Radiomics analysis to predict pulmonary nodule malignancy using machine learning approaches.

BACKGROUND: Low-dose CT screening can reduce lung cancer-related mortality. However, most screen-detected pulmonary abnormalities do not develop into cancer and it often remains challenging to identify malignant nodules, particularly among indeterminate nodules. We aimed to develop and assess prediction models based on radiological features to discriminate between benign and malignant pulmonary lesions detected on a baseline screen. METHODS: Using four international lung cancer screening studies, we extracted 2060 radiomic features for each of 16 797 nodules (513 malignant) among 6865 participants. After filtering out low-quality radiomic features, 642 radiomic and 9 epidemiological features remained for model development. We used cross-validation and grid search to assess three machine learning (ML) models (eXtreme Gradient Boosted Trees, random forest, least absolute shrinkage and selection operator (LASSO)) for their ability to accurately predict risk of malignancy for pulmonary nodules. We report model performance based on the area under the curve (AUC) and calibration metrics in the held-out test set. RESULTS: The LASSO model yielded the best predictive performance in cross-validation and was fit in the full training set based on optimised hyperparameters. Our radiomics model had a test-set AUC of 0.93 (95% CI 0.90 to 0.96) and outperformed the established Pan-Canadian Early Detection of Lung Cancer model (AUC 0.87, 95% CI 0.85 to 0.89) for nodule assessment. Our model performed well among both solid (AUC 0.93, 95% CI 0.89 to 0.97) and subsolid nodules (AUC 0.91, 95% CI 0.85 to 0.95). CONCLUSIONS: We developed highly accurate ML models based on radiomic and epidemiological features from four international lung cancer screening studies that may be suitable for assessing indeterminate screen-detected pulmonary nodules for risk of malignancy.

Kindergarten-age neurocognitive, functional, and quality-of-life outcomes after liver transplantation at under 6 years of age.

BACKGROUND: We aimed to describe school-entry age neurocognitive, functional, and HRQL outcomes and their predictors after liver transplant done at age <6 years. METHODS: A prospective cohort of all (n = 69) children surviving liver transplant from 1999 to 2014 were assessed at age 55.4 (SD 7.2) months and 38.6 (12.4) months after transplant. Assessment included: the Wechsler Preschool and Primary Scales of Intelligence, Beery-Buktenica Developmental Test of VMI, Adaptive Behavior Assessment System caregiver-completed questionnaire, and PedsQL 4.0 Generic Core Scales. Univariate and multiple linear regression determined predictors of outcomes at P < .05. RESULTS: Neurocognitive and functional outcomes were on average within 1 SD of population norms, although shifted to the left (P ≤ .03), with more patients than expected having scores >2 (3.7-5.9 times more, P ≤ .007) SD below population norms. Total and Summary HRQL scores were statistically significantly lower than the healthy normative population (P ≤ .02) and a congenital heart disease group (P ≤ .02), but similar to children with other chronic health conditions; differences often exceeded the MCID and were lowest in the School functioning domain. There were few predictors on multiple linear regressions, and we could not confirm previous studies that suggested various inconsistent predictors of outcomes. Neurocognitive and functional outcomes scores were highly correlated with HRQL scores except for the School functioning domain, but did not fully explain them. CONCLUSIONS: Long-term follow-up of this vulnerable population is important in order to facilitate support for the patient and family, and early intervention for any difficulties identified.

Longitudinal linear combination test for gene set analysis.

BACKGROUND: Although microarray studies have greatly contributed to recent genetic advances, lack of replication has been a continuing concern in this area. Complex study designs have the potential to address this concern, though they remain undervalued by investigators due to the lack of proper analysis methods. The primary challenge in the analysis of complex microarray study data is handling the correlation structure within data while also dealing with the combination of large number of genetic measurements and small number of subjects that are ubiquitous even in standard microarray studies. Motivated by the lack of available methods for analysis of repeatedly measured phenotypic or transcriptomic data, herein we develop a longitudinal linear combination test (LLCT). RESULTS: LLCT is a two-step method to analyze multiple longitudinal phenotypes when there is high dimensionality in response and/or explanatory variables. Alternating between calculating within-subjects and between-subjects variations in two steps, LLCT examines if the maximum possible correlation between a linear combination of the time trends and a linear combination of the predictors given by the gene expressions is statistically significant. A generalization of this method can handle family-based study designs when the subjects are not independent. This method is also applicable to time-course microarray, with the ability to identify gene sets that exhibit significantly different expression patterns over time. Based on the results from a simulation study, LLCT outperformed its alternative: pathway analysis via regression. LLCT was shown to be very powerful in the analysis of large gene sets even when the sample size is small. CONCLUSIONS: This self-contained pathway analysis method is applicable to a wide range of longitudinal genomics, proteomics, metabolomics (OMICS) data, allows adjusting for potentially time-dependent covariates and works well with unbalanced and incomplete data. An important potential application of this method could be time-course linkage of OMICS, an attractive possibility for future genetic researchers. AVAILABILITY: R package of LLCT is available at: https://github.com/its-likeli-jeff/LLCT.

Avoiding Furosemide Ototoxicity Associated With Single-Ventricle Repair in Young Infants.

OBJECTIVE: To reduce bilateral delayed-onset progressive sensory permanent hearing loss using a systems-wide quality improvement project with adherence to best practice for the administration of furosemide. DESIGN: Prospective cohort study with regular audiologic follow-up assessment of survivors both before and after a 2007-2008 quality improvement practice change. SETTING: The referral center in Western Canada for complex cardiac surgery, with comprehensive multidisciplinary follow-up by the Complex Pediatric Therapies Follow-up Program. PATIENTS: All consecutive patients having single-ventricle palliative cardiac surgery at age 6 weeks old or younger. INTERVENTIONS: A 2007-2008 quality improvement practice change consisted of a Parenteral Drug Monograph revision indicating slow IV administration of furosemide, an educational program, and an evaluation. MEASUREMENTS AND MAIN RESULTS: The outcome measure was the prevalence of permanent hearing loss by 4 years old. Firth multiple logistic regression compared pre (1996-2008) to post (2008-2012) practice change occurrence of permanent hearing loss, adjusting for confounding variables, including all hospital days, extracorporeal membrane oxygenation, cardiopulmonary bypass time, age at first surgery, dialysis, and sepsis. From 1996 to 2012, 259 infants had single-ventricle palliative surgery at age 6 weeks old or younger, with 173 (64%) surviving to age 4 years. Of survivors, 106 (61%) were male, age at surgery was 11.6 days (9.0 d), and total hospitalization days by age 4 years were 64 (42); 18 (10%) had cardiopulmonary resuscitation and 38 (22%) had sepsis at any time. All 173 (100%) had 4-year follow-up. Pre- to postpractice change permanent hearing loss dropped from 17/100 (17%) to 0/73 (0%) of survivors. On Firth multiple logistic regression, the only variable statistically associated with permanent hearing loss was the pre- to postpractice change time period (odds ratio, 0.03; 95% CI, 0-0.35; p = 0.001). CONCLUSIONS: A practice change to ensure slow IV administration of furosemide eliminated permanent hearing loss. Centers caring for critically ill infants, particularly those with single-ventricle anatomy or hypoxia, should review their drug administration guidelines and adhere to best practice for administration of IV furosemide.

The impact of rural health system reform on hospitalization rates in the Islamic Republic of Iran: an interrupted time series.

OBJECTIVE: To assess the effects on hospital utilization rates of a major health system reform - a family physician programme and a social protection scheme - undertaken in rural areas of the Islamic Republic of Iran in 2005. METHODS: A "tracer" province that was not a patient referral hub was selected for the collection of monthly hospitalization data over a period of about 10 years, beginning two years before the rural health system reform (the "intervention") began. An interrupted time series analysis was conducted and segmented regression analysis was used to assess the immediate and gradual effects of the intervention on hospitalization rates in an intervention group composed of rural residents and a comparison group composed of urban residents primarily. FINDINGS: Before the intervention, the hospitalization rate in the rural population was significantly lower than in the comparison group. Although there was no significant increase or decline in hospitalization rates in the intervention or comparison group before the intervention, after the intervention a significant increase in the hospitalization rate - of 4.6 hospitalizations per 100 000 insured persons per month on average - was noted in the intervention group (P < 0.001). The monthly increase in the hospitalization rate continued for over a year and stabilized thereafter. No increase in the hospitalization rate was observed in the comparison group. CONCLUSION: The primary health-care programme instituted as part of the health system reform process has increased access to hospital care in a population that formerly underutilized hospital services. It has not reduced hospitalizations or hospitalization-related expenditure.

Circulating proteome for pulmonary nodule malignancy.

BACKGROUND: Although lung cancer screening with low-dose computed tomography is rolling out in many areas of the world, differentiating indeterminate pulmonary nodules remains a major challenge. We conducted one of the first systematic investigations of circulating protein markers to differentiate malignant from benign screen-detected pulmonary nodules. METHODS: Based on 4 international low-dose computed tomography screening studies, we assayed 1078 protein markers using prediagnostic blood samples from 1253 participants based on a nested case-control design. Protein markers were measured using proximity extension assays, and data were analyzed using multivariable logistic regression, random forest, and penalized regressions. Protein burden scores (PBSs) for overall nodule malignancy and imminent tumors were estimated. RESULTS: We identified 36 potentially informative circulating protein markers differentiating malignant from benign nodules, representing a tightly connected biological network. Ten markers were found to be particularly relevant for imminent lung cancer diagnoses within 1 year. Increases in PBSs for overall nodule malignancy and imminent tumors by 1 standard deviation were associated with odds ratios of 2.29 (95% confidence interval: 1.95 to 2.72) and 2.81 (95% confidence interval: 2.27 to 3.54) for nodule malignancy overall and within 1 year of diagnosis, respectively. Both PBSs for overall nodule malignancy and for imminent tumors were substantially higher for those with malignant nodules than for those with benign nodules, even when limited to Lung Computed Tomography Screening Reporting and Data System (LungRADS) category 4 (P < .001). CONCLUSIONS: Circulating protein markers can help differentiate malignant from benign pulmonary nodules. Validation with an independent computed tomographic screening study will be required before clinical implementation.

Developing a Socioeconomic Status Index for Chronic Disease Prevention Research in Canada.

Capturing socioeconomic inequalities in relation to chronic disease is challenging since socioeconomic status (SES) encompasses many aspects. We constructed a comprehensive individual-level SES index based on a broad set of social and demographic indicators (gender, education, income adequacy, occupational prestige, employment status) and examined its relationship with smoking, a leading chronic disease risk factor. Analyses were based on baseline data from 17,371 participants of Alberta's Tomorrow Project (ATP), a prospective cohort of adults aged 35−69 years with no prior personal history of cancer. To construct the SES index, we used principal component analysis (PCA) and to illustrate its utility, we examined the association with smoking intensity and smoking history using multiple regression models, adjusted for age and gender. Two components were retained from PCA, which explained 61% of the variation. The SES index was best aligned with educational attainment and occupational prestige, and to a lesser extent, with income adequacy. In the multiple regression analysis, the SES index was negatively associated with smoking intensity (p < 0.001). Study findings highlight the potential of using individual-level SES indices constructed from a broad set of social and demographic indicators in epidemiological research.

Considerations of biomarker application for cancer continuum in the era of precision medicine.

Purpose of the review: The goal of this review is to highlight emerging biomarker research by the key phases of the cancer continuum and outline the methodological considerations for biomarker application. Recent findings: While biomarkers have an established role in targeted therapy and to some extent, disease monitoring, their role in early detection and survivorship remains to be elucidated. With the advent of omics technology, the discovery of biomarkers has been accelerated exponentially, therefore careful consideration to ensure an unbiased study design and robust validity is crucial. Summary: The rigor of biomarker research holds the key to the success of precision health care. The potential clinical utility and the feasibility of implementation should be central to future biomarker research study design.

Early-childhood cytomegalovirus infection and children's neurocognitive development.

BACKGROUND: Despite a clear association seen in congenitally infected children, the effect of postnatal cytomegalovirus (CMV) infection during early childhood on cognitive development has not yet been determined. METHODS: CMV-infection status was obtained based on serological measurements when children were 7 years old. Using population-based longitudinal data, we employed multivariate Poisson regression with a robust variance estimator to characterize the relationship between childhood CMV infection and adverse neurocognitive outcomes in children. Suboptimal neurocognitive outcomes were compared between CMV-positive and CMV-negative children using various cognitive assessments from 8 to 15 years of age. Children were evaluated on the cognitive domains of language, reading, memory and general intelligence, with a suboptimal score being >2 standard deviations lower than the mean score. Approximate Bayes factor (ABF) analysis was used to determine the level of evidence for the observed associations. RESULTS: With adjustment for potential confounders, we observed that early-childhood CMV infection was associated with suboptimal total intelligence quotient (IQ) at 8 years of age [incidence-rate ratio (IRR) = 2.50, 95% confidence interval (CI) 1.35-4.62, ABF = 0.08], but not with suboptimal total IQ at 15 years of age (IRR = 0.97, 95% CI 0.43-2.19, ABF = 1.68). Suboptimal attentional control at 8 years (IRR = 1.74, 95% CI 1.13-2.68, ABF = 0.18) and reading comprehension at 9 years (IRR = 1.93, 95% CI 1.12-3.33, ABF = 0.24) were also associated with CMV infection. ABF analysis provided strong evidence for the association between CMV infection and total IQ at 8 years, and only anecdotal evidence for attentional control at 8 years and reading comprehension at 9 years. All other cognitive measures assessed were not associated with CMV infection. CONCLUSION: In this large-scale prospective cohort, we observed some evidence for adverse neurocognitive effects of postnatal CMV infection on general intelligence during early childhood, although not with lasting effect. If confirmed, these results could support the implementation of preventative measures to combat postnatal CMV infection.

Co-consumption of Vegetables and Fruit, Whole Grains, and Fiber Reduces the Cancer Risk of Red and Processed Meat in a Large Prospective Cohort of Adults from Alberta's Tomorrow Project.

We examined whether co-consumption of red and processed meat with key foods items and food constituents recommended for cancer prevention (vegetables and fruit, whole grains, and fiber) mitigates cancer incidence. In a prospective cohort of 26,218 adults aged 35-69 years at baseline, dietary intake was collected through 124-item past-year food frequency questionnaire. Incidence of all-cause and 15 cancers previously linked to red and processed meat intake was obtained through data linkage with a cancer registry (average follow-up 13.5 years). Competing risk Cox Proportional Hazard models estimated cancer risk and Accelerated Failure Time models estimated time-to-cancer occurrence for different combinations of intake levels while considering mortality from vital statistics and established confounders. Co-consumption of low vegetables and fruit intake with high processed meat was associated with higher incidence of all-cause and 15 cancers (men: HR = 1.85, 1.91; women: HR = 1.44, 1.49) and accelerated time-to-cancer occurrence (men: 6.5 and 7.1 years and women: 5.6 and 6.3 years, respectively), compared to high vegetables and fruit with low processed meat intake. Less pronounced and less consistent associations were observed for whole grains and fiber and for red meat. The findings provide initial evidence toward refining existing cancer prevention recommendations to optimize the intake and combination of foods in the general adult population.

Nicotine dependence as a risk factor for upper aerodigestive tract (UADT) cancers: A mediation analysis.

PURPOSE: This study investigated nicotine dependence as an independent risk factor for upper aerodigestive tract (UADT) cancers, including lung and head and neck cancers (HNC). The study aimed to isolate the direct effect of nicotine dependence, independent of tobacco smoking. METHODS: A case-control study with a total of 4957 participants was conducted in Ontario, Canada, of which 2964 categorized as either current or former smokers were used in the analysis. Nicotine dependence of ever-smokers (2360 UADT cases and 604 controls) was measured using the Fagerström Test for Nicotine Dependence. Using mediation analyses and adjusted logistic regression models, we decomposed the direct effect of nicotine dependence and the mediated effect of smoking duration to quantify the risks of lung and HNC. The role of human papillomavirus (HPV) and cancer subtypes were assessed. RESULTS: Most individual nicotine dependence behaviours showed positive associations with lung cancer with approximately 1.8 to 3.5-fold risk increase, and to lesser extent with 1.4 to 2.3-fold risk for HNC. Nicotine dependence is partially accountable for increased risks of lung cancer (OR = 1.20, 95%CI = 1.13-1.28) and HNC (1.12, 95%CI = 1.04-1.19). Nicotine dependence had a greater effect on the risk of HPV-negative oropharyngeal cancer (OR = 3.06, 95%CI = 1.65-5.66) in comparison to HPV-positive oropharyngeal cancer (OR = 1.05, 95%CI = 0.67-1.65). The direct effects of nicotine dependence remained significant after accounting for cumulative tobacco exposures. CONCLUSION: Nicotine dependence increases the risks of lung and HNC cancers after accounting for tobacco smoking, suggesting potential toxic effects of nicotine. These results are informative for the safety consideration of nicotine exposures.

Longitudinal analysis reveals early-pregnancy associations between perfluoroalkyl sulfonates and thyroid hormone status in a Canadian prospective birth cohort.

Serum perfluoroalkyl acids (PFAAs) have been linked to disruption of maternal thyroid hormone homeostasis, but results have varied between studies which we hypothesized was due to timing of the thyroid hormone measurements, variability in PFAA isomer patterns, or presence of other stressors. In a longitudinal study design, we investigated the time-dependency of associations between PFAA isomers and thyroid hormones during pregnancy and post-partum while considering thyroid peroxidase antibody (TPOAb) status and mercury (Hg) co-exposure. In participants of a prospective Canadian birth cohort (n = 494), free thyroxine (FT4), free triiodothyronine (FT3), thyroid stimulating hormone (TSH) and TPOAb were quantified in maternal plasma collected in each trimester and 3-months postpartum, and 25 PFAAs (15 linear and 10 branched) and Hg were quantified in samples collected during the second trimester. Perfluorohexane sulfonate (PFHxS) and total branched isomers of perfluorooctane sulfonate (PFOS) were positively associated with TSH in mixed-effect models, with strongest associations early in gestation. Throughout pregnancy and post-partum, PFHxS was inversely associated with FT4, consistent with elevated TSH, while Hg was inversely associated with FT3. In TPOAb-positive women, negative associations were found between PFUnA and FT4, and 1m-PFOS and TSH, supporting previous studies that thyroid disorder could increase susceptibility to PFAA-mediated hormone dysregulation. Hg did not confound associations but was a significant interaction term, revealing further positive associations between PFOS isomers (∑3m+4m-PFOS) and TSH. Higher perfluoroalkyl sulfonate exposures were associated with higher TSH and/or lower FT4, strongly suggestive that PFHxS and branched PFOS isomers are risk factors for subclinical maternal hypothyroidism. Isomer-specific analysis is important in future studies, as crude measures of 'total-PFOS' masked the associations of branched isomers. A concerning result was for PFHxS which had consistent negative associations with FT4 at all time points and a positive association with TSH in early pregnancy when fetal development is most sensitive to disruption.

Association between bivariate expression of key oncogenes and metabolic phenotypes of patients with prostate cancer.

BACKGROUND: AKT and MYC are two of the most prevalent oncogenes associated with prostate cancer. The precise effects of overexpression of these two key oncogenes on the regulation of metabolic pathways in prostate cancer are under active investigation; however, few studies have investigated their bivariate oncogene-pair expressions in metabolic prostate cancer phenotypes. This is primarily due to the lack of a suitable statistical method to analyze the data in the presence of oncogene interactions and within-metabolite-set correlations. METHODS: We analyzed data on the expressions of phosphorylated AKT1 and MYC and the concentrations of 228 metabolites from 60 human prostate tumor samples and 16 normal tissue samples. The metabolomic data allowed us to study not only the measurement of individual metabolites, which can exhibit a dynamic range, but the enriched phenotypes in terms of "metabolite sets" that come from known metabolic pathways. We studied 71 metabolite sets defined by KEGG annotation. We used a modification of linear combination test (LCT) for multiple continuous outcomes to find associations between metabolite sets and oncogenic expressions, after accounting for the correlation between AKT1 and MYC expressions and the correlation between metabolites in a metabolite set. The LCT performance was evaluated using a simulation study. RESULTS: Through a comprehensive analytical method, our study linked oncogenomics and metabolomics data from patients to improve our understanding of the interconnected mechanisms underlying prostate cancer. This study showed that dysregulations of AKT1 and MYC significantly alter the metabolic pathways activated by nonglucose nutrient sources and their downstream targets. Our findings highlighted the role of MYC as the leading, but not the only, oncogene in prostate oncogenesis. In our simulation study, the LCT performed better than the known alternative method, gene-set enrichment analysis (GSEA). CONCLUSIONS: Our study offers a solution for linking genomics and metabolomics, working directly with multiple continuous and correlated measurements.

Schroth Physiotherapeutic Scoliosis-Specific Exercises Added to the Standard of Care Lead to Better Cobb Angle Outcomes in Adolescents with Idiopathic Scoliosis - an Assessor and Statistician Blinded Randomized Controlled Trial.

BACKGROUND: The North American non-surgical standard of care for adolescent idiopathic scoliosis (AIS) includes observation and bracing, but not exercises. Schroth physiotherapeutic scoliosis-specific exercises (PSSE) showed promise in several studies of suboptimal methodology. The Scoliosis Research Society calls for rigorous studies supporting the role of exercises before including it as a treatment recommendation for scoliosis. OBJECTIVES: To determine the effect of a six-month Schroth PSSE intervention added to standard of care (Experimental group) on the Cobb angle compared to standard of care alone (Control group) in patients with AIS. METHODS: Fifty patients with AIS aged 10-18 years, with curves of 10°-45° and Risser grade 0-5 were recruited from a single pediatric scoliosis clinic and randomized to the Experimental or Control group. Outcomes included the change in the Cobb angles of the Largest Curve and Sum of Curves from baseline to six months. The intervention consisted of a 30-45 minute daily home program and weekly supervised sessions. Intention-to-treat and per protocol linear mixed effects model analyses are reported. RESULTS: In the intention-to-treat analysis, after six months, the Schroth group had significantly smaller Largest Curve than controls (-3.5°, 95% CI -1.1° to -5.9°, p = 0.006). Likewise, the between-group difference in the square root of the Sum of Curves was -0.40°, (95% CI -0.03° to -0.8°, p = 0.046), suggesting that an average patient with 51.2° at baseline, will have a 49.3° Sum of Curves at six months in the Schroth group, and 55.1° in the control group with the difference between groups increasing with severity. Per protocol analyses produced similar, but larger differences: Largest Curve = -4.1° (95% CI -1.7° to -6.5°, p = 0.002) and [Formula: see text] (95% CI -0.8 to 0.2, p = 0.006). CONCLUSION: Schroth PSSE added to the standard of care were superior compared to standard of care alone for reducing the curve severity in patients with AIS. TRIAL REGISTRATION: NCT01610908.