Sample size considerations using mathematical models: an example with Chlamydia trachomatis infection and its sequelae pelvic inflammatory disease.

BACKGROUND The success of an intervention to prevent the complications of an infection is influenced by the natural history of the infection. Assumptions about the temporal relationship between infection and the development of sequelae can affect the predicted effect size of an intervention and the sample size calculation. This study investigates how a mathematical model can be used to inform sample size calculations for a randomised controlled trial (RCT) using the example of Chlamydia trachomatis infection and pelvic inflammatory disease (PID). METHODS We used a compartmental model to imitate the structure of a published RCT. We considered three different processes for the timing of PID development, in relation to the initial C. trachomatis infection: immediate, constant throughout, or at the end of the infectious period. For each process we assumed that, of all women infected, the same fraction would develop PID in the absence of an intervention. We examined two sets of assumptions used to calculate the sample size in a published RCT that investigated the effect of chlamydia screening on PID incidence. We also investigated the influence of the natural history parameters of chlamydia on the required sample size. RESULTS The assumed event rates and effect sizes used for the sample size calculation implicitly determined the temporal relationship between chlamydia infection and PID in the model. Even small changes in the assumed PID incidence and relative risk (RR) led to considerable differences in the hypothesised mechanism of PID development. The RR and the sample size needed per group also depend on the natural history parameters of chlamydia. CONCLUSIONS Mathematical modelling helps to understand the temporal relationship between an infection and its sequelae and can show how uncertainties about natural history parameters affect sample size calculations when planning a RCT.

Pre-clinical in vivo models for the screening of bone biomaterials for oral/craniofacial indications: focus on small-animal models.

Preclinical in vivo experimental studies are performed for evaluating proof-of-principle concepts, safety and possible unwanted reactions of candidate bone biomaterials before proceeding to clinical testing. Specifically, models involving small animals have been developed for screening bone biomaterials for their potential to enhance bone formation. No single model can completely recreate the anatomic, physiologic, biomechanic and functional environment of the human mouth and jaws. Relevant aspects regarding physiology, anatomy, dimensions and handling are discussed in this paper to elucidate the advantages and disadvantages of small-animal models. Model selection should be based not on the 'expertise' or capacities of the team, but rather on a scientifically solid rationale, and the animal model selected should reflect the question for which an answer is sought. The rationale for using heterotopic or orthotopic testing sites, and intraosseous, periosseous or extraskeletal defect models, is discussed. The paper also discusses the relevance of critical size defect modeling, with focus on calvarial defects in rodents. In addition, the rabbit sinus model and the capsule model in the rat mandible are presented and discussed in detail. All animal experiments should be designed with care and include sample-size and study-power calculations, thus allowing generation of meaningful data. Moreover, animal experiments are subject to ethical approval by the relevant authority. All procedures and the postoperative handling and care, including postoperative analgesics, should follow best practice.

A Non-Synonymous HMGA2 Variant Decreases Height in Shetland Ponies and Other Small Horses

The identification of quantitative trait loci (QTL) such as height and their underlying causative variants is still challenging and often requires large sample sizes. In humans hundreds of loci with small effects control the heritable portion of height variability. In domestic animals, typically only a few loci with comparatively large effects explain a major fraction of the heritability. We investigated height at withers in Shetland ponies and mapped a QTL to ECA 6 by genome-wide association (GWAS) using a small cohort of only 48 animals and the Illumina equine SNP70 BeadChip. Fine-mapping revealed a shared haplotype block of 793 kb in small Shetland ponies. The HMGA2 gene, known to be associated with height in horses and many other species, was located in the associated haplotype. After closing a gap in the equine reference genome we identified a non-synonymous variant in the first exon of HMGA2 in small Shetland ponies. The variant was predicted to affect the functionally important first AT-hook DNA binding domain of the HMGA2 protein (c.83G>A; p.G28E). We assessed the functional impact and found impaired DNA binding of a peptide with the mutant sequence in an electrophoretic mobility shift assay. This suggests that the HMGA2 variant also affects DNA binding in vivo and thus leads to reduced growth and a smaller stature in Shetland ponies. The identified HMGA2 variant also segregates in several other pony breeds but was not found in regular-sized horse breeds. We therefore conclude that we identified a quantitative trait nucleotide for height in horses.

Reconstruction of in-situ porosity and porewater compositions of low-permeability crystalline rocks: Magnitude of artefacts induced by drilling and sample recovery

Geological site characterisation programmes typically rely on drill cores for direct information on subsurface rocks. However, porosity, transport properties and porewater composition measured on drill cores can deviate from in-situ values due to two main artefacts caused by drilling and sample recovery: (1) mechanical disruption that increases porosity and (2) contamination of the porewater by drilling fluid. We investigated the effect and magnitude of these perturbations on large drill core samples (12–20 cm long, 5 cmdiameter) of high-grade, granitic gneisses obtained from 350 to 600 m depth in a borehole on Olkiluoto Island (SW Finland). The drilling fluid was traced with sodium–iodide. By combining out-diffusion experiments, gravimetry, UV-microscopy and iodide mass balance calculations, we successfully quantified the magnitudes of the artefacts: 2–6% increase in porosity relative to the bulk connected porosity and 0.9 to 8.9 vol.% contamination by drilling fluid. The spatial distribution of the drilling-induced perturbations was revealed by numerical simulations of 2D diffusion matched to the experimental data. This showed that the rims of the samples have a mechanically disrupted zone 0.04 to 0.22 cm wide, characterised by faster transport properties compared to the undisturbed centre (1.8 to 7.7 times higher pore diffusion coefficient). Chemical contamination was shown to affect an even wider zone in all samples, ranging from 0.15 to 0.60 cm, inwhich iodide enrichmentwas up to 180 mg/kgwater, compared to 0.5 mg/kgwater in the uncontaminated centre. For all samples in the present case study, it turned out that the magnitude of the artefacts caused by drilling and sample recovery is so small that no correction is required for their effects. Therefore, the standard laboratory measurements of porosity, transport properties and porewater composition can be taken as valid in-situ estimates. However, it is clear that the magnitudes strongly depend on site- and drilling-specific factors and therefore our results cannot be transferred simply to other locations. We recommend the approach presented in this study as a route to obtain reliable values in future drilling campaigns aimed at characterising in-situ bedrock properties.

Effects of sample mass and macrofossil type on radiocarbon dating of arctic and boreal lake sediments

Dating lake sediments by accelerator mass spectrometry (AMS) 14C analysis of terrestrial plant macrofossils overcomes one of the main problems associated with dating bulk sediment samples, i.e., the presence of old organic matter. Even so, many AMS dates from arctic and boreal sites appear to misrepresent the age of the sediment. To understand the nature of these apparent dating anomalies better, we conducted a series of 14C dating experiments using samples from Alaskan and Siberian lake-sediment cores. First, to test whether our analytical procedures introduced a sample-mass bias, we obtained 14C dates for different-sized pieces of single woody macrofossils. In these sample-mass experiments, statistically equivalent ages were found for samples as small as 0.05 mg C. Secondly, to assess whether macrofossil type influenced dating results, we conducted sample-type experiments in which 14C dates were obtained for different macrofossil types sieved from the same depth in the sediment. We dated materials from multiple levels in sediment cores from Upper Capsule Lake (North Slope, northern Alaska) and Grizzly Lake (Copper River Basin, southern Alaska) and from single depths in other records from northern Alaska. In several of the experiments there were significant discrepancies between dates for different plant tissues, and in most cases wood and charcoal were older than other macrofossil types, usually by several hundred years. This pattern suggests that 14C dates for woody macrofossils may misrepresent the age of the sediment by centuries, perhaps because of their longer terrestrial residence time and the potential in-built age of longlived plants. This study identifies why some 14C dates appear to be inconsistent with the overall age-depth trend of a lake-sediment record, and it may guide the selection of 14C samples in future studies.

Canopy gaps promote selective stem-cutting by small mammals of two dominant tree species in an African lowland forest: the importance of seedling chemistry

Small mammals can impede tree regeneration by injuring seedlings and saplings in several ways. One fatal way is by severing their stems, but apparently this type of predation is not well-studied in tropical rain forest. Here, we report on the incidence of 'stem-cutting' to new, wild seedlings of two locally dominant, canopy tree species monitored in 40 paired forest understorey and gap-habitat areas in Korup, Cameroon following a 2007 masting event. In gap areas, which are required for the upward growth and sapling recruitment of both species, 137 seedlings of the long-lived, light-demanding, fast-growing large tropical tree (Microberlinia bisulcata) were highly susceptible to stem-cutting (83% of deaths) - it killed 39% of all seedlings over a c. 2-y period. In stark contrast, seedlings of the more shade-tolerant, slower-growing tree species (Tetraberlinia bifoliolata) were hardly attacked (4.3%). In the understorey, however, stem-cutting was virtually absent. Across the gap areas, the incidence of stem-cutting of M. bisulcata seedlings showed significant spatial variation that could not be explained significantly by either canopy openness or Janzen-Connell type effects (proximity and basal area of conspecific adult trees). To examine physical and chemical traits that might explain the species difference to being cut, bark and wood tissues were collected from a separate sample of seedlings in gaps (i.e. not monitored for stem-cutting). These analyses suggested that, compared with T. bifoliolata, the lower stem density, higher Mg and K and fatty acid concentrations in bark, and fewer phenolic and terpene compounds in M. bisulcata seedlings made them more palatable and attractive to small-mammal predators, likely rodents. We conclude that selective stem-cutting is a potent countervailing force to the current local canopy dominance of the grove-forming M. bisulcata by limiting the recruitment and abundance of its saplings. Given the ubiquity of gaps and ground-dwelling rodents in pantropical forests, it would be surprising if this form of lethal browsing was restricted to Korup.

Coxiella burnetii Infections in Small Ruminants and Humans in Switzerland

The recent Q fever epidemic in the Netherlands raised concerns about the potential risk of outbreaks in other European countries. In Switzerland, the prevalence of Q fever in animals and humans has not been studied in recent years. In this study, we describe the current situation with respect to Coxiella (C.) burnetii infections in small ruminants and humans in Switzerland, as a basis for future epidemiological investigations and public health risk assessments. Specific objectives of this cross-sectional study were to (i) estimate the seroprevalence of C. burnetii in sheep and goats, (ii) quantify the amount of bacteria shed during abortion and (iii) analyse temporal trends in human C. burnetii infections. The seroprevalence of C. burnetii in small ruminants was determined by commercial ELISA from a representative sample of 100 sheep flocks and 72 goat herds. Herd-level seroprevalence was 5.0% (95% CI: 1.6-11.3) for sheep and 11.1% (95% CI: 4.9-20.7) for goats. Animal-level seroprevalence was 1.8% (95% CI: 0.8-3.4) for sheep and 3.4% (95% CI: 1.7-6) for goats. The quantification of C. burnetii in 97 ovine and caprine abortion samples by real-time PCR indicated shedding of >10(4) bacteria/g in 13.4% of all samples tested. To our knowledge, this is the first study reporting C. burnetii quantities in a large number of small ruminant abortion samples. Annual human Q fever serology data were provided by five major Swiss laboratories. Overall, seroprevalence in humans ranged between 1.7% and 3.5% from 2007 to 2011, and no temporal trends were observed. Interestingly, the two laboratories with significantly higher seroprevalences are located in the regions with the largest goat populations as well as, for one laboratory, with the highest livestock density in Switzerland. However, a direct link between animal and human infection data could not be established in this study.

Multiple imputation for missing continuous and dichotomous data: The effects of patterns of missingness and variable correlations on type I error with small samples

The purpose of this study is to investigate the effects of predictor variable correlations and patterns of missingness with dichotomous and/or continuous data in small samples when missing data is multiply imputed. Missing data of predictor variables is multiply imputed under three different multivariate models: the multivariate normal model for continuous data, the multinomial model for dichotomous data and the general location model for mixed dichotomous and continuous data. Subsequent to the multiple imputation process, Type I error rates of the regression coefficients obtained with logistic regression analysis are estimated under various conditions of correlation structure, sample size, type of data and patterns of missing data. The distributional properties of average mean, variance and correlations among the predictor variables are assessed after the multiple imputation process. ^ For continuous predictor data under the multivariate normal model, Type I error rates are generally within the nominal values with samples of size n = 100. Smaller samples of size n = 50 resulted in more conservative estimates (i.e., lower than the nominal value). Correlation and variance estimates of the original data are retained after multiple imputation with less than 50% missing continuous predictor data. For dichotomous predictor data under the multinomial model, Type I error rates are generally conservative, which in part is due to the sparseness of the data. The correlation structure for the predictor variables is not well retained on multiply-imputed data from small samples with more than 50% missing data with this model. For mixed continuous and dichotomous predictor data, the results are similar to those found under the multivariate normal model for continuous data and under the multinomial model for dichotomous data. With all data types, a fully-observed variable included with variables subject to missingness in the multiple imputation process and subsequent statistical analysis provided liberal (larger than nominal values) Type I error rates under a specific pattern of missing data. It is suggested that future studies focus on the effects of multiple imputation in multivariate settings with more realistic data characteristics and a variety of multivariate analyses, assessing both Type I error and power. ^

The Real Exchange Rate in Small Open Developed Economies: Evidence from Cointegration Analysis

We examine the effects of the terms of trade and the expected real interest rate differential on the real exchange rate in a sample of small open developed economies. We employ cointegration analysis to search for possible long-term linkages. We find that while both the terms of trade and the expected real interest rate differentials affect the real exchange rate in the long run, the role of the terms of trade generally proves more consistent across countries. The speed of adjustment for the expected real interest rate differential in the error-correction model, however, is quantitatively larger than it is for the terms of trade.

Evaluation of the causal pathways through which subjective norms influence mammography intention and stage of change in a national sample of women veterans

Few studies have investigated causal pathways linking psychosocial factors to each other and to screening mammography. Conflicting hypotheses exist in the theoretic literature regarding the role and importance of subjective norms, a person's perceived social pressure to perform the behavior and his/her motivation to comply. The Theory of Reasoned Action (TRA) hypothesizes that subjective norms directly affect intention; while the Transtheoretical Model (TTM) hypothesizes that attitudes mediate the influence of subjective norms on stage of change. No one has examined which hypothesis best predicts the effect of subjective norms on mammography intention and stage of change. Two statistical methods are available for testing mediation, sequential regression analysis (SRA) and latent variable structural equation modeling (LVSEM); however, software to apply LVSEM to dichotomous variables like intention has only recently become available. No one has compared the methods to determine whether or not they yield similar results for dichotomous variables. ^ Study objectives were to: (1) determine whether the effect of subjective norms on mammography intention and stage of change are mediated by pros and cons; and (2) compare mediation results from the SRA and LVSEM approaches when the outcome is dichotomous. We conducted a secondary analysis of data from a national sample of women veterans enrolled in Project H.O.M.E. (H&barbelow;ealthy O&barbelow;utlook on the M&barbelow;ammography E&barbelow;xperience), a behavioral intervention trial. ^ Results showed that the TTM model described the causal pathways better than the TRA one; however, we found support for only one of the TTM causal mechanisms. Cons was the sole mediator. The mediated effect of subjective norms on intention and stage of change by cons was very small. These findings suggest that interventionists focus their efforts on reducing negative attitudes toward mammography when resources are limited. ^ Both the SRA and LVSEM methods provided evidence for complete mediation, and the direction, magnitude, and standard errors of the parameter estimates were very similar. Because SRA parameter estimates were not biased toward the null, we can probably assume negligible measurement error in the independent and mediator variables. Simulation studies are needed to further our understanding of how these two methods perform under different data conditions. ^

Health, education and community: Assets and needs of a small urban community in north east Tamaulipas, Mexico: A community based participatory research assessment

This study explored the health, education, social assets, needs, attitudes, and behaviors of residents of Ferrocarril #4, a small urban community in Tamaulipas, Mexico. A collaborative Participatory Action Research approach was used to emphasize community involvement. Using Triangulation to ensure validity, qualitative methods included key informant in depth interviews, participant observation and participatory discussion groups with women and men. A personal interview with a probability sample of women was done. The median age of interviewees was 37 years. The majority was married or had a partner. Over half of respondents completed grades 6-9. Employed women (25%) earned a median weekly salary equivalent to ∼56 USD. Women with health insurance (67.7%) were covered mainly through Social Security and Seguro Popular. One in 5 reported bad health. Barriers to care were primarily money and transportation. To improve health care, women wanted a full service clinic in or close to the community and affordable health care. Socially, 28% of respondents had no close friends in the community and most did not participate in beneficial community activities. Many women did not socialize with others and help from neighbors was situational. Primary school teachers lacked parental support and it interfered with classroom efforts. Healthy community discussion groups focused on personal and environmental hygiene and safety. Valuable assets exist in the community. To date, collaborative efforts resulted in a school First Aid station, a school nurse visit weekly, posting of emergency contact phone numbers in the school and community center, and development of a student health information form. ^

HPV vaccine coverage in Texas counties: An application of multilevel, small area estimation

Health departments, research institutions, policy-makers, and healthcare providers are often interested in knowing the health status of their clients/constituents. Without the resources, financially or administratively, to go out into the community and conduct health assessments directly, these entities frequently rely on data from population-based surveys to supply the information they need. Unfortunately, these surveys are ill-equipped for the job due to sample size and privacy concerns. Small area estimation (SAE) techniques have excellent potential in such circumstances, but have been underutilized in public health due to lack of awareness and confidence in applying its methods. The goal of this research is to make model-based SAE accessible to a broad readership using clear, example-based learning. Specifically, we applied the principles of multilevel, unit-level SAE to describe the geographic distribution of HPV vaccine coverage among females aged 11-26 in Texas.^ Multilevel (3 level: individual, county, public health region) random-intercept logit models of HPV vaccination (receipt of ≥ 1 dose Gardasil® ) were fit to data from the 2008 Behavioral Risk Factor Surveillance System (outcome and level 1 covariates) and a number of secondary sources (group-level covariates). Sampling weights were scaled (level 1) or constructed (levels 2 & 3), and incorporated at every level. Using the regression coefficients (and standard errors) from the final models, I simulated 10,000 datasets for each regression coefficient from the normal distribution and applied them to the logit model to estimate HPV vaccine coverage in each county and respective demographic subgroup. For simplicity, I only provide coverage estimates (and 95% confidence intervals) for counties.^ County-level coverage among females aged 11-17 varied from 6.8-29.0%. For females aged 18-26, coverage varied from 1.9%-23.8%. Aggregated to the state level, these values translate to indirect state estimates of 15.5% and 11.4%, respectively; both of which fall within the confidence intervals for the direct estimates of HPV vaccine coverage in Texas (Females 11-17: 17.7%, 95% CI: 13.6, 21.9; Females 18-26: 12.0%, 95% CI: 6.2, 17.7).^ Small area estimation has great potential for informing policy, program development and evaluation, and the provision of health services. Harnessing the flexibility of multilevel, unit-level SAE to estimate HPV vaccine coverage among females aged 11-26 in Texas counties, I have provided (1) practical guidance on how to conceptualize and conduct modelbased SAE, (2) a robust framework that can be applied to other health outcomes or geographic levels of aggregation, and (3) HPV vaccine coverage data that may inform the development of health education programs, the provision of health services, the planning of additional research studies, and the creation of local health policies.^

A SIMULATION STUDY OF THE SIZE AND POWER OF SOME TWO-SAMPLE TESTS FOR UNEQUALLY CENSORED SURVIVAL DATA (RANK TESTS, LOGRANK, PROPORTIONAL HAZARDS, WILCOXON TESTS, EXPONENTIAL DISTRIBUTION)

Sizes and power of selected two-sample tests of the equality of survival distributions are compared by simulation for small samples from unequally, randomly-censored exponential distributions. The tests investigated include parametric tests (F, Score, Likelihood, Asymptotic), logrank tests (Mantel, Peto-Peto), and Wilcoxon-Type tests (Gehan, Prentice). Equal sized samples, n = 18, 16, 32 with 1000 (size) and 500 (power) simulation trials, are compared for 16 combinations of the censoring proportions 0%, 20%, 40%, and 60%. For n = 8 and 16, the Asymptotic, Peto-Peto, and Wilcoxon tests perform at nominal 5% size expectations, but the F, Score and Mantel tests exceeded 5% size confidence limits for 1/3 of the censoring combinations. For n = 32, all tests showed proper size, with the Peto-Peto test most conservative in the presence of unequal censoring. Powers of all tests are compared for exponential hazard ratios of 1.4 and 2.0. There is little difference in power characteristics of the tests within the classes of tests considered. The Mantel test showed 90% to 95% power efficiency relative to parametric tests. Wilcoxon-type tests have the lowest relative power but are robust to differential censoring patterns. A modified Peto-Peto test shows power comparable to the Mantel test. For n = 32, a specific Weibull-exponential comparison of crossing survival curves suggests that the relative powers of logrank and Wilcoxon-type tests are dependent on the scale parameter of the Weibull distribution. Wilcoxon-type tests appear more powerful than logrank tests in the case of late-crossing and less powerful for early-crossing survival curves. Guidelines for the appropriate selection of two-sample tests are given. ^

TECHNOLOGY ASSESSMENT: SMALL-VOLUME INFUSION PUMPS

"Technology assessment is a comprehensive form of policy research that examines the short- and long-term social consequences of the application or use of technology" (US Congress 1967).^ This study explored a research methodology appropriate for technology assessment (TA) within the health industry. The case studied was utilization of external Small-Volume Infusion Pumps (SVIP) at a cancer treatment and research center. Primary and secondary data were collected in three project phases. In Phase I, hospital prescription records (N = 14,979) represented SVIP adoption and utilization for the years 1982-1984. The Candidate Adoption-Use (CA-U) diffusion paradigm developed for this study was germane. Compared to classic and unorthodox curves, CA-U more accurately simulated empiric experience. The hospital SVIP 1983-1984 trends denoted assurance in prescribing chemotherapy and concomitant balloon SVIP efficacy and efficiency. Abandonment of battery pumps was predicted while exponential demand for balloon SVIP was forecast for 1985-1987. In Phase II, patients using SVIP (N = 117) were prospectively surveyed from July to October 1984; the data represented a single episode of therapy. The questionnaire and indices, specifically designed to measure the impact of SVIP, evinced face validity. Compeer group data were from pre-SVIP case reviews rather than from an inpatient sample. Statistically significant results indicated that outpatients using SVIP interacted socially more than inpatients using the alternative technology. Additionally, the hospital's education program effectively taught clients to discriminate between self care and professional SVIP services. In these contexts, there was sufficient evidence that the alternative technology restricted patients activity whereas SVIP permitted patients to function more independently and in a social lifestyle, thus adding quality to life. In Phase III, diffusion forecast and patient survey findings were combined with direct observation of clinic services to profile some economic dimensions of SVIP. These three project phases provide a foundation for executing: (1) cost effectiveness analysis of external versus internal infusors, (2) institutional resource allocation, and (3) technology deployment to epidemiology-significant communities. The models and methods tested in this research of clinical technology assessment are innovative and do assess biotechnology. ^

Genetic Predictors of Clinical Outcomes in Non-Small Cell Lung Cancer Patients

Lung cancer is the leading cause of cancer-related mortality in the US. Emerging evidence has shown that host genetic factors can interact with environmental exposures to influence patient susceptibility to the diseases as well as clinical outcomes, such as survival and recurrence. We aimed to identify genetic prognostic markers for non-small cell lung cancer (NSCLC), a major (85%) subtype of lung cancer, and also in other subgroups. With the fast evolution of genotyping technology, genetic association studies have went through candidate gene approach, to pathway-based approach, to the genome wide association study (GWAS). Even in the era of GWAS, pathway-based approach has its own advantages on studying cancer clinical outcomes: it is cost-effective, requiring a smaller sample size than GWAS easier to identify a validation population and explore gene-gene interactions. In the current study, we adopted pathway-based approach focusing on two critical pathways - miRNA and inflammation pathways. MicroRNAs (miRNA) post-transcriptionally regulate around 30% of human genes. Polymorphisms within miRNA processing pathways and binding sites may influence patients’ prognosis through altered gene regulation. Inflammation plays an important role in cancer initiation and progression, and also has shown to impact patients’ clinical outcomes. We first evaluated 240 single nucleotide polymorphisms (SNPs) in miRNA biogenesis genes and predicted binding sites in NSCLC patients to determine associations with clinical outcomes in early-stage (stage I and II) and late-stage (stage III and IV) lung cancer patients, respectively. First, in 535 early-stage patients, after correcting multiple comparisons, FZD4:rs713065 (hazard ratio [HR]:0.46, 95% confidence interval [CI]:0.32-0.65) showed a significant inverse association with survival in early stage surgery-only patients. SP1:rs17695156 (HR:2.22, 95% CI:1.44-3.41) and DROSHA:rs6886834 (HR:6.38, 95% CI:2.49-16.31) conferred increased risk of progression in the all patients and surgery-only populations, respectively. FAS:rs2234978 was significantly associated with improved survival in all patients (HR:0.59, 95% CI:0.44-0.77) and in the surgery plus chemotherapy populations (HR:0.19, 95% CI:0.07-0.46).. Functional genomics analysis demonstrated that this variant creates a miR-651 binding site resulting in altered miRNA regulation of FAS, providing biological plausibility for the observed association. We then analyzed these associations in 598 late-stage patients. After multiple comparison corrections, no SNPs remained significant in the late stage group, while the top SNP NAT1:rs15561 (HR=1.98, 96%CI=1.32-2.94) conferred a significantly increased risk of death in the chemotherapy subgroup. To test the hypothesis that genetic variants in the inflammation-related pathways may be associated with survival in NSCLC patients, we first conducted a three-stage study. In the discovery phase, we investigated a comprehensive panel of 11,930 inflammation-related SNPs in three independent lung cancer populations. A missense SNP (rs2071554) in HLA-DOB was significantly associated with poor survival in the discovery population (HR: 1.46, 95% CI: 1.02-2.09), internal validation population (HR: 1.51, 95% CI: 1.02-2.25), and external validation (HR: 1.52, 95% CI: 1.01-2.29) population. Rs2900420 in KLRK1 was significantly associated with a reduced risk for death in the discovery (HR: 0.76, 95% CI: 0.60-0.96) and internal validation (HR: 0.77, 95% CI: 0.61-0.99) populations, and the association reached borderline significance in the external validation population (HR: 0.80, 95% CI: 0.63-1.02). We also evaluated these inflammation-related SNPs in NSCLC patients in never smokers. Lung cancer in never smokers has been increasingly recognized as distinct disease from that in ever-smokers. A two-stage study was performed using a discovery population from MD Anderson (411 patients) and a validation population from Mayo Clinic (311 patients). Three SNPs (IL17RA:rs879576, BMP8A:rs698141, and STK:rs290229) that were significantly associated with survival were validated (pCD74:rs1056400 and CD38:rs10805347) were borderline significant (p=0.08) in the Mayo Clinic population. In the combined analysis, IL17RA:rs879576 resulted in a 40% reduction in the risk for death (p=4.1 × 10-5 [p=0.61, heterogeneity test]). We also validated a survival tree created in MD Anderson population in the Mayo Clinic population. In conclusion, our results provided strong evidence that genetic variations in specific pathways that examined (miRNA and inflammation pathways) influenced clinical outcomes in NSCLC patients, and with further functional studies, the novel loci have potential to be translated into clinical use.