) and Tromsø (p < 0.05) cohorts. Furthermore, detection rate (number of participants selected to detect one LC case) was significantly better for the HUNT Lung-SNP vs. HUNT LCM in both cohorts (42 vs. 48, p = 0.003 and 11 vs. 14, p = 0.025, respectively) as well as versus the NLST, NELSON and 2021 USPSTF criteria. The area under the receiver operating characteristic curve (AUC) was higher for the HUNT Lung-SNP in both cohorts, but significant only in HUNT2 (AUC 0.875 vs. 0.844, p < 0.001). However, the integrated discrimination improvement index (IDI) indicates a significant improvement of LC risk stratification by the HUNT Lung-SNP in both cohorts (IDI 0.019, p < 0.001 (HUNT2) and 0.013, p < 0.001 (Tromsø)).

CONCLUSION

The HUNT Lung-SNP model could have a clinical impact on LC screening and has the potential to replace the HUNT LCM as well as the NLST, NELSON and 2021 USPSTF criteria in a screening setting. However, the model should be further validated in other populations and evaluated in a prospective trial setting.

Polygenic inheritance and its interplay with smoking history in predicting lung cancer diagnosis: a French-Canadian case-control cohort

BACKGROUND

The most near-term clinical application of genome-wide association studies in lung cancer is a polygenic risk score (PRS).

METHODS

A case-control dataset was generated consisting of 4002 lung cancer cases from the LORD project and 20,010 ethnically matched controls from CARTaGENE. A genome-wide PRS including >1.1 million genetic variants was derived and validated in UK Biobank (n = 5419 lung cancer cases). The predictive ability and diagnostic discrimination performance of the PRS was tested in LORD/CARTaGENE and benchmarked against previous PRSs from the literature. Stratified analyses were performed by smoking status and genetic risk groups defined as low (<20th percentile), intermediate (20-80th percentile) and high (>80th percentile) PRS.

FINDINGS

The phenotypic variance explained and the effect size of the genome-wide PRS numerically outperformed previous PRSs. Individuals with high genetic risk had a 2-fold odds of lung cancer compared to low genetic risk. The PRS was an independent predictor of lung cancer beyond conventional clinical risk factors, but its diagnostic discrimination performance was incremental in an integrated risk model. Smoking increased the odds of lung cancer by 7.7-fold in low genetic risk and by 11.3-fold in high genetic risk. Smoking with high genetic risk was associated with a 17-fold increase in the odds of lung cancer compared to individuals who never smoked and with low genetic risk.

INTERPRETATION

Individuals at low genetic risk are not protected against the smoking-related risk of lung cancer. The joint multiplicative effect of PRS and smoking increases the odds of lung cancer by nearly 20-fold.

FUNDING

This work was supported by the CQDM and the IUCPQ Foundation owing to a generous donation from Mr. Normand Lord.

Interaction between Continuous Pack-Years Smoked and Polygenic Risk Score on Lung Cancer Risk: Prospective Results from the Framingham Heart Study

BACKGROUND

Lung cancer risk attributable to smoking is dose dependent, yet few studies examining a polygenic risk score (PRS) by smoking interaction have included comprehensive lifetime pack-years smoked.

METHODS

We analyzed data from participants of European ancestry in the Framingham Heart Study Original (n = 454) and Offspring (n = 2,470) cohorts enrolled in 1954 and 1971, respectively, and followed through 2018. We built a PRS for lung cancer using participant genotyping data and genome-wide association study summary statistics from a recent study in the OncoArray Consortium. We used Cox proportional hazards regression models to assess risk and the interaction between pack-years smoked and genetic risk for lung cancer adjusting for European ancestry, age, sex, and education.

RESULTS

We observed a significant submultiplicative interaction between pack-years and PRS on lung cancer risk (P = 0.09). Thus, the relative risk associated with each additional 10 pack-years smoked decreased with increasing genetic risk (HR = 1.56 at one SD below mean PRS, HR = 1.48 at mean PRS, and HR = 1.40 at one SD above mean PRS). Similarly, lung cancer risk per SD increase in the PRS was highest among those who had never smoked (HR = 1.55) and decreased with heavier smoking (HR = 1.32 at 30 pack-years).

CONCLUSIONS

These results suggest the presence of a submultiplicative interaction between pack-years and genetics on lung cancer risk, consistent with recent findings. Both smoking and genetics were significantly associated with lung cancer risk.

IMPACT

These results underscore the contributions of genetics and smoking on lung cancer risk and highlight the negative impact of continued smoking regardless of genetic risk.

Functional studies of lung cancer GWAS beyond association

Fourteen years after the first genome-wide association study (GWAS) of lung cancer was published, approximately 45 genomic loci have now been significantly associated with lung cancer risk. While functional characterization was performed for several of these loci, a comprehensive summary of the current molecular understanding of lung cancer risk has been lacking. Further, many novel computational and experimental tools now became available to accelerate the functional assessment of disease-associated variants, moving beyond locus-by-locus approaches. In this review, we first highlight the heterogeneity of lung cancer GWAS findings across histological subtypes, ancestries and smoking status, which poses unique challenges to follow-up studies. We then summarize the published lung cancer post-GWAS studies for each risk-associated locus to assess the current understanding of biological mechanisms beyond the initial statistical association. We further summarize strategies for GWAS functional follow-up studies considering cutting-edge functional genomics tools and providing a catalog of available resources relevant to lung cancer. Overall, we aim to highlight the importance of integrating computational and experimental approaches to draw biological insights from the lung cancer GWAS results beyond association.

Prediction of lung cancer risk in Chinese population with genetic-environment factor using extreme gradient boosting

BACKGROUND

Detecting early-stage lung cancer is critical to reduce the lung cancer mortality rate; however, existing models based on germline variants perform poorly, and new models are needed. This study aimed to use extreme gradient boosting to develop a predictive model for the early diagnosis of lung cancer in a multicenter case-control study.

MATERIALS AND METHODS

A total of 974 cases and 1005 controls in Shanghai and Taizhou were recruited, and 61 single nucleotide polymorphisms (SNPs) were genotyped. Multivariate logistic regression was used to calculate the association between signal SNPs and lung cancer risk. Logistic regression (LR) and extreme gradient boosting (XGBoost) algorithms, a large-scale machine learning algorithm, were adopted to build the lung cancer risk model. In both models, 10-fold cross-validation was performed, and model predictive performance was evaluated by the area under the curve (AUC).

RESULTS

After FDR adjustment, TYMS rs3819102 and BAG6 rs1077393 were significantly associated with lung cancer risk (p < 0.05). For lung cancer risk prediction, the model predicted only with epidemiology attained an AUC of 0.703 for LR and 0.744 for XGBoost. Compared with the LR model predicted only with epidemiology, further adding SNPs and applying XGBoost increased the AUC to 0.759 (p < 0.001) in the XGBoost model. BAG6 rs1077393 was the most important predictor among all SNPs in the lung cancer prediction XGBoost model, followed by TERT rs2735845 and CAMKK1 rs7214723. Further stratification in lung adenocarcinoma (ADC) showed a significantly elevated performance from 0.639 to 0.699 (p = 0.009) when applying XGBoost and adding SNPs to the model, while the best model for lung squamous cell carcinoma (SCC) prediction was the LR model predicted with epidemiology and SNPs (AUC = 0.833), compared with the XGBoost model (AUC = 0.816).

CONCLUSION

Our lung cancer risk prediction models in the Chinese population have a strong predictive ability, especially for SCC. Adding SNPs and applying the XGBoost algorithm to the epidemiologic-based logistic regression risk prediction model significantly improves model performance.

Association of smoking and polygenic risk with the incidence of lung cancer: a prospective cohort study

Background

Genetic variation increases the risk of lung cancer, but the extent to which smoking amplifies this effect remains unknown. Therefore, we aimed to investigate the risk of lung cancer in people with different genetic risks and smoking habits.

Methods

This prospective cohort study included 345,794 European ancestry participants from the UK Biobank and followed up for 7.2 [6.5–7.8] years.

Results

Overall, 26.2% of the participants were former smokers, and 9.8% were current smokers. During follow-up, 1687 (0.49%) participants developed lung cancer. High genetic risk and smoking were independently associated with an increased risk of incident lung cancer. Compared with never-smokers, HR per standard deviation of the PRS increase was 1.16 (95% CI, 1.11–1.22), and HR of heavy smokers (≥40 pack-years) was 17.89 (95% CI, 15.31–20.91). There were no significant interactions between the PRS and the smoking status or pack-years. Population-attributable fraction analysis showed that smoking cessation might prevent 76.4% of new lung cancers.

Conclusions

Both high genetic risk and smoking were independently associated with higher lung cancer risk, but the increased risk of smoking was much more significant than heredity. The combination of traditional risk factors and additional PRS provides realistic application prospects for precise prevention.

Targeting lung cancer screening to individuals at greatest risk: the role of genetic factors

Lung cancer (LC) is the most common global cancer. An individual’s risk of developing LC is mediated by an array of factors, including family history of the disease. Considerable research into genetic risk factors for LC has taken place in recent years, with both low-penetrance and high-penetrance variants implicated in increasing or decreasing a person’s risk of the disease. LC is the leading cause of cancer death worldwide; poor survival is driven by late onset of non-specific symptoms, resulting in late-stage diagnoses. Evidence for the efficacy of screening in detecting cancer earlier, thereby reducing lung-cancer specific mortality, is now well established. To ensure the cost-effectiveness of a screening programme and to limit the potential harms to participants, a risk threshold for screening eligibility is required. Risk prediction models (RPMs), which provide an individual’s personal risk of LC over a particular period based on a large number of risk factors, may improve the selection of high-risk individuals for LC screening when compared with generalised eligibility criteria that only consider smoking history and age. No currently used RPM integrates genetic risk factors into its calculation of risk. This review provides an overview of the evidence for LC screening, screening related harms and the use of RPMs in screening cohort selection. It gives a synopsis of the known genetic risk factors for lung cancer and discusses the evidence for including them in RPMs, focusing in particular on the use of polygenic risk scores to increase the accuracy of targeted lung cancer screening.

Individual and joint contributions of genetic and methylation risk scores for enhancing lung cancer risk stratification: data from a population-based cohort in Germany

BACKGROUND

Risk stratification for lung cancer (LC) screening is so far mostly based on smoking history. This study aimed to assess if and to what extent such risk stratification could be enhanced by additional consideration of genetic risk scores (GRSs) and epigenetic risk scores defined by DNA methylation.

METHODS

We conducted a nested case-control study of 143 incident LC cases and 1460 LC-free controls within a prospective cohort of 9949 participants aged 50-75 years with 14-year follow-up. Lifetime smoking history was obtained in detail at recruitment. We built a GRS based on 31 previously identified LC-associated single-nucleotide polymorphisms (SNPs) and a DNA methylation score (MRS) based on methylation of 151 previously identified smoking-associated cytosine-phosphate-guanine (CpG) loci. We evaluated associations of GRS and MRS with LC incidence by logistic regression models, controlling for age, sex, smoking status, and pack-years. We compared the predictive performance of models based on pack-years alone with models additionally including GRS and/or MRS using the area under the receiver operating characteristic curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI).

RESULTS

GRS and MRS showed moderate and strong associations with LC risk even after comprehensive adjustment for smoking history (adjusted odds ratio [95% CI] comparing highest with lowest quartile 1.93 [1.05-3.71] and 5.64 [2.13-17.03], respectively). Similar associations were also observed within the risk groups of ever and heavy smokers. Addition of GRS and MRS furthermore strongly enhanced LC prediction beyond prediction by pack-years (increase of optimism-corrected AUC among heavy smokers from 0.605 to 0.654, NRI 26.7%, p = 0.0106, IDI 3.35%, p = 0.0036), the increase being mostly attributable to the inclusion of MRS.

CONCLUSIONS

Consideration of MRS, by itself or in combination with GRS, may strongly enhance LC risk stratification.

SNP eQTL status and eQTL density in the adjacent region of the SNP are associated with its statistical significance in GWA studies

Background

Over the relatively short history of Genome Wide Association Studies (GWASs), hundreds of GWASs have been published and thousands of disease risk-associated SNPs have been identified. Summary statistics from the conducted GWASs are often available and can be used to identify SNP features associated with the level of GWAS statistical significance. Those features could be used to select SNPs from gray zones (SNPs that are nominally significant but do not reach the genome-wide level of significance) for targeted analyses.

Methods

We used summary statistics from recently published breast and lung cancer and scleroderma GWASs to explore the association between the level of the GWAS statistical significance and the expression quantitative trait loci (eQTL) status of the SNP. Data from the Genotype-Tissue Expression Project (GTEx) were used to identify eQTL SNPs.

Results

We found that SNPs reported as eQTLs were more significant in GWAS (higher -log₁₀p) regardless of the tissue specificity of the eQTL. Pan-tissue eQTLs (those reported as eQTLs in multiple tissues) tended to be more significant in the GWAS compared to those reported as eQTL in only one tissue type. eQTL density in the ±5 kb adjacent region of a given SNP was also positively associated with the level of GWAS statistical significance regardless of the eQTL status of the SNP. We found that SNPs located in the regions of high eQTL density were more likely to be located in regulatory elements (transcription factor or miRNA binding sites).

When SNPs were stratified by the level of statistical significance, the proportion of eQTLs was positively associated with the mean level of statistical significance in the group. The association curve reaches a plateau around -log₁₀p ≈ 5. The observed associations suggest that quasi-significant SNPs (10^− 5 < p < 5 × 10^− 8) and SNPs at the genome wide level of statistical significance (p < 5 × 10^− 8) may have a similar proportions of risk associated SNPs.

Conclusions

The results of this study indicate that the SNP’s eQTL status, as well as eQTL density in the adjacent region are positively associated with the level of statistical significance of the SNP in GWAS.

Identification of risk loci and a polygenic risk score for lung cancer: a large-scale prospective cohort study in Chinese populations

BACKGROUND

Genetic variation has an important role in the development of non-small-cell lung cancer (NSCLC). However, genetic factors for lung cancer have not been fully identified, especially in Chinese populations, which limits the use of existing polygenic risk scores (PRS) to identify subpopulations at high risk of lung cancer for prevention. We therefore aimed to identify novel loci associated with NSCLC risk, and generate a PRS and evaluate its utility and effectiveness in the prediction of lung cancer risk in Chinese populations.

METHODS

To systematically identify genetic variants for NSCLC risk, we newly genotyped 19 546 samples from Chinese NSCLC cases and controls from the Nanjing Medical University Global Screening Array Project and did a meta-analysis of genome-wide association studies (GWASs) of 27 120 individuals with NSCLC and 27 355 without NSCLC (13 327 cases and 13 328 controls of Chinese descent as well as 13 793 cases and 14 027 controls of European descent). We then built a PRS for Chinese populations from all reported single-nucleotide polymorphisms that have been reported to be associated with lung cancer risk at genome-wide significance level. We evaluated the utility and effectiveness of the generated PRS in predicting subpopulations at high-risk of lung cancer in an independent prospective cohort of 95 408 individuals from the China Kadoorie Biobank (CKB) with more than 10 years' follow-up.

FINDINGS

We identified 19 susceptibility loci to be significantly associated with NSCLC risk at p≤5·0 × 10-8, including six novel loci. When applied to the CKB cohort, the PRS of the risk loci successfully predicted lung cancer incident cases in a dose-response manner in participants at a high genetic risk (top 10%) than those at a low genetic risk (bottom 10%; adjusted hazard ratio 1·96, 95% CI 1·53-2·51; ptrend=2·02 × 10-9). Specially, we observed consistently separated curves of lung cancer events in individuals at low, intermediate, and high genetic risk, respectively, and PRS was an independent effective risk stratification indicator beyond age and smoking pack-years.

INTERPRETATION

We have shown for the first time that GWAS-derived PRS can be effectively used in discriminating subpopulations at high risk of lung cancer, who might benefit from a practically feasible PRS-based lung cancer screening programme for precision prevention in Chinese populations.

FUNDING

National Natural Science Foundation of China, the Priority Academic Program for the Development of Jiangsu Higher Education Institutions, National Key R&D Program of China, Science Foundation for Distinguished Young Scholars of Jiangsu, and China's Thousand Talents Program.

Using germline variants to estimate glioma and subtype risks

BACKGROUND

Twenty-five single nucleotide polymorphisms (SNPs) are associated with adult diffuse glioma risk. We hypothesized that the inclusion of these 25 SNPs with age at diagnosis and sex could estimate risk of glioma as well as identify glioma subtypes.

METHODS

Case-control design and multinomial logistic regression were used to develop models to estimate the risk of glioma development while accounting for histologic and molecular subtypes. Case-case design and logistic regression were used to develop models to predict isocitrate dehydrogenase (IDH) mutation status. A total of 1273 glioma cases and 443 controls from Mayo Clinic were used in the discovery set, and 852 glioma cases and 231 controls from UCSF were used in the validation set. All samples were genotyped using a custom Illumina OncoArray.

RESULTS

Patients in the highest 5% of the risk score had more than a 14-fold increase in relative risk of developing an IDH mutant glioma. Large differences in lifetime absolute risk were observed at the extremes of the risk score percentile. For both IDH mutant 1p/19q non-codeleted glioma and IDH mutant 1p/19q codeleted glioma, the lifetime risk increased from almost null to 2.3% and almost null to 1.7%, respectively. The SNP-based model that predicted IDH mutation status had a validation concordance index of 0.85.

CONCLUSIONS

These results suggest that germline genotyping can provide new tools for the initial management of newly discovered brain lesions. Given the low lifetime risk of glioma, risk scores will not be useful for population screening; however, they may be useful in certain clinically defined high-risk groups.

Genetic variant of IRAK2 in the toll-like receptor signaling pathway and survival of non-small cell lung cancer

The toll-like receptor (TLR) signaling pathway plays an important role in the innate immune responses and antigen-specific acquired immunity. Aberrant activation of the TLR pathway has a significant impact on carcinogenesis or tumor progression. Therefore, we hypothesize that genetic variants in the TLR signaling pathway genes are associated with overall survival (OS) of patients with non-small cell lung cancer (NSCLC). To test this hypothesis, we first performed Cox proportional hazards regression analysis to evaluate associations between genetic variants of 165 TLR signaling pathway genes and NSCLC OS using the genome-wide association study (GWAS) dataset from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO). The results were further validated by the Harvard Lung Cancer Susceptibility GWAS dataset. Specifically, we identified IRAK2 rs779901 C > T as a predictor of NSCLC OS, with a variant-allele (T) attributed hazards ratio (HR) of 0.78 [95% confidence interval (CI) = 0.67-0.91, P = 0.001] in the PLCO dataset, 0.84 (0.72-0.98, 0.031) in the Harvard dataset, and 0.81 (0.73-0.90, 1.08x10-4 ) in the meta-analysis of these two GWAS datasets. In addition, the T allele was significantly associated with an increased mRNA expression level of IRAK2. Our findings suggest that IRAK2 rs779901 C > T may be a promising prognostic biomarker for NSCLC OS.

A Decade of GWAS Results in Lung Cancer

Genome-wide association studies (GWAS) were successful to identify genetic factors robustly associated with lung cancer. This review aims to synthesize the literature in this field and accelerate the translation of GWAS discoveries into results that are closer to clinical applications. A chronologic presentation of published GWAS on lung cancer susceptibility, survival, and response to treatment is presented. The most important results are tabulated to provide a concise overview in one read. GWAS have reported 45 lung cancer susceptibility loci with varying strength of evidence and highlighted suspected causal genes at each locus. Some genetic risk loci have been refined to more homogeneous subgroups of lung cancer patients in terms of histologic subtypes, smoking status, gender, and ethnicity. Overall, these discoveries are an important step for future development of new therapeutic targets and biomarkers to personalize and improve the quality of care for patients. GWAS results are on the edge of offering new tools for targeted screening in high-risk individuals, but more research is needed if GWAS are to pay off the investment. Complementary genomic datasets and functional studies are needed to refine the underlying molecular mechanisms of lung cancer preliminarily revealed by GWAS and reach results that are medically actionable. Cancer Epidemiol Biomarkers Prev; 27(4); 363-79. ©2018 AACRSee all articles in this CEBP Focus section, "Genome-Wide Association Studies in Cancer."

Genetics and epigenetics of NAFLD and NASH: Clinical impact

Non-alcoholic fatty liver disease (NAFLD) is now recognised as the most common liver disease worldwide. It encompasses a broad spectrum of conditions, from simple steatosis, through non-alcoholic steatohepatitis, to fibrosis and ultimately cirrhosis and hepatocellular carcinoma. A hallmark of NAFLD is the substantial inter-patient variation in disease progression. NAFLD is considered a complex disease trait such that interactions between the environment and a susceptible polygenic host background determine disease phenotype and influence progression. Recent years have witnessed multiple genome-wide association and large candidate gene studies, which have enriched our understanding of the genetic basis of NAFLD. Notably, the I148M PNPLA3 variant has been identified as the major common genetic determinant of NAFLD. Variants with moderate effect size in TM6SF2, MBOAT7 and GCKR have also been shown to have a significant contribution. The premise for this review is to discuss the status of research into important genetic and epigenetic modifiers of NAFLD progression. The potential to translate the accumulating wealth of genetic data into the design of novel therapeutics and the clinical implementation of diagnostic/prognostic biomarkers will be explored. Finally, personalised medicine and the opportunities for future research and challenges in the immediate post genetics era will be illustrated and discussed.

High-order gene interactions between the genetic polymorphisms in Wnt and AhR pathway in modulating lung cancer susceptibility

AIM

Genetic variations present within Wnt and AhR pathway might be related to the lung cancer susceptibility.

METHODS

A total of 555 subjects were genotyped using PCR-RFLP technique for polymorphic sites in DKK4, DKK3, DKK2, sFRP3, sFRP4, Axin2 and AhR. Multifactor dimensionality reduction method and classification and regression tree analysis was used.

RESULTS

Overall sFRP4_rs1802073 which has a cross validation consistency of 10/10, prediction error = 0.43 (p > 0.0001) is the best factor model. The second best model was sFRP4_rs1802073 and DKK2_rs419558 with cross validation consistency of 9/10 and prediction error = 0.40. In classification and regression tree analysis, DKK2 _rs419558 came out to be a significant factor; DKK2_rs17037102 (M)/DKK2_rs419558 (M) showed a tenfold risk of acquiring lung cancer, p = 0.0001. DKK2_rs17037102 (M)/AhR_rs2066853 (W)/AhR_rs10250822 (M) showed an 11-fold risk of developing lung cancer, p = 0.00001.

CONCLUSION

Both DKK2 and sFRP4 polymorphisms are found to play a crucial role; especially for smokers towards modulating risk for lung cancer. AhR variants are contributing maximally toward lung cancer risk.

Applying Risk Prediction Models to Optimize Lung Cancer Screening: Current Knowledge, Challenges, and Future Directions

Purpose of review

Risk prediction models may be useful for facilitating effective and high-quality decision-making at critical steps in the lung cancer screening process. This review provides a current overview of published lung cancer risk prediction models and their applications to lung cancer screening and highlights both challenges and strategies for improving their predictive performance and use in clinical practice.

Recent findings

Since the 2011 publication of the National Lung Screening Trial results, numerous prediction models have been proposed to estimate the probability of developing or dying from lung cancer or the probability that a pulmonary nodule is malignant. Respective models appear to exhibit high discriminatory accuracy in identifying individuals at highest risk of lung cancer or differentiating malignant from benign pulmonary nodules. However, validation and critical comparison of the performance of these models in independent populations are limited. Little is also known about the extent to which risk prediction models are being applied in clinical practice and influencing decision-making processes and outcomes related to lung cancer screening.

Summary

Current evidence is insufficient to determine which lung cancer risk prediction models are most clinically useful and how to best implement their use to optimize screening effectiveness and quality. To address these knowledge gaps, future research should be directed toward validating and enhancing existing risk prediction models for lung cancer and evaluating the application of model-based risk calculators and its corresponding impact on screening processes and outcomes.

A lung cancer risk classifier comprising genome maintenance genes measured in normal bronchial epithelial cells

Background

Annual low dose CT (LDCT) screening of individuals at high demographic risk reduces lung cancer mortality by more than 20%. However, subjects selected for screening based on demographic criteria typically have less than a 10% lifetime risk for lung cancer. Thus, there is need for a biomarker that better stratifies subjects for LDCT screening. Toward this goal, we previously reported a lung cancer risk test (LCRT) biomarker comprising 14 genome-maintenance (GM) pathway genes measured in normal bronchial epithelial cells (NBEC) that accurately classified cancer (CA) from non-cancer (NC) subjects. The primary goal of the studies reported here was to optimize the LCRT biomarker for high specificity and ease of clinical implementation.

Methods

Targeted competitive multiplex PCR amplicon libraries were prepared for next generation sequencing (NGS) analysis of transcript abundance at 68 sites among 33 GM target genes in NBEC specimens collected from a retrospective cohort of 120 subjects, including 61 CA cases and 59 NC controls. Genes were selected for analysis based on contribution to the previously reported LCRT biomarker and/or prior evidence for association with lung cancer risk. Linear discriminant analysis was used to identify the most accurate classifier suitable to stratify subjects for screening.

Results

After cross-validation, a model comprising expression values from 12 genes (CDKN1A, E2F1, ERCC1, ERCC4, ERCC5, GPX1, GSTP1, KEAP1, RB1, TP53, TP63, and XRCC1) and demographic factors age, gender, and pack-years smoking, had Receiver Operator Characteristic area under the curve (ROC AUC) of 0.975 (95% CI: 0.96–0.99). The overall classification accuracy was 93% (95% CI 88%–98%) with sensitivity 93.1%, specificity 92.9%, positive predictive value 93.1% and negative predictive value 93%. The ROC AUC for this classifier was significantly better (p < 0.0001) than the best model comprising demographic features alone.

Conclusions

The LCRT biomarker reported here displayed high accuracy and ease of implementation on a high throughput, quality-controlled targeted NGS platform. As such, it is optimized for clinical validation in specimens from the ongoing LCRT blinded prospective cohort study. Following validation, the biomarker is expected to have clinical utility by better stratifying subjects for annual lung cancer screening compared to current demographic criteria alone.

Electronic supplementary material

The online version of this article (doi:10.1186/s12885-017-3287-4) contains supplementary material, which is available to authorized users.

Familial Lung Cancer: A Brief History from the Earliest Work to the Most Recent Studies

Lung cancer is the deadliest cancer in the United States, killing roughly one of four cancer patients in 2016. While it is well-established that lung cancer is caused primarily by environmental effects (particularly tobacco smoking), there is evidence for genetic susceptibility. Lung cancer has been shown to aggregate in families, and segregation analyses have hypothesized a major susceptibility locus for the disease. Genetic association studies have provided strong evidence for common risk variants of small-to-moderate effect. Rare and highly penetrant alleles have been identified by linkage studies, including on 6q23-25. Though not common, some germline mutations have also been identified via sequencing studies. Ongoing genomics studies aim to identify additional high penetrance germline susceptibility alleles for this deadly disease.

Risk Prediction Using Genome-Wide Association Studies on Type 2 Diabetes

The success of genome-wide association studies (GWASs) has enabled us to improve risk assessment and provide novel genetic variants for diagnosis, prevention, and treatment. However, most variants discovered by GWASs have been reported to have very small effect sizes on complex human diseases, which has been a big hurdle in building risk prediction models. Recently, many statistical approaches based on penalized regression have been developed to solve the “large p and small n” problem. In this report, we evaluated the performance of several statistical methods for predicting a binary trait: stepwise logistic regression (SLR), least absolute shrinkage and selection operator (LASSO), and Elastic-Net (EN). We first built a prediction model by combining variable selection and prediction methods for type 2 diabetes using Affymetrix Genome-Wide Human SNP Array 5.0 from the Korean Association Resource project. We assessed the risk prediction performance using area under the receiver operating characteristic curve (AUC) for the internal and external validation datasets. In the internal validation, SLR-LASSO and SLR-EN tended to yield more accurate predictions than other combinations. During the external validation, the SLR-SLR and SLR-EN combinations achieved the highest AUC of 0.726. We propose these combinations as a potentially powerful risk prediction model for type 2 diabetes.

Pathway-based approach using hierarchical components of collapsed rare variants

MOTIVATION

To address 'missing heritability' issue, many statistical methods for pathway-based analyses using rare variants have been proposed to analyze pathways individually. However, neglecting correlations between multiple pathways can result in misleading solutions, and pathway-based analyses of large-scale genetic datasets require massive computational burden. We propose a Pathway-based approach using HierArchical components of collapsed RAre variants Of High-throughput sequencing data (PHARAOH) for the analysis of rare variants by constructing a single hierarchical model that consists of collapsed gene-level summaries and pathways and analyzes entire pathways simultaneously by imposing ridge-type penalties on both gene and pathway coefficient estimates; hence our method considers the correlation of pathways without constraint by a multiple testing problem.

RESULTS

Through simulation studies, the proposed method was shown to have higher statistical power than the existing pathway-based methods. In addition, our method was applied to the large-scale whole-exome sequencing data with levels of a liver enzyme using two well-known pathway databases Biocarta and KEGG. This application demonstrated that our method not only identified associated pathways but also successfully detected biologically plausible pathways for a phenotype of interest. These findings were successfully replicated by an independent large-scale exome chip study.

AVAILABILITY AND IMPLEMENTATION

An implementation of PHARAOH is available at http://statgen.snu.ac.kr/software/pharaoh/

CONTACT

tspark@stats.snu.ac.kr

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.