Genetic Evidence for a Causal Link between Intra-Pancreatic Fat Deposition and Pancreatitis: a Mendelian Randomization Study.

Introduction: Recent associative studies have linked intra-pancreatic fat deposition (IPFD) with risk of pancreatitis, but the causal relationship remains unclear. Methods: Utilizing Mendelian randomization, we evaluated the causal association between genetically predicted IPFD and pancreatitis. This approach utilized genetic variants from genome-wide association studies of IPFD (n=25,617), acute pancreatitis (n=6,787 cases/361,641 controls), and chronic pancreatitis (n=3,875 cases/361,641 controls). Results: Genetically predicted IPFD was significantly associated with acute pancreatitis (OR per 1-SD increase: 1.40[95%CI:1.12-1.76], p=0.0032) and chronic pancreatitis (OR:1.64[95%CI:1.13-2.39], p=0.0097). Discussion: Our findings support a causal role of IPFD in pancreatitis, suggesting that reducing IPFD could lower the risk of pancreatitis.

Genetically Determined Circulating Lactase/Phlorizin Hydrolase Concentrations and Risk of Colorectal Cancer: A Two-Sample Mendelian Randomization Study.

Previous research has found that milk is associated with a decreased risk of colorectal cancer (CRC). However, it is unclear whether the milk digestion by the enzyme lactase-phlorizin hydrolase (LPH) plays a role in CRC susceptibility. Our study aims to investigate the direct causal relationship of CRC risk with LPH levels by applying a two-sample Mendelian Randomization (MR) strategy. Genetic instruments for LPH were derived from the Fenland Study, and CRC-associated summary statistics for these instruments were extracted from the FinnGen Study, PLCO Atlas Project, and Pan-UK Biobank. Primary MR analyses focused on a cis-variant (rs4988235) for LPH levels, with results integrated via meta-analysis. MR analyses using all variants were also undertaken. This analytical approach was further extended to assess CRC subtypes (colon and rectal). Meta-analysis across the three datasets illustrated an inverse association between genetically predicted LPH levels and CRC risk (OR: 0.92 [95% CI, 0.89-0.95]). Subtype analyses revealed associations of elevated LPH levels with reduced risks for both colon (OR: 0.92 [95% CI, 0.89-0.96]) and rectal cancer (OR: 0.92 [95% CI, 0.87, 0.98]). Consistency was observed across varied analytical methods and datasets. Further exploration is warranted to unveil the underlying mechanisms and validate LPH's potential role in CRC prevention.

Epigenome-wide association study of total nicotine equivalents in multiethnic current smokers from three prospective cohorts.

The impact of tobacco exposure on health varies by race and ethnicity and is closely tied to internal nicotine dose, a marker of carcinogen uptake. DNA methylation is strongly responsive to smoking status and may mediate health effects, but study of associations with internal dose is limited. We performed a blood leukocyte epigenome-wide association study (EWAS) of urinary total nicotine equivalents (TNEs; a measure of nicotine uptake) and DNA methylation measured using the MethylationEPIC v1.0 BeadChip (EPIC) in six racial and ethnic groups across three cohort studies. In the Multiethnic Cohort Study (discovery, n = 1994), TNEs were associated with differential methylation at 408 CpG sites across >250 genomic regions (p < 9 × 10-8). The top significant sites were annotated to AHRR, F2RL3, RARA, GPR15, PRSS23, and 2q37.1, all of which had decreasing methylation with increasing TNEs. We identified 45 novel CpG sites, of which 42 were unique to the EPIC array and eight annotated to genes not previously linked with smoking-related DNA methylation. The most significant signal in a novel gene was cg03748458 in MIR383;SGCZ. Fifty-one of the 408 discovery sites were validated in the Singapore Chinese Health Study (n = 340) and the Southern Community Cohort Study (n = 394) (Bonferroni corrected p < 1.23 × 10-4). Significant heterogeneity by race and ethnicity was detected for CpG sites in MYO1G and CYTH1. Furthermore, TNEs significantly mediated the association between cigarettes per day and DNA methylation at 15 sites (average 22.5%-44.3% proportion mediated). Our multiethnic study highlights the transethnic and ethnic-specific methylation associations with internal nicotine dose, a strong predictor of smoking-related morbidities.

Evidence for a causal link between intra-pancreatic fat deposition and pancreatic cancer: A prospective cohort and Mendelian randomization study.

Prior observational studies suggest an association between intra-pancreatic fat deposition (IPFD) and pancreatic ductal adenocarcinoma (PDAC); however, the causal relationship is unclear. To elucidate causality, we conduct a prospective observational study using magnetic resonance imaging (MRI)-measured IPFD data and also perform a Mendelian randomization study using genetic instruments for IPFD. In the observational study, we use UK Biobank data (N = 29,463, median follow-up: 4.5 years) and find that high IPFD (>10%) is associated with PDAC risk (adjusted hazard ratio [HR]: 3.35, 95% confidence interval [95% CI]: 1.60-7.00). In the Mendelian randomization study, we leverage eight out of nine IPFD-associated genetic variants (p < 5 × 10-8) from a genome-wide association study in the UK Biobank (N = 25,617) and find that genetically determined IPFD is associated with PDAC (odds ratio [OR] per 1-standard deviation [SD] increase in IPFD: 2.46, 95% CI: 1.38-4.40) in the Pancreatic Cancer Cohort Consortium I, II, III (PanScan I-III)/Pancreatic Cancer Case-Control Consortium (PanC4) dataset (8,275 PDAC cases and 6,723 non-cases). This study provides evidence for a potential causal role of IPFD in the pathogenesis of PDAC. Thus, reducing IPFD may lower PDAC risk.

Genetic Evidence Causally Linking Pancreas Fat to Pancreatic Cancer: A Mendelian Randomization Study.

Background & Aims: Pancreatic ductal adenocarcinoma (PDAC) is highly lethal, and any clues to understanding its elusive etiology could lead to breakthroughs in prevention, early detection, or treatment. Observational studies have shown a relationship between pancreas fat accumulation and PDAC, but the causality of this link is unclear. We therefore investigated whether pancreas fat is causally associated with PDAC using two-sample Mendelian randomization. Methods: We leveraged eight genetic variants associated with pancreas fat (P<5×10 -8 ) from a genome-wide association study (GWAS) in the UK Biobank (25,617 individuals), and assessed their association with PDAC in the Pancreatic Cancer Cohort Consortium I-III and the Pancreatic Cancer Case-Control Consortium dataset (8,275 PDAC cases and 6,723 non-cases). Causality was assessed using the inverse-variance weighted method. Although none of these genetic variants were associated with body mass index (BMI) at genome-wide significance, we further conducted a sensitivity analysis excluding genetic variants with a nominal BMI association in GWAS summary statistics from the UK Biobank and the Genetic Investigation of Anthropometric Traits consortium dataset (806,834 individuals). Results: Genetically determined higher levels of pancreas fat using the eight genetic variants was associated with increased risk of PDAC. For one standard deviation increase in pancreas fat levels (i.e., 7.9% increase in pancreas fat fraction), the odds ratio of PDAC was 2.46 (95%CI:1.38-4.40, P=0.002). Similar results were obtained after excluding genetic variants nominally linked to BMI (odds ratio:3.79, 95%CI:1.66-8.65, P=0.002). Conclusions: This study provides genetic evidence for a causal role of pancreas fat in the pathogenesis of PDAC. Thus, reducing pancreas fat could lower the risk of PDAC.

The benefits and harms of adjuvant chemotherapy for non-small cell lung cancer in patients with major comorbidities: A simulation study.

BACKGROUND: Randomized controlled trials (RCTs) have demonstrated a survival benefit for adjuvant platinum-based chemotherapy after resection of locoregional non-small cell lung cancer (NSCLC). The relative benefits and harms and optimal approach to treatment for NSCLC patients who have major comorbidities (chronic obstructive pulmonary disease [COPD], coronary artery disease [CAD], and congestive heart failure [CHF]) are unclear, however. METHODS: We used a simulation model to run in-silico comparative trials of adjuvant chemotherapy versus observation in locoregional NSCLC in patients with comorbidities. The model estimated quality-adjusted life years (QALYs) gained by each treatment strategy stratified by age, comorbidity, and stage. The model was parameterized using outcomes and quality-of-life data from RCTs and primary analyses from large cancer databases. RESULTS: Adjuvant chemotherapy was associated with clinically significant QALY gains for all patient age/stage combinations with COPD except for patients >80 years old with Stage IB and IIA cancers. For patients with CHF and Stage IB and IIA disease, adjuvant chemotherapy was not advantageous; in contrast, it was associated with QALY gains for more advanced stages for younger patients with CHF. For stages IIB and IIIA NSCLC, most patient groups benefited from adjuvant chemotherapy. However, In general, patients with multiple comorbidities benefited less from adjuvant chemotherapy than those with single comorbidities and women with comorbidities in older age categories benefited more from adjuvant chemotherapy than their male counterparts. CONCLUSIONS: Older, multimorbid patients may derive QALY gains from adjuvant chemotherapy after NSCLC surgery. These results help extend existing clinical trial data to specific unstudied, high-risk populations and may reduce the uncertainty regarding adjuvant chemotherapy use in these patients.

The impact of stay-at-home orders on the rate of emergency department child maltreatment diagnoses.

BACKGROUND: There is limited data regarding the rates and severity of child maltreatment in medical settings during the COVID-19 pandemic, and the reports are somewhat contradictory. OBJECTIVE: To examine the rates of emergency department (ED) child maltreatment (CM) diagnosis before and after the California statewide stay-at-home order, as well as potential disparities by age, gender, race/ethnicity, and Medicaid status. METHODS: A retrospective pre-post interrupted time series was conducted using data from the electronic health records of children (<18 years) with at least one emergency department visit between January 1, 2019 and September 30, 2021. Enactment of the stay-at-home order in California, March 2020 was used to determine a change in trend of rates of diagnosis of CM in the ED. RESULTS: Overall the study included 407,228 pediatric ED visits. There was a significant change in the percentage of CM visits immediately after the stay-at-home order, followed by small month to month decreases returning to near pre-stay-at-home order levels. This significant increase was driven by higher risk for children <4 years old. The increased rate of CM in the first month after the stay-at-home order was also elevated for female, Black, and Hispanic children. CONCLUSIONS: Our results indicated the rates of CM diagnoses in the ED doubled after the March 2020 stay-at-home order in California. Additionally, our findings suggest that some children may be at higher risk than others, which supports the importance of social safety nets for children in times of national emergency.

Ambient Air Pollutant Exposures and COVID-19 Severity and Mortality in a Cohort of Patients with COVID-19 in Southern California.

Rationale: Ecological studies have shown air pollution associations with coronavirus disease (COVID-19) outcomes. However, few cohort studies have been conducted. Objectives: To conduct a cohort study investigating the association between air pollution and COVID-19 severity using individual-level data from the electronic medical record. Methods: This cohort included all individuals who received diagnoses of COVID-19 from Kaiser Permanente Southern California between March 1 and August 31, 2020. One-year and 1-month averaged ambient air pollutant (particulate matter ⩽2.5 µm in aerodynamic diameter [PM2.5], NO2, and O3) exposures before COVID-19 diagnosis were estimated on the basis of residential address history. Outcomes included COVID-19-related hospitalizations, intensive respiratory support (IRS), and ICU admissions within 30 days and mortality within 60 days after COVID-19 diagnosis. Covariates included socioeconomic characteristics and comorbidities. Measurements and Main Results: Among 74,915 individuals (mean age, 42.5 years; 54% women; 66% Hispanic), rates of hospitalization, IRS, ICU admission, and mortality were 6.3%, 2.4%, 1.5%, and 1.5%, respectively. Using multipollutant models adjusted for covariates, 1-year PM2.5 and 1-month NO2 average exposures were associated with COVID-19 severity. The odds ratios associated with a 1-SD increase in 1-year PM2.5 (SD, 1.5 µg/m3) were 1.24 (95% confidence interval [CI], 1.16-1.32) for COVID-19-related hospitalization, 1.33 (95% CI, 1.20-1.47) for IRS, and 1.32 (95% CI, 1.16-1.51) for ICU admission; the corresponding odds ratios associated with 1-month NO2 (SD, 3.3 ppb) were 1.12 (95% CI, 1.06-1.17) for hospitalization, 1.18 (95% CI, 1.10-1.27) for IRS, and 1.21 (95% CI, 1.11-1.33) for ICU admission. The hazard ratios for mortality were 1.14 (95% CI, 1.02-1.27) for 1-year PM2.5 and 1.07 (95% CI, 0.98-1.16) for 1-month NO2. No significant interactions with age, sex or ethnicity were observed. Conclusions: Ambient PM2.5 and NO2 exposures may affect COVID-19 severity and mortality.

Ambient air pollution and COVID-19 incidence during four 2020-2021 case surges.

BACKGROUND: Air pollution exposure may make people more vulnerable to COVID-19 infection. However, previous studies in this area mostly focused on infection before May 2020 and long-term exposure. OBJECTIVE: To assess both long-term and short-term exposure to air pollution and COVID-19 incidence across four case surges from 03/1/2020 to 02/28/2021. METHODS: The cohort included 4.6 million members from a large integrated health care system in southern California with comprehensive electronic medical records (EMR). COVID-19 cases were identified from EMR. Incidence of COVID-19 was computed at the census tract-level among members. Prior 1-month and 1-year averaged air pollutant levels (PM2.5, NO2, and O3) at the census tract-level were estimated based on hourly and daily air quality data. Data analyses were conducted by each wave: 3/1/2020-5/31/2020, 6/1/202-9/30/2020, 10/1/2020-12/31/2020, and 1/1/2021-2/28/2021 and pooled across waves using meta-analysis. Generalized linear mixed effects models with Poisson distribution and spatial autocorrelation were used with adjustment for meteorological factors and census tract-level social and health characteristics. Results were expressed as relative risk (RR) per 1 standard deviation. RESULTS: The cohort included 446,440 COVID-19 cases covering 4609 census tracts. The pooled RRs (95% CI) of COVID-19 incidence associated with 1-year exposures to PM2.5, NO2, and O3 were 1.11 (1.04, 1.18) per 2.3 µg/m3,1.09 (1.02, 1.17) per 3.2 ppb, and 1.06 (1.00, 1.12) per 5.5 ppb respectively. The corresponding RRs (95% CI) associated with prior 1-month exposures were 1.11 (1.03, 1.20) per 5.2 µg/m3 for PM2.5, 1.09 (1.01, 1.17) per 6.0 ppb for NO2 and 0.96 (0.85, 1.08) per 12.0 ppb for O3. CONCLUSION: Long-term PM2.5 and NO2 exposures were associated with increased risk of COVID-19 incidence across all case surges before February 2021. Short-term PM2.5 and NO2 exposures were also associated. Our findings suggest that air pollution may play a role in increasing the risk of COVID-19 infection.

Healthcare Utilization Among Patients Diagnosed with COVID-19 in a Large Integrated Health System.

BACKGROUND: The demands for healthcare resources following a COVID-19 diagnosis are substantial, but not currently quantified. OBJECTIVE: To describe trends in healthcare utilization within 180 days for patients diagnosed with COVID-19 and identify patient factors associated with increased healthcare use. DESIGN: Observational cohort study. PATIENTS: A total of 64,011 patients with a test-confirmed COVID-19 diagnosis from March to September 2020 in a large integrated healthcare system in Southern California. MAIN MEASURES: Overall healthcare utilization during the 180 days following COVID-19 diagnosis, as well as encounter types and reasons for visits during the first 30 days. Poisson regression was used to identify patient factors associated with higher utilization. Analyses were performed separately for patients who were and were not hospitalized for COVID-19. KEY RESULTS: Healthcare utilization was about twice as high for hospitalized patients compared to non-hospitalized patients in all time periods. The average number of visits was highest in the first 30 days (hospitalized: 12.3 visits/30 person-days; non-hospitalized: 6.6) and gradually decreased over time. In the first 30 days, the majority of healthcare visits were telehealth encounters (hospitalized: 9.0 visits; non-hospitalized: 5.6 visits), and the most prevalent reasons for visits were COVID-related diagnoses, COVID-related symptoms, and respiratory-related conditions. For hospitalized patients, older age (≥65: RR 1.27, 95% CI 1.15-1.41), female gender (RR 1.07, 95% CI 1.05-1.09), and higher BMI (≥40: RR 1.07, 95% CI 1.03-1.10) were associated with higher total utilization. For non-hospitalized patients, older age, female gender, higher BMI, non-white race/ethnicity, former smoking, and greater number of pre-existing comorbidities were all associated with increased utilization. CONCLUSIONS: Patients with COVID-19 seek healthcare frequently within 30 days of diagnosis, placing high demands on health systems. Identifying ways to support patients diagnosed with COVID-19 while adequately providing the usual recommended care to our communities will be important as we recover from the pandemic.

The Probability of Lung Cancer in Patients With Incidentally Detected Pulmonary Nodules: Clinical Characteristics and Accuracy of Prediction Models.

BACKGROUND: The frequency of cancer and accuracy of prediction models have not been studied in large, population-based samples of patients with incidental pulmonary nodules measuring > 8 mm in diameter. RESEARCH QUESTIONS: How does the frequency of cancer vary by size and smoking history among patients with incidental nodules? How accurate are two widely used models for identifying cancer in these patients? STUDY DESIGN AND METHODS: We assembled a retrospective cohort of individuals with incidental nodules measuring > 8 mm in diameter identified by chest CT imaging between 2006 and 2016. We used a validated natural language processing algorithm to identify nodules and their characteristics by scanning the text of dictated radiology reports. We reported patient and nodule characteristics stratified by the presence or absence of a lung cancer diagnosis within 27 months of nodule identification and estimated the area under the receiver operating characteristic curve (AUC) to compare the accuracy of the Mayo Clinic and Brock models for identifying cancer. RESULTS: The sample included 23,780 individuals with a nodule measuring > 8 mm, including 2,356 patients (9.9%) with a lung cancer diagnosis within 27 months of nodule identification. Cancer was diagnosed in 5.4% of never smokers, 12.2% of former smokers, and 17.7% of current smokers. Cancer was diagnosed in 5.7% of patients with nodules measuring 9 to 15 mm, 12.1% of patients with nodules > 15 to 20 mm, and 18.4% of patients with nodules > 20 to 30 mm. In the full sample, the Mayo Clinic model (AUC, 0.747; 95% CI, 0.737-0.757) was more accurate than the Brock model (AUC, 0.713; 95% CI, 0.702-0.724; P < .0001). When restricted to ever smokers, the Mayo Clinic model was still more accurate. Both models overestimated the probability of cancer. INTERPRETATION: Almost 10% of patients with an incidental pulmonary nodule measuring > 8 mm in diameter will receive a lung cancer diagnosis. Existing prediction models have only fair accuracy and overestimate the probability of cancer.

Near-roadway air pollution associated with COVID-19 severity and mortality - Multiethnic cohort study in Southern California.

BACKGROUND: Air pollution exposure has been associated with increased risk of COVID-19 incidence and mortality by ecological analyses. Few studies have investigated the specific effect of traffic-related air pollution on COVID-19 severity. OBJECTIVE: To investigate the associations of near-roadway air pollution (NRAP) exposure with COVID-19 severity and mortality using individual-level exposure and outcome data. METHODS: The retrospective cohort includes 75,010 individuals (mean age 42.5 years, 54% female, 66% Hispanic) diagnosed with COVID-19 at Kaiser Permanente Southern California between 3/1/2020-8/31/2020. NRAP exposures from both freeways and non-freeways during 1-year prior to the COVID-19 diagnosis date were estimated based on residential address history using the CALINE4 line source dispersion model. Primary outcomes include COVID-19 severity defined as COVID-19-related hospitalizations, intensive respiratory support (IRS), intensive care unit (ICU) admissions within 30 days, and mortality within 60 days after COVID-19 diagnosis. Covariates including socio-characteristics and comorbidities were adjusted for in the analysis. RESULT: One standard deviation (SD) increase in 1-year-averaged non-freeway NRAP (0.5 ppb NOx) was associated with increased odds of COVID-19-related IRS and ICU admission [OR (95% CI): 1.07 (1.01, 1.13) and 1.11 (1.04, 1.19) respectively] and increased risk of mortality (HR = 1.10, 95% CI = 1.03, 1.18). The associations of non-freeway NRAP with COVID-19 outcomes were largely independent of the effect of regional fine particulate matter and nitrogen dioxide exposures. These associations were generally consistent across age, sex, and race/ethnicity subgroups. The associations of freeway and total NRAP with COVID-19 severity and mortality were not statistically significant. CONCLUSIONS: Data from this multiethnic cohort suggested that NRAP, particularly non-freeway exposure in Southern California, may be associated with increased risk of COVID-19 severity and mortality among COVID-19 infected patients. Future studies are needed to assess the impact of emerging COVID-19 variants and chemical components from freeway and non-freeway NRAP.

Asthma Disease Status, COPD, and COVID-19 Severity in a Large Multiethnic Population.

BACKGROUND: Current studies of asthma history on coronavirus disease 2019 (COVID-19) outcomes are limited and lack consideration of disease status. OBJECTIVE: To conduct a population-based study to assess asthma disease status and chronic obstructive pulmonary disease (COPD) in relation to COVID-19 severity. METHODS: Patients diagnosed with COVID-19 (n = 61,338) in a large, diverse integrated health care system were identified. Asthma/COPD history, medication use, and covariates were extracted from electronic medical records. Asthma patients were categorized into those with and without clinical visits for asthma 12 or fewer months prior to COVID-19 diagnosis and labeled as active and inactive asthma, respectively. Primary outcomes included COVID-19-related hospitalizations, intensive respiratory support (IRS), and intensive care unit admissions within 30 days, and mortality within 60 days after COVID-19 diagnosis. Logistic and Cox regression were used to relate COVID-19 outcomes to asthma/COPD history. RESULTS: The cohort was 53.9% female and 66% Hispanic and had a mean age of 43.9 years. Patients with active asthma had increased odds of hospitalization, IRS, and intensive care unit admission (odds ratio 1.47-1.66; P < .05) compared with patients without asthma or COPD. No increased risks were observed for patients with inactive asthma. Chronic obstructive pulmonary disease was associated with increased risks of hospitalization, IRS, and mortality (odds ratio and hazard ratio 1.27-1.67; P < .05). Among active asthma patients, those using asthma medications had greater than 25% lower odds for COVID-19 outcomes than those without medication. CONCLUSIONS: Patients with asthma who required clinical care 12 or fewer months prior to COVID-19 or individuals with COPD history are at increased risk for severe COVID-19 outcomes. Proper medication treatment for asthma may lower this risk.

Natural Language Processing to Identify Pulmonary Nodules and Extract Nodule Characteristics From Radiology Reports.

BACKGROUND: There is an urgent need for population-based studies on managing patients with pulmonary nodules. RESEARCH QUESTION: Is it possible to identify pulmonary nodules and associated characteristics using an automated method? STUDY DESIGN AND METHODS: We revised and refined an existing natural language processing (NLP) algorithm to identify radiology transcripts with pulmonary nodules and greatly expanded its functionality to identify the characteristics of the largest nodule, when present, including size, lobe, laterality, attenuation, calcification, and edge. We compared NLP results with a reference standard of manual transcript review in a random test sample of 200 radiology transcripts. We applied the final automated method to a larger cohort of patients who underwent chest CT scan in an integrated health care system from 2006 to 2016, and described their demographic and clinical characteristics. RESULTS: In the test sample, the NLP algorithm had very high sensitivity (98.6%; 95% CI, 95.0%-99.8%) and specificity (100%; 95% CI, 93.9%-100%) for identifying pulmonary nodules. For attenuation, edge, and calcification, the NLP algorithm achieved similar accuracies, and it correctly identified the diameter of the largest nodule in 135 of 141 cases (95.7%; 95% CI, 91.0%-98.4%). In the larger cohort, the NLP found 217,771 reports with nodules among 717,304 chest CT reports (30.4%). From 2006 to 2016, the number of reports with nodules increased by 150%, and the mean size of the largest nodule gradually decreased from 11 to 8.9 mm. Radiologists documented the laterality and lobe (90%-95%) more often than the attenuation, calcification, and edge characteristics (11%-14%). INTERPRETATION: The NLP algorithm identified pulmonary nodules and associated characteristics with high accuracy. In our community practice settings, the documentation of nodule characteristics is incomplete. Our results call for better documentation of nodule findings. The NLP algorithm can be used in population-based studies to identify pulmonary nodules, avoiding labor-intensive chart review.

Red meat consumption, cooking mutagens, NAT1/2 genotypes and pancreatic cancer risk in two ethnically diverse prospective cohorts.

There is limited evidence on the association between red meat consumption and pancreatic cancer among ethnic minorities. We assessed this relationship in two large prospective cohorts: the Multiethnic Cohort Study (MEC) and the Southern Community Cohort Study (SCCS). Demographic, dietary and other risk factor data were collected at cohort entry. Red meat intake was assessed using cohort-specific validated food frequency questionnaires. Incident pancreatic cancer cases were identified via linkages to state cancer registries. Cox regression was used to calculate relative risks (RRs) and 95% confidence intervals (CIs) for the association of red meat intake with pancreatic cancer risk in each cohort. We performed additional analyses to evaluate cooking methods, mutagens and effect modification by NAT1/2 genotypes. From a total of 184 542 (MEC) and 66 793 (SCCS) at-risk participants, we identified 1618 (MEC) and 266 (SCCS) incident pancreatic cancer cases. Red meat consumption was associated with pancreatic cancer risk in the MEC (RRQ4vsQ1 1.18, 95% CI 1.02-1.37) and with borderline statistical significance in the SCCS (RRQ4vsQ1 1.31, 95% CI 0.93-1.86). This association was significant in African Americans (RRQ4vsQ1 1.49, 95% CI 1.06-2.11) and Latinos (RRQ4vsQ1 1.44, 95% CI 1.02-2.04) in the MEC, and among African Americans (RRQ4vsQ1 1.55, 95% CI 1.03-2.33) in the SCCS. NAT2 genotypes appeared to modify the relationship between red meat and pancreatic cancer in the MEC (pinteraction = 0.03). Our findings suggest that the associations for red meat may be strongest in African Americans and Latinos. The mechanisms underlying the increased risk for these populations should be further investigated.

Machine Learning for Early Lung Cancer Identification Using Routine Clinical and Laboratory Data.

Rationale: Most lung cancers are diagnosed at an advanced stage. Presymptomatic identification of high-risk individuals can prompt earlier intervention and improve long-term outcomes. Objectives: To develop a model to predict a future diagnosis of lung cancer on the basis of routine clinical and laboratory data by using machine learning. Methods: We assembled data from 6,505 case patients with non-small cell lung cancer (NSCLC) and 189,597 contemporaneous control subjects and compared the accuracy of a novel machine learning model with a modified version of the well-validated 2012 Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial risk model (mPLCOm2012), by using the area under the receiver operating characteristic curve (AUC), sensitivity, and diagnostic odds ratio (OR) as measures of model performance. Measurements and Main Results: Among ever-smokers in the test set, a machine learning model was more accurate than the mPLCOm2012 for identifying NSCLC 9-12 months before clinical diagnosis (P < 0.00001) and demonstrated an AUC of 0.86, a diagnostic OR of 12.3, and a sensitivity of 40.1% at a predefined specificity of 95%. In comparison, the mPLCOm2012 demonstrated an AUC of 0.79, an OR of 7.4, and a sensitivity of 27.9% at the same specificity. The machine learning model was more accurate than standard eligibility criteria for lung cancer screening and more accurate than the mPLCOm2012 when applied to a screening-eligible population. Influential model variables included known risk factors and novel predictors such as white blood cell and platelet counts. Conclusions: A machine learning model was more accurate for early diagnosis of NSCLC than either standard eligibility criteria for screening or the mPLCOm2012, demonstrating the potential to help prevent lung cancer deaths through early detection.