An empirical evaluation of the Random Forests classifier models for variable selection in a large-scale lung cancer case-control study

Random Forests™ is reported to be one of the most accurate classification algorithms in complex data analysis. It shows excellent performance even when most predictors are noisy and the number of variables is much larger than the number of observations. In this thesis Random Forests was applied to a large-scale lung cancer case-control study. A novel way of automatically selecting prognostic factors was proposed. Also, synthetic positive control was used to validate Random Forests method. Throughout this study we showed that Random Forests can deal with large number of weak input variables without overfitting. It can account for non-additive interactions between these input variables. Random Forests can also be used for variable selection without being adversely affected by collinearities. ^ Random Forests can deal with the large-scale data sets without rigorous data preprocessing. It has robust variable importance ranking measure. Proposed is a novel variable selection method in context of Random Forests that uses the data noise level as the cut-off value to determine the subset of the important predictors. This new approach enhanced the ability of the Random Forests algorithm to automatically identify important predictors for complex data. The cut-off value can also be adjusted based on the results of the synthetic positive control experiments. ^ When the data set had high variables to observations ratio, Random Forests complemented the established logistic regression. This study suggested that Random Forests is recommended for such high dimensionality data. One can use Random Forests to select the important variables and then use logistic regression or Random Forests itself to estimate the effect size of the predictors and to classify new observations. ^ We also found that the mean decrease of accuracy is a more reliable variable ranking measurement than mean decrease of Gini. ^

A case-control study of tea/coffee consumption and lung cancer risk

Background. Despite the increasing attention to the effects of dietary factors on lung cancer risk, epidemiological research on the role of black/green tea and coffee intake and lung cancer risk is scarce. The purpose of this study was to explore the following three hypotheses: (1) the preventive (protective) effect from lung cancer is higher in green tea than in black tea and coffee consumption. (2) brewed tea (either black or green) daily drinkers have lower odds of lung cancer than non-drinkers of brewed tea (3) regular black and green tea have more preventive effect against lung cancer than decaffeinated teas due to the synergistic effect of caffeine and other tea components. ^ Methods. Data on 1,088 lung cancer cases and 1,127 controls from an ongoing epidemiological study of lung cancer by the Department of Epidemiology of the University of Texas M.D. Anderson Cancer were analyzed. Multiple logistic regressions were performed for testing associations between frequency of specific types of tea/coffee consumption and the risk of lung cancer. ^ Results. We observed that more than a cup a week of green tea and decaffeinated black tea were significantly associated with reduced odds of lung cancer by 64% for green tea (adjusted OR = 0.44; 95% CI = 0.31–0.64), 36% for decaffeinated black tea (OR = 0.64; 95% CI = 0.45–0.90), when compared with non-drinkers and those who drank less than a cup a week. On the other hand, increasing intake of regular coffee (more than 3 cups a day) was associated with a 30% higher odds ratio of lung cancer (OR = 1.30; 95% CI = 1.01–1.09). No association was found between regular black tea, decaffeinated coffee consumption and the odds ratio of lung cancer. However, when drinkers of other tea/coffee beverages were excluded from each model in order to explore the independent effect of each type of tea/coffee, green tea and decaffeinated black tea-lung cancer associations remained but no association was observed for drinkers of regular coffee. ^ Conclusion. We report the chemopreventive effects of more than a cup a week of green tea and decaffeinated black tea on lung cancer. ^

The association between smoking marijuana and the occurrence of lung cancer

It is still unclear if an association exists between the smoking of marijuana and the occurrence of lung cancer although one clearly exists between cigarette smoking and lung cancer. The objective of this study was to systematically review the literature in order to assess the impact of marijuana smoking, which is increasingly becoming a significant public health issue, on the occurrence of the number one killing cancer, lung cancer. ^ Method. Selected studies in the English language conducted on humans that assess the impact of marijuana smoking on lung cancer were identified from EBSCO MEDLINE, PUBMED, and GOOGLE databases. The search keywords were marijuana, cannabis, hashish, kif, and lung cancer with selection criteria including studies in the English language identified from 1950 to April 2008. Excluded were non-research studies such as editorials, letters, and reviews. Also excluded were studies that did not involve humans with direct intentional marijuana smoking or in vivo studies with mice. Case report studies or case series studies involving less than 10 patients were also excluded as well as studies that did not examine lung conditions related to premalignant or cancerous changes. ^ Results. Ten studies met the selection criteria and were analyzed. The ten studies in this review overall offer biological evidence of the potential association between marijuana smoking and premalignant lung findings but no overwhelming conclusive evidence for the association between marijuana smoking and the occurrence of lung cancer. Two of the observational studies [1, 2] failed to demonstrate an association between marijuana and lung cancer, but the remaining studies supported an association between marijuana smoking and premalignant or malignant lung findings. ^ Conclusion. It must, therefore, be concluded that no convincing evidence exists, based on the existing data, for an association between marijuana smoking and the occurrence of lung cancer even though the few observational studies failing to report such an association may be due to certain limitations particularly the relative young age of the participants precluding sufficient lag time for the identification of lung cancer outcome as explained in other sections of this paper. ^ Further research is, therefore, necessary to better evaluate this critical issue, while recommendations against smoking marijuana because of its potential harmful effects, including the potential for premalignant lung changes as noted in this review, should continue to be made. ^ In the future, large prospective studies with study participants representing a much wider spectrum of ages, and longer follow-up periods, with detailed assessment of marijuana exposure and definitive pathologic diagnosis of lung cancer are necessary^

Hormone therapy and physical activity as predictors of lung cancer risk and survival: The California Teachers Study

Cigarette smoking is responsible for the majority of lung cancer cases worldwide; however, a proportion of never smokers still develop lung cancer over their lifetime, prompting investigation into additional factors that may modify lung cancer incidence, as well as mortality. Although hormone therapy (HT), physical activity (PA), and lung cancer have been previously examined, the associations remain unclear. This study investigated exposure to HT and PA that may modulate underlying mechanisms of lung cancer etiology and progression among women by using existing, de-identified data from the California Teachers Study (CTS).^ The CTS cohort, established in 1995–1996, has 133,479 active and retired female teachers and administrators, recruited through the California State Teachers Retirement System, and followed annually for cancer diagnosis, death, and change of address. Each woman enrolled in the CTS returned a questionnaire covering a wide variety of issues related to cancer risk and women's health, including recent and past HT use and physical activity, as well as active and environmental cigarette smoke exposure. Complete data to assess the associations between HT and lung cancer risk and survival were available for 60,592 postmenopausal women. Between 1995 and 2007, 727 of these women were diagnosed with invasive lung cancer; 441 of these died. Complete data to assess the associations between PA and lung cancer risk and survival were available for 118,513 women. Between 1995 and 2007, 853 of these women were diagnosed with invasive lung cancer; 516 of these died.^ After careful adjustment for smoking habits and other potential confounders, no measure of HT use was associated with lung cancer risk; however, any HT use (vs. no use) was associated with a decrease in lung-cancer-specific mortality. Specifically, among women who only used estrogen (E-only), decreases in lung cancer mortality were seen for recent use, but not for former use; no association was observed for estrogen plus progestin (E+P). Furthermore, among former users of HT, a statistically significant decrease in lung cancer mortality was observed for E-only use within 5 years prior to baseline, but not for E-only use >5 years prior to baseline. Neither long-term recreational PA nor recent recreational PA alone were associated with lung cancer risk; however, among women with a BMI<25 and ever smokers, high long-term moderate+strenuous PA was associated with a decrease in lung cancer risk. Women with non-local disease showed a decrease in lung cancer mortality associated with increasing duration of strenuous long-term activity, and 1.50-3.00 h/wk/y of recent moderate or recent strenuous PA. Long-term moderate PA was associated with decreased lung cancer mortality in never smokers, whereas recent moderate PA was associated with increased lung cancer mortality in current smokers. ^ Placing our findings in the context of the current literature, HT does not appear to be associated with lung cancer risk and previous studies reporting a protective effect of HT use on lung cancer risk may be subject to residual confounding by smoking. Looking at our findings regarding PA overall, the evidence still remains inconclusive regarding whether or not physical activity influence lung cancer risk or mortality. Our results suggest that recreational PA may associated with decreased lung cancer risk among women with BMI<25 and ever smoking-women; however, residual confounding by smoking should be strongly considered. To our knowledge, this is the first study to investigate lifetime recreational PA and lung cancer mortality among women. Our results contribute to the growing body of knowledge regarding non-smoking-related risk factors for lung cancer incidence and mortality among women. Given the potential clinical and interventional significance, further study and validation of these findings is warranted.^

Genetic susceptibility, cigarette smoking, and wood dust exposure in lung cancer: Case control analyses

This case control study was conducted to assess the association between lung cancer risk, mutagen sensitivity (a marker of cancer susceptibility), and a putative lung carcinogen, wood dust exposure. There were 165 cases (98 African-Americans, 67 Mexican-Americans) with newly diagnosed, previously untreated lung cancer, and 239 controls, frequency-matched on age, sex, and ethnicity.^ Mutagen sensitivity ($\ge$1 break/cell) was associated with a statistically significant elevated risk for lung cancer (odds ratio (OR) = 4.1, 95% confidence limits (CL) = 2.3,7.2). Wood dust exposure was also a significant predictor of risk (OR = 2.8, 95% CL = 1.2,6.6) after controlling for smoking and mutagen sensitivity. When stratified by ethnicity, wood dust exposure was a significant risk factor for African-Americans (OR = 4.0, 95% CL = 1.4,11.5), but not for Mexican-Americans (OR = 1.5, 95% CL = 0.3,7.1). Stratified analysis suggested a greater than multiplicative interaction between wood dust exposure and both mutagen sensitivity and smoking.^ The cases had significantly more breaks on chromosomes 4 and 5 than the controls did with ORs of 4.9 (95% CL = 2.0, 11.7) and 3.9 (95% CL = 1.6, 9.3), respectively. Breaks at 4p14, 4q27, 4q31, 5q21-22, 5q31, and 5q33 were significantly more common in lung cancer patients than in controls. Lung cancer risk had a dose-response relationship with breaks on chromosomes 4 and 5. Cigarette smoking had a strong interaction with breaks on chromosomes 2, 4, and 5.^ In a molecular cytogenetic study, using chromosome painting and G-banding, we showed that: (1) the proportion of chromosome 5 abnormalities surviving as chromosome-type aberrations remained significantly higher in cells of lung cancer cases (14%) than in controls (5%) (P $<$ 0.001). However, no significant differences were detected in chromosome 4 abnormalities between cases and controls; (2) the proportion of chromosome 5q13-22 abnormalities was 5.3% in the cases and 0.7% in the controls (P $<$ 0.001). 5q13-22 regions represented 40% of all abnormalities on chromosome 5 in the cases and only 14% in the controls.^ This study suggests that mutagen sensitivity, wood dust exposure, and cigarette smoking were independent risk factors for lung cancer, and the susceptibility of particular chromosome loci to mutagenic damage may be a genetic marker for specific types of lung cancer. ^

LUNG CANCER AMONG TEXAS MALE OIL REFINERY WORKERS, 1976 - 1980 (CASE-CONTROL STUDY)

The apparent excess in the rate of lung cancer among the population of Texas coastal counties has prompted this study. The main objective was to assess the risk of lung cancer among employees of oil refining industries. Data collected by UTSPH and NCI were used for this research. A non significant overall low risk of lung cancer was observed among workers ever employed in oil refining (COR = 0.84). A lower but not significant risk of lung cancer was detected among the same workers when classified by their usual industry of employment (COR = 0.77). An overall non significant crosstime decline in the risk of lung cancer was observed among most of the occupational groups within the oil refining industry, with the exception of professional/technical and clerical/sales occupations where a non significant crosstime increase in the risk was observed. ^

SABR for early-stage lung cancer: Early-adoption and future directions

Early-stage lung cancer incidence among older adults is expected to increase due to demographic trends and CT-based screening, yet optimal treatment of lung cancer in the elderly remains controversial. There are several accepted strategies for treating lung cancer including surgery, conventional radiation, and stereotactic ablative body radiotherapy (SABR). However, there are currently no randomized controlled trials to help distinguish the comparative effectiveness of these various strategies. This is an unfortunate omission as lung cancer causes the most deaths among all cancers in the United States (as well as the entire world). SABR holds particular promise as it is a completely non-invasive, ambulatory technique for achieving cure without an operation, thus avoiding the risks of surgery and the associated pre-operative and post-operative costs. To provide fair view of the potential effect on SABR on controlling lung cancer in the United States, a systematic review of SABR with a focus on its achieved outcomes, toxicities, and comparison to conventional radiation and surgical options is presented. ^

Non-p53 p53RE binding protein, a human transcription factor functionally analogous to P53

The transactivation activity of the p53 tumor suppressor protein is critical for regulating cell growth and apoptosis. We describe the identification of a transcription factor that is functionally similar to p53 and contains the same DNA binding and transcription activities specific for the p53 responsive DNA element (p53RE). This protein was highly purified through chromatography from HeLa cell extracts. The purified protein was able to bind specifically to the p53RE derived from a p21waf1 promoter and to stimulate p53RE-dependent transcription but not basal transcription in vitro. Its DNA-binding activity was inhibited by the wild type but not mutant p53RE-containing DNA oligomers. Also, this p53RE-binding activity was found in human p53 null Saos-2 osteosarcoma and H1299 small cell lung carcinoma cells. Interestingly, this activity exhibited a p53RE sequence preference that was distinct from the p53 protein. The activity is neither p53 nor p73, because anti-p53 or anti-73 antibodies were unable to detect this purified protein nor were the antibodies able to alter the p53-like activity, the p53RE-protein complex. These results demonstrate that, besides p73, an additional p53-like protein exists in cells, which is named NBP for non-p53, p53RE binding protein.

A dominant form of inherited retinal degeneration caused by a non-photoreceptor cell-specific mutation

We have isolated a dominant mutation, night blindness a (nba), that causes a slow retinal degeneration in zebrafish. Heterozygous nba fish have normal vision through 2–3 months of age but subsequently become night blind. By 9.5 months of age, visual sensitivity of affected fish may be decreased more than two log units, or 100-fold, as measured behaviorally. Electroretinographic (ERG) thresholds of mutant fish are also raised significantly, and the ERG b-wave shows a delayed implicit time. These defects are due primarily to a late-onset photoreceptor cell degeneration involving initially the rods but eventually the cones as well. Homozygous nba fish display an early-onset neuronal degeneration throughout the retina and elsewhere in the central nervous system. As a result, animals develop with small eyes and die by 4–5 days postfertilization (pf). These latter data indicate that the mutation affecting nba fish is not in a photoreceptor cell-specific gene.

Statistical modeling and projections of lung cancer mortality in 4 industrialized countries

The purpose of this work was to model lung cancer mortality as a function of past exposure to tobacco and to forecast age-sex-specific lung cancer mortality rates. A 3-factor age-period-cohort (APC) model, in which the period variable is replaced by the product of average tar content and adult tobacco consumption per capita, was estimated for the US, UK, Canada and Australia by the maximum likelihood method. Age- and sex-specific tobacco consumption was estimated from historical data on smoking prevalence and total tobacco consumption. Lung cancer mortality was derived from vital registration records. Future tobacco consumption, tar content and the cohort parameter were projected by autoregressive moving average (ARIMA) estimation. The optimal exposure variable was found to be the product of average tar content and adult cigarette consumption per capita, lagged for 2530 years for both males and females in all 4 countries. The coefficient of the product of average tar content and tobacco consumption per capita differs by age and sex. In all models, there was a statistically significant difference in the coefficient of the period variable by sex. In all countries, male age-standardized lung cancer mortality rates peaked in the 1980s and declined thereafter. Female mortality rates are projected to peak in the first decade of this century. The multiplicative models of age, tobacco exposure and cohort fit the observed data between 1950 and 1999 reasonably well, and time-series models yield plausible past trends of relevant variables. Despite a significant reduction in tobacco consumption and average tar content of cigarettes sold over the past few decades, the effect on lung cancer mortality is affected by the time lag between exposure and established disease. As a result, the burden of lung cancer among females is only just reaching, or soon will reach, its peak but has been declining for I to 2 decades in men. Future sex differences in lung cancer mortality are likely to be greater in North America than Australia and the UK due to differences in exposure patterns between the sexes. (c) 2005 Wiley-Liss, Inc.

Secure Full-Duplex Small-Cell Networks in a Spectrum Sharing Environment

In this paper, we propose three relay selection schemes for full-duplex heterogeneous networks in the presence of multiple cognitive radio eavesdroppers. In this setup, the cognitive small-cell nodes (secondary network) can share the spectrum licensed to the macro-cell system (primary network) on the condition that the quality-of-service of the primary network is always satisfied subjected to its outage probability constraint. The messages are delivered from one small-cell base station to the destination with the help of full-duplex small-cell base stations, which act as relay nodes. Based on the availability of the network’s channel state information at the secondary information source, three different selection criteria for full-duplex relays, namely: 1) partial relay selection; 2) optimal relay selection; and 3) minimal self-interference relay selection, are proposed. We derive the exact closed-form and asymptotic expressions of the secrecy outage probability for the three criteria under the attack of non-colluding/colluding eavesdroppers. We demonstrate that the optimal relay selection scheme outperforms the partial relay selection and minimal self-interference relay selection schemes at the expense of acquiring full channel state information knowledge. In addition, increasing the number of the full-duplex small-cell base stations can improve the security performance. At the illegitimate side, deploying colluding eavesdroppers and increasing the number of eavesdroppers put the confidential information at a greater risk. Besides, the transmit power and the desire outage probability of the primary network have great influences on the secrecy outage probability of the secondary network. 

Systems Modeling of Lung Cancer Screening Programs to Improve Quality

Thesis (Master's)--University of Washington, 2016-08

Childhood Non-Langerhans Cell Histiocytosis – An Atypical Case

Introduction: Histiocytoses correspond to a heterogeneous group of disorders characterized by the proliferation andinfiltration of histiocytes in tissues. For years, many of the histiocytoses where known by numerous names, reflecting the lack of understanding regarding their origin. Case report: A previously healthy, 4-month-old infant presented a cutaneous lesion of progressive growth on his nose since 1 month ago. The mother has a diagnosis of Sjögren syndrome. On examination, there was a firm well-defined erythematous plaque with 30 mm of diameter and raised edges, on the bridge of the nose. At the left cheek, left thoracic region and right arm there were also small erythematous papules with 5 days of evolution. The analytical study showed elevated antinuclear antibodies and anti-SS-A. Histopathological examination confirmed the diagnosis of non-Langerhans cells histiocytosis. Progressively, more annular erythematous papules and plaques appeared scattered over the face, trunk and limbs with spontaneous resolution of the first lesion on the nose. About 2 months after the initial evaluation, all lesions had regressed spontaneously, with normalization of analytical study at one year-old. Conclusion: Many authors believe that the different subtypes of histiocytosis represent only different stages of the same disease. Immunophenotyping of histiocytosis does not always identify the subtype involved, making the occurrence of atypical cases frequent.

Exposure to Stressful Life Events and Lung Cancer Risk

Objectif: Examiner l’association entre l'exposition aux évènements stressants de la vie et le cancer du poumon. Méthodes: Les données proviennent d’une étude cas-témoins, menée chez les hommes et les femmes vivant dans la région métropolitaine de Montréal entre 1996 et 2001. Le cancer du poumon d’un cas éligible devait être confirmé histologiquement à l’un des 18 hôpitaux de cette région. Les témoins ont été sélectionnés aléatoirement de la liste électorale du Québec et ont été appariés au cas par fréquence de groupes d'âge et par sexe. Un questionnaire a été administré en entrevue pour recueillir les données, dont l’évaluation de huit évènements stressants de la vie par le participant. Si le participant avait vécu un évènement stressant ciblé durant les six dernières années, il devait aussi coter cet évènement sur une échelle de trois points. La régression logistique non conditionnelle a été utilisée pour estimer les rapports de cotes ainsi que leurs intervalles de confiance à 95%. Des analyses par sexe, niveau de tabagisme et par type histologique ont été réalisées. Nous avons aussi analysé l’association entre le cancer du poumon et le nombre total d'évènements, les évènements de perte et les évènements socioéconomiques, ainsi que chaque évènement individuellement. Les analyses des scores d'impact autoévalués et avec un score externe de perception, ont également été menées. Résultats: La population de ce projet comprend 1061 cas et 1422 témoins, âgés de 35 à 70 ans. Les participants inclus avaient répondu aux sections du questionnaire portant sur les facteurs de style de vie et sur l'historique de tabagisme. Dans l'ensemble, nous n’avons pas observé d’association entre le cancer du poumon et l'exposition aux évènements stressants de la vie. Nous avons observé une diminution du risque pour les évènements socioéconomiques autoévalués comme peu stressants (RC=0,50; IC 95%= 0,31 - 0,81). Conclusion: Nos résultats suggèrent que les évènements socioéconomiques sont associées à un risque réduit si ces évènements sont considérés comme peu stressant.