High-Dimensional Mediation Analysis With Confounders in Survival Models

Mediation analysis is a common statistical method for investigating the mechanism of environmental exposures on health outcomes. Previous studies have extended mediation models with a single mediator to high-dimensional mediators selection. It is often assumed that there are no confounders that influence the relations among the exposure, mediator, and outcome. This is not realistic for the observational studies. To accommodate the potential confounders, we propose a concise and efficient high-dimensional mediation analysis procedure using the propensity score for adjustment. Results from simulation studies demonstrate the proposed procedure has good performance in mediator selection and effect estimation compared with methods that ignore all confounders. Of note, as the sample size increases, the performance of variable selection and mediation effect estimation is as well as the results shown in the method which include all confounders as covariates in the mediation model. By applying this procedure to a TCGA lung cancer data set, we find that lung cancer patients who had serious smoking history have increased the risk of death via the methylation markers cg21926276 and cg20707991 with significant hazard ratios of 1.2093 (95% CI: 1.2019–1.2167) and 1.1388 (95% CI: 1.1339–1.1438), respectively.

Download Full-text

The mediation effect of serum metabolites on the relationship between long-term smoking exposure and esophageal squamous cell carcinoma

BMC Cancer ◽

10.1186/s12885-021-08151-6 ◽

2021 ◽

Vol 21 (1) ◽

Author(s):

Mengke Wei ◽

Lihong Zhao ◽

Jiali Lv ◽

Xia Li ◽

Guangshuai Zhou ◽

...

Keyword(s):

Squamous Cell Carcinoma ◽

Cigarette Smoking ◽

Esophageal Squamous Cell Carcinoma ◽

Cholic Acid ◽

Mediation Analysis ◽

Mediation Effect ◽

High Dimensional ◽

Smoking Exposure ◽

Smoking Index

Abstract Background Long-term smoking exposure will increase the risk of esophageal squamous cell carcinoma (ESCC), whereas the mechanism is still unclear. We conducted a cross-sectional study to explore whether serum metabolites mediate the occurrence of ESCC caused by cigarette smoking. Methods Serum metabolic profiles and lifestyle information of 464 participants were analyzed. Multiple logistic regression was used to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) of smoking exposure to ESCC risk. High-dimensional mediation analysis and univariate mediation analysis were performed to screen potential intermediate metabolites of smoking exposure for ESCC. Results Ever smoking was associated with a 3.11-fold increase of ESCC risk (OR = 3.11, 95% CI 1.63–6.05), and for each cigarette-years increase in smoking index, ESCC risk increased by 56% (OR = 1.56, 95% CI 1.18–2.13). A total of 5 metabolites were screened as mediators by high-dimensional mediation analysis. In addition, glutamine, histidine, and cholic acid were further proved existing mediation effects according to univariate mediation analysis. And the proportions of mediation of histidine and glutamine were 40.47 and 30.00%, respectively. The mediation effect of cholic acid was 8.98% according to the analysis of smoking index. Conclusions Our findings suggest that cigarette smoking contributed to incident ESCC, which may be mediated by glutamine, histidine and cholic acid.

Download Full-text

Investigation of the international comparability of population-based routine hospital data set derived comorbidity scores for patients with lung cancer

Thorax ◽

10.1136/thoraxjnl-2017-210362 ◽

2017 ◽

Vol 73 (4) ◽

pp. 339-349 ◽

Cited By ~ 4

Author(s):

Margreet Lüchtenborg ◽

Eva J A Morris ◽

Daniela Tataru ◽

Victoria H Coupland ◽

Andrew Smith ◽

...

Keyword(s):

Lung Cancer ◽

Cancer Survival ◽

Population Based ◽

Data Sets ◽

Hospital Data ◽

Data Set ◽

Discharge Data ◽

International Differences ◽

Cancer Data ◽

Comorbidity Scores

IntroductionThe International Cancer Benchmarking Partnership (ICBP) identified significant international differences in lung cancer survival. Differing levels of comorbid disease across ICBP countries has been suggested as a potential explanation of this variation but, to date, no studies have quantified its impact. This study investigated whether comparable, robust comorbidity scores can be derived from the different routine population-based cancer data sets available in the ICBP jurisdictions and, if so, use them to quantify international variation in comorbidity and determine its influence on outcome.MethodsLinked population-based lung cancer registry and hospital discharge data sets were acquired from nine ICBP jurisdictions in Australia, Canada, Norway and the UK providing a study population of 233 981 individuals. For each person in this cohort Charlson, Elixhauser and inpatient bed day Comorbidity Scores were derived relating to the 4–36 months prior to their lung cancer diagnosis. The scores were then compared to assess their validity and feasibility of use in international survival comparisons.ResultsIt was feasible to generate the three comorbidity scores for each jurisdiction, which were found to have good content, face and concurrent validity. Predictive validity was limited and there was evidence that the reliability was questionable.ConclusionThe results presented here indicate that interjurisdictional comparability of recorded comorbidity was limited due to probable differences in coding and hospital admission practices in each area. Before the contribution of comorbidity on international differences in cancer survival can be investigated an internationally harmonised comorbidity index is required.

Download Full-text

Comparison of haplotype-based tests for detecting gene–environment interactions with rare variants

Briefings in Bioinformatics ◽

10.1093/bib/bbz031 ◽

2019 ◽

Vol 21 (3) ◽

pp. 851-862 ◽

Cited By ~ 1

Author(s):

Charalampos Papachristou ◽

Swati Biswas

Keyword(s):

Lung Cancer ◽

Rare Variants ◽

Practical Interest ◽

Serum Triglyceride ◽

Error Rates ◽

Type I ◽

Rare Haplotype ◽

Data Set ◽

Cancer Data ◽

Gene Environment

Abstract Dissecting the genetic mechanism underlying a complex disease hinges on discovering gene–environment interactions (GXE). However, detecting GXE is a challenging problem especially when the genetic variants under study are rare. Haplotype-based tests have several advantages over the so-called collapsing tests for detecting rare variants as highlighted in recent literature. Thus, it is of practical interest to compare haplotype-based tests for detecting GXE including the recent ones developed specifically for rare haplotypes. We compare the following methods: haplo.glm, hapassoc, HapReg, Bayesian hierarchical generalized linear model (BhGLM) and logistic Bayesian LASSO (LBL). We simulate data under different types of association scenarios and levels of gene–environment dependence. We find that when the type I error rates are controlled to be the same for all methods, LBL is the most powerful method for detecting GXE. We applied the methods to a lung cancer data set, in particular, in region 15q25.1 as it has been suggested in the literature that it interacts with smoking to affect the lung cancer susceptibility and that it is associated with smoking behavior. LBL and BhGLM were able to detect a rare haplotype–smoking interaction in this region. We also analyzed the sequence data from the Dallas Heart Study, a population-based multi-ethnic study. Specifically, we considered haplotype blocks in the gene ANGPTL4 for association with trait serum triglyceride and used ethnicity as a covariate. Only LBL found interactions of haplotypes with race (Hispanic). Thus, in general, LBL seems to be the best method for detecting GXE among the ones we studied here. Nonetheless, it requires the most computation time.

Download Full-text

Survival after cancer diagnosis among patients with higher income and education in Israel, a country with highly appreciated health service.

Journal of Clinical Oncology ◽

10.1200/jco.2013.31.15_suppl.1598 ◽

2013 ◽

Vol 31 (15_suppl) ◽

pp. 1598-1598

Author(s):

Yakir Rottenberg ◽

Aviad Zick ◽

Tamar Peretz

Keyword(s):

Lung Cancer ◽

Cancer Diagnosis ◽

Proportional Hazards ◽

Cox Proportional Hazards ◽

Lower Socioeconomic Status ◽

Cancer Data ◽

Risk Of Death ◽

Lower Socioeconomic ◽

Treatment Disparities ◽

After Cancer

1598 Background: In recent years, the 5 year survival following cancer diagnosis is about two thirds. Among patients with various chronic diseases, improved survival is known to be associated with higher income and education. The aim of the current study is to assess the influence of income and education on survival following cancer diagnosis in Israel. Methods: Retrospective cohort study, using baseline measurement from the 1995 census conducted by the Central Bureau of Statistics in Israel. Cancer data were obtained from the Israel Cancer Registry. Cox proportional hazards ratios were calculated for mortality among cancer patients and adjusted for age, sex, religious, income and education years. The first model excluded cancers associated with early detection (breast, prostate, colorectal and cervix), and a second model excluded also lung cancer in order to control for smoking which is common in lower socioeconomic status. Results: A total of 3,712 cases of cancer and 1,252 deaths were reported during the study period. Higher income (HR=0.985 per 1000NIS, approximately 330$ in 1995's value, p=0.016) and education (HR=0.957 per year of education, p<0.001) were associated with decreased risk of death after cancer diagnosis. Jews had better prognosis than non-Jews following cancer diagnosis (HR=0.62, p<0.001), while males (HR=1.54, p<0.001) and age (HR=1.036 per year, p<0.001) had been associated with worse prognosis. The association between higher income and education was not changed in a model which excluded lung cancer. Conclusions: Higher income and education are associated with improved survival after cancer diagnosis. In the light of current study, further studies are needed to depict the variation in cancer incidence, stage at diagnosis and treatment disparities related to socioeconomic variables.

Download Full-text

Variable selection for ultra-high dimensional quantile regression with missing data and measurement error

Statistical Methods in Medical Research ◽

10.1177/0962280220941533 ◽

2020 ◽

pp. 096228022094153

Author(s):

Yongxin Bai ◽

Maozai Tian ◽

Man-Lai Tang ◽

Wing-Yan Lee

Keyword(s):

Missing Data ◽

Measurement Error ◽

Variable Selection ◽

Quantile Regression ◽

Measurement Errors ◽

Estimation Method ◽

High Dimensional ◽

Data Set ◽

Cancer Data ◽

Selection For

In this paper, we consider variable selection for ultra-high dimensional quantile regression model with missing data and measurement errors in covariates. Specifically, we correct the bias in the loss function caused by measurement error by applying the orthogonal quantile regression approach and remove the bias caused by missing data using the inverse probability weighting. A nonconvex Atan penalized estimation method is proposed for simultaneous variable selection and estimation. With the proper choice of the regularization parameter and under some relaxed conditions, we show that the proposed estimate enjoys the oracle properties. The choice of smoothing parameters is also discussed. The performance of the proposed variable selection procedure is assessed by Monte Carlo simulation studies. We further demonstrate the proposed procedure with a breast cancer data set.

Download Full-text

A Generic Data Harmonization Process for Cross-linked Research and Network Interaction

Methods of Information in Medicine ◽

10.3414/me14-02-0030 ◽

2015 ◽

Vol 54 (05) ◽

pp. 455-460 ◽

Cited By ~ 4

Author(s):

M. Ganzinger ◽

T. Muley ◽

M. Thomas ◽

P. Knaup ◽

D. Firnkorn

Keyword(s):

Lung Cancer ◽

Research Networks ◽

Data Set ◽

Common Data Elements ◽

Cancer Data ◽

Data Harmonization ◽

Wide Range ◽

Harmonization Process ◽

Cancer Phenotype ◽

Data Elements

Summary Objective: Joint data analysis is a key requirement in medical research networks. Data are available in heterogeneous formats at each network partner and their harmonization is often rather complex. The objective of our paper is to provide a generic approach for the harmonization process in research networks. We applied the process when harmonizing data from three sites for the Lung Cancer Phenotype Database within the German Center for Lung Research. Methods: We developed a spreadsheet-based solution as tool to support the harmonization process for lung cancer data and a data integration procedure based on Talend Open Studio. Results: The harmonization process consists of eight steps describing a systematic approach for defining and reviewing source data elements and standardizing common data elements. The steps for defining common data elements and harmonizing them with local data definitions are repeated until consensus is reached. Application of this process for building the phenotype database led to a common basic data set on lung cancer with 285 structured parameters. The Lung Cancer Phenotype Database was realized as an i2b2 research data warehouse. Conclusion: Data harmonization is a challenging task requiring informatics skills as well as domain knowledge. Our approach facilitates data harmonization by providing guidance through a uniform process that can be applied in a wide range of projects.

Download Full-text

Multifactorial Analysis of Mortality in Screening Detected Lung Cancer

Journal of Oncology ◽

10.1155/2018/1296246 ◽

2018 ◽

Vol 2018 ◽

pp. 1-7 ◽

Cited By ~ 5

Author(s):

Subba R. Digumarthy ◽

Ruben De Man ◽

Rodrigo Canellas ◽

Alexi Otrakji ◽

Ge Wang ◽

...

Keyword(s):

Lung Cancer ◽

Muscle Mass ◽

Low Dose ◽

Subcutaneous Fat ◽

Lung Cancer Mortality ◽

Data Access ◽

Smoking History ◽

Quantitative Prediction ◽

Cancer Data ◽

Significant Difference

We hypothesized that severity of coronary artery calcification (CAC), emphysema, muscle mass, and fat attenuation can help predict mortality in patients with lung cancer participating in the National Lung Screening Trial (NLST). Following regulatory approval from the Cancer Data Access System (CDAS), all patients diagnosed with lung cancer at the time of the screening study were identified. These subjects were classified into two groups: survivors and nonsurvivors at the conclusion of the NLST trial. These groups were matched based on their age, gender, body mass index (BMI), smoking history, lung cancer stage, and survival time. CAC, emphysema, muscle mass, and subcutaneous fat attenuation were quantified on baseline low-dose chest CT (LDCT) for all patients in both groups. Nonsurvivor group had significantly greater CAC, decreased muscle mass, and higher fat attenuation compared to the survivor group (p<0.01). No significant difference in severity of emphysema was noted between the two groups (p>0.1). We thus conclude that it is possible to create a quantitative prediction model for lung cancer mortality for subjects with lung cancer detected on screening low-dose CT (LDCT).

Download Full-text

GOMPERTZ REGRESSION MODEL WITH GAMMA FRAILTY: A STUDY ON THE APPLICATION IN LUNG CANCER

REVISTA BRASILEIRA DE BIOMETRIA ◽

10.28951/rbb.v36i4.312 ◽

2018 ◽

Vol 36 (4) ◽

pp. 860

Author(s):

Vera Lucia Damasceno TOMAZELLA ◽

Eder Ângelo MILANI ◽

Teresa Cristina Martins DIAS

Keyword(s):

Lung Cancer ◽

Survival Function ◽

Random Effect ◽

Gompertz Model ◽

Survival Models ◽

Cancer Data ◽

Gamma Frailty ◽

Maximum Likelihood Methods ◽

The Laplace Transform ◽

The Individual

Survival models with frailty are used when some variables are non-available to explain the occurrence time of an event of interest. This non-availability may be considered as a random effect related to unobserved covariates, or that cannot be measured, such as environmental or genetic factors. This paper focuses on the Gamma-Gompertz (denoted by G-G) model that is one of a class of models that investigate the effects of unobservable heterogeneity. We assume that the baseline mortality rate in the G-G model is the Gompertz model, in which mortality increases exponentially with age and the frailty is a fixed property of the individual, and the distribution of frailty is a gamma distribution. The proposed methodology uses the Laplace transform to find the unconditional survival function in the individual frailty. Estimation is based on maximum likelihood methods and this distribution is compared with its particular case. A simulation study examines the bias, the mean squared errors and the coverage probabilities considering various samples sizes and censored data. A real example with lung cancer data illustrates the applicability of the methodology, where we compared the G-G and without frailty models via criteria which select thebest fitted model to the data.

Download Full-text

Risk of SARS-CoV2-related Mortality in Non-small Cell Lung Cancer Patients Treated With First-line Immunotherapy Alone or in Combination With Chemotherapy

10.21203/rs.3.rs-167896/v1 ◽

2021 ◽

Author(s):

Giuseppe Luigi Banna ◽

Ornella Cantale ◽

Alex Friedlaender ◽

Harliana Yusof ◽

Alfredo Addeo

Keyword(s):

Lung Cancer ◽

Small Cell Lung Cancer ◽

Cell Lung Cancer ◽

Small Cell ◽

Smoking History ◽

Small Cell Lung ◽

First Line ◽

Risk Of Death ◽

The Impact ◽

Related Mortality

Abstract Background: Patients with cancer are vulnerable to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), although the impact of solid cancer types and systemic anticancer treatments on its related mortality is still debatable.Methods: To weigh the real impact of immune-checkpoint inhibitors (ICIs) by exploring the risk of SARS-CoV-2-related mortality in a retrospective analysis of patients with non-small-cell lung cancer (NSCLC) treated with first-line Pembrolizumab or in combination with chemotherapy (ChT) during the first surge of the pandemic. Results: The risk of death was significantly higher in the group of patients treated with ChT+Pembrolizumab than with Pembrolizumab alone (OR 2.43 (1.23-4.82, p=0.01). The SARS-CoV-2-related mortality rate was 8% and significantly associated with ChT+Pembrolizumab as compared to Pembrolizumab alone (18% vs. 0%, respectively, p=0.03). Patients dead because of SARS-CoV-2 were older than 70 years (100 vs. 34%, respectively, p=0.03) and tended to have a heavier smoking history (67 vs. 29% of current smokers, respectively, p=0.17). Higher baseline values of neutrophil-to-lymphocyte ratio (NLR) (with 67 vs. 50% ≥ 4.0, p=0.58) and systemic immune-inflammation index (SII) (with 67 vs. 50% ≥ 1236, p=0.58) were observed in patients dead due to the SARS-CoV-2.Conclusions: Immunotherapy might not impact the risk of SARS-CoV-2-related mortality, whilst the addition of ChT was either associated with an overall increased risk of mortality and to the risk of SARS-CoV-2-related mortality. The co-existence of other clinical factors may have contributed to the latter.

Download Full-text

Lung cancer in African-Americans and analysis of estrogen plus progestin use.

Journal of Clinical Oncology ◽

10.1200/jco.2019.37.15_suppl.e18258 ◽

2019 ◽

Vol 37 (15_suppl) ◽

pp. e18258-e18258

Author(s):

Idoroenyi Usua Amanam ◽

Rowan T. Chlebowski ◽

Rebecca A. Nelson ◽

Ravi Salgia

Keyword(s):

Lung Cancer ◽

Cox Regression ◽

Controlled Trial ◽

African Ancestry ◽

Snp Array ◽

Smoking History ◽

Risk Of Death ◽

Increased Risk ◽

Estrogen Plus Progestin ◽

The Impact

e18258 Background: The 15-year Women’s Health Initiative (WHI) sponsored by the NIH has provided a robust dataset on health risks for post-menopausal black women (BW), including the impact of hormone therapy (HRT) on cancer risk. Women enrolled in the WHI randomized, placebo-controlled trial and taking HRT demonstrated no increase in lung cancer incidence, but a statistically significant increase in mortality. However, effects of estrogen plus progestin on non-small cell lung cancer (NSCLC) incidence and outcomes has not been extensively examined, especially in African Ancestry (AA) and smoking history. Methods: Study participants were identified who met WHI clinical trial entry criteria. Cox regression models and Kaplan-Meier method plots were utilized. Analyses adjusted for age, BMI, education, smoking status, alcohol use, health status, and physical activity. A secondary analysis was performed on BW based on AA via Affymetrix Human SNP Array. Results: 161, 808 pts were enrolled from October 1993 to December 1998 (after exclusions total analytic cohort = 142,503). 128,682 (90%) were white (WW) and 13,821 (10%) were BW. BW had lower incidence of NSCLC compared to WW (HR 0.68; P < .0001). HRT participants had a 55% increase in incidence of NSCLC (p < .0001). Former alcohol users had highest risk of NSCLC incidence (HR 2.72; p < 0.0001). Age groups (55-59 years; 60-69 years; 70-79 years) were significantly less associated with BW compared to the youngest(50-54 years; P < .0001). HRT participants were more likely BW (OR 1.17; p < .0001). More current smokers were BW compared to WW (OR 1.75; p < .0001). HRT participants had increased risk of death to NSCLC (HR 1.29; p < .001). There was a trend for survival (p = 0.3667) in WW participants compared to BW (32 vs 28.0 months, respectively). BW who had > 80% AA had a decreased incidence NSCLC trend compared to BW with < 80% AA (HR 0.81; p = 0.2806). Conclusions: BW, especially those with high levels of AA had decreased incidence of NSCLC. Those patients who received HRT had higher incidence and death from NSCLC. Further investigations are required to understand the mechanisms that AA and HRT alter risks associated with NSCLC.

Download Full-text