Disparities in esophageal cancer care based on race: a National Cancer Database analysis.

Esophageal cancer is one of the most common cancer killers in our country. The effects of racial disparities on care for esophageal cancer patients are incompletely understood. Using the National Cancer Database, we investigated racial disparities in treatment and outcome of esophageal cancer patients. The National Cancer Database was queried from 2004 to 2017. Logistic regression and survival analysis were used to determine racial differences in access, treatment and outcome. A total of 127,098 patients were included. All minority groups were more likely to be diagnosed at advanced stages versus Caucasians after adjusting for covariates (African American OR-1.64 [95% confidence interval 1.53-1.76], Hispanic OR-1.19 [1.08-1.32], Asian OR-1.78 [1.55-2.06]). After adjustment, all minorities were less likely at every stage to receive surgery. Despite these disparities, Hispanics and Asians had improved survival compared with Caucasians. African Americans had worse survival. Racial disparities for receiving surgery were present in both academic and community institutions, and at high-volume and low-volume institutions. Surgery partially mediated the survival difference between African Americans and Caucasians (HR-1.13 [1.10-1.16] and HR-1.04 [1.02-1.07], without and with adjustment of surgery).There are racial disparities in the treatment of esophageal cancer. Despite these disparities, Hispanics and Asians have improved overall survival versus Caucasians. African Americans have the worst overall survival. Racial disparities likely affect outcome in esophageal cancer. But other factors, such as epigenetics and tumor biology, may correlate more strongly with outcome for patients with esophageal cancer.

COVID-19 Pandemic and Surgical Oncology: Preserving the Academic Mission.

BACKGROUND: The global pandemic of respiratory disease cause by the novel human coronavirus (SARS-CoV-2) has caused untold suffering, loss of life and upheaval in society. The pandemic has lead to massive redirection of health care resources to treat the surge of COVID-19 patients, and enforcement of social distancing to reduce the rate of transmission. METHODS: Editorial Board members provided observations of the implications of the pandemic on academic surgical oncology. RESULTS: Delivery of health care to other populations including cancer patients has been significantly disrupted. The implications both short term and long term threaten preservation of the academic mission in medicine at large, and certainly in the field of surgical oncology. CONCLUSIONS: The effects on surgical oncology training, research and clinical trials are major.

Effects of proton pump inhibitor use on the esophageal microbial community.

BACKGROUND: Changes in the esophageal microbiome correlate with esophageal disease, but the effects of proton pump inhibitor (PPI) drugs are incompletely characterized. Our objective was to identify the effects of PPI use on the microbial community of the esophagus. METHODS: Mucosal biopsies of the distal esophagus were analyzed using a customized esophageal microbiome qPCR panel array (EMB). Patient demographics, use of PPIs, duration of use and dose were recorded. RESULTS: Fifty-eight patients were included. Mean age was 60.5 years. Ninety percent (52/58) of patients were on PPIs. Mean dose was 42.7 mg. Mean duration of use was 2.5 years. The use of PPIs led to a significant difference in absolute levels of only one organism, Actinomyces, in the entire array (p < 0.01). Among patients who used proton pump inhibitors, there was no significant association between dose and absolute levels of any organism. Similarly, there was no association between duration of use and absolute levels of any organism. CONCLUSIONS: PPI use does not seem to cause significant changes in the distal esophageal microbial community. Future studies with larger sample sizes and esophageal pH testing should be performed to determine the level of acidity and its relationship to the microbial community.

Regional trends in diagnosis of advanced lung cancer in Michigan over 33 years.

Background: Lung cancer is the most common cancer killer worldwide. Nearly 80 percent of lung cancers are diagnosed at advanced stages. Lack of access to medical care and undwerutilized lung cancer screening are key reasons for advanced diagnoses. We sought to understand the regional differences in presentation of lung cancer across Michigan. Utilizing a comprehensive cancer registry over 33 years, our goal was to examine associations between sociodemographic patient factors and diagnoses at advanced stages. Methods: The Michigan Cancer Registry was queried from 1985 to 2018 to include all new diagnoses of non-small cell lung cancer (NSCLC) using International Classification of Diseases for Oncology (ICD-O) version 3 codes. NSCLC was categorized as early, regional and distant disease. Advanced disease was considered to be any disease that was regional or distant. NSCLC rates were calculated and mapped at the zip code level using the 2010 population as the denominator and spatial empirical Bayes methodology. Regional hospital service areas were constructed using travel time to treatment from the patient's zip code centroid. Logistic regression models were estimated to investigate the significance of rural vs. urban and travel time on level of disease at presentation. Kaplan-Meier and multivariate survival analysis was performed to evaluate the association between distance from the nearest medical center and length of survival controlling for known risk factors for lung cancer. Results: From 1985 to 2018, there were 141,977 patients in Michigan diagnosed with NSCLC. In 1985, men were 2.2 times more likely than women to be diagnosed but by 2018 women and men developed disease at equal rates. Mean age was 67.8 years. Among all patients with known stage of disease, 72.5% of patients were diagnosed with advanced disease. Regional and distant NSCLC rates were both higher in the northern parts of the state. Longer drive times in rural regions also significantly increased the likelihood of advanced NSCLC diagnoses, in particular regional lung cancer. Patients with longer drive times also experienced overall worse survival after controlling for other factors. Conclusions: Regional disparities exist in Michigan for diagnoses of NSCLC at advanced stages. Factors such as lack of screening in urban regions and distances to treating institutions in rural areas likely contribute to the increased likelihood of advanced NSCLC. Future interventions should target the specific needs of residents to detect disease at earlier stages and improve overall outcomes.

Support service use and interest in support services among distressed family caregivers of lung cancer patients.

OBJECTIVE: This study examined support service use and interest in support services among distressed family caregivers of patients recently entering comprehensive cancer care facilities. METHODS: Primary family caregivers of lung cancer patients (N=83) were recruited from three medical centers within 12 weeks of the patient's new visit to the oncology clinic. All family caregivers were screened for psychological distress, and those reporting significant anxiety or depressive symptoms were eligible for this study. Caregivers completed a baseline assessment of support service use (i.e., use of mental health services and complementary and alternative medicine [CAM]) and interest in support services. Support service use was also assessed 3 months later. RESULTS: Although all caregivers reported clinically meaningful distress, only 26% used mental health and 39% used CAM services during the 3-month study period. Patients' receipt of chemotherapy was positively associated with caregivers' mental health service use, whereas greater education and receiving assistance with caregiving tasks were associated with CAM use. Forty percent of caregivers who did not use CAM at baseline were interested in CAM. In addition, 29% of caregivers who did not receive mental health services at baseline were interested in professional psychosocial support, and 29% of caregivers who did not receive staff assistance with practical needs at baseline were interested in this service. CONCLUSIONS: Findings suggest that distressed family caregivers of lung cancer patients underuse mental health services and that a sizable minority are interested in professional help with psychosocial and practical needs.

Economic and social changes among distressed family caregivers of lung cancer patients.

PURPOSE: Although costs of lung cancer care have been documented, economic and social changes among lung cancer patients' family caregivers have yet to be fully examined. In addition, research has not focused on caregivers with greater need for support services. This study examined various economic and social changes among distressed family caregivers of lung cancer patients during the initial months of cancer care in the USA. METHODS: Lung cancer patients' primary family caregivers with significant anxiety or depressive symptoms were recruited from three medical centers within 12 weeks of the patient's new oncology visit. Caregivers (N = 83) reported demographic and medical information and caregiving burden at baseline. Seventy-four caregivers reported anxiety and depressive symptoms and economic and social changes 3 months later. RESULTS: Seventy-four percent of distressed caregivers experienced one or more adverse economic or social changes since the patient's illness. Common changes included caregivers' disengagement from most social and leisure activities (56%) and, among employed caregivers (n = 49), reduced hours of work (45%). In 18% of cases, a family member quit work or made another major lifestyle change due to caregiving. Additionally, 28% of caregivers reported losing the main source of family income, and 18% reported losing most or all of the family savings. Loss of the main source of family income and disengagement from most social and leisure activities predicted greater caregiver distress. CONCLUSIONS: Findings suggest that distressed caregivers of lung cancer patients experience high rates of adverse economic and social changes that warrant clinical and research attention.

The growing impact of air quality on lung-related illness: a narrative review.

Background and Objective: Poor air quality can be harmful to human well-being. There are a variety of respiratory disorders associated with toxins present within the atmosphere, such as bronchitis and asthma, which eventually lead to heart or lung complications over time. Fine particles like particulate matter 2.5 (PM2.5) accumulate in the small airways of the lung. These irritants can cause epigenetic modifications in gene regulation, leading to changes responsible for both benign and malignant lung diseases. In this review we will discuss known associations between environmental factors and pulmonary complications, consider preventative measures and offer further areas for future investigation. This review presents a summary of the literature outlining the current work done on air quality and its effects on lung-related illnesses. We discuss regional differences in air quality and consider the causes, such as manufacturing, traffic density, increase in fuel usage and natural events. We further explore disparities based on geography, race, and other social determinants. Methods: A comprehensive literature review was performed using keywords related to air quality, pollution and lung disease within the PubMed database as well as MEDLINE and Google Scholar. Key Content and Findings: The Clean Air Act of 1970 marked an essential transition for air quality improvement. The legislation led to decreased emissions and control measures to address atmosphere contamination. Despite these actions, poor atmospheric conditions still persist today and have become an ongoing issue. These poor conditions affect individuals living in metropolitan areas more significantly than suburban or rural areas. Pollution from industrial operations and transportation vehicles have led to increased emission outputs recently. Climate change further aggravates air quality problems by raising pollutant and allergen concentrations. The detrimental consequences of poor air quality include increased incidence of disease processes like asthma, chronic obstructive pulmonary disease (COPD) and lung cancer. To keep up with the well-being of people globally, it is important that actions be taken to battle contamination in the climate so its impact on public health can be limited. Conclusions: Poor air quality and recent worsening of industrial emissions have had a negative impact on lung-related illnesses. Future mitigation strategies should be taken to reduce pollution and treat diseases earlier in their course. Some of these strategies include more reliance on alternative energy sources, creation of mass transit systems and increased rates of recycling.

The Role of Sublobar Resection for the Surgical Treatment of Non-Small Cell Lung Cancer.

Lung cancer is the most common cancer killer in the world. The standard of care for surgical treatment of non-small cell lung cancer has been lobectomy. Recent studies have identified that sublobar resection has non-inferior survival rates compared to lobectomy, however. Sublobar resection may increase the number of patients who can tolerate surgery and reduce postoperative pulmonary decline. Sublobar resection appears to have equivalent results to surgery in patients with small, peripheral tumors and no lymph node disease. As the utilization of segmentectomy increases, there may be some centers that perform this operation more than other centers. Care must be taken to ensure that all patients have access to this modality. Future investigations should focus on examining the outcomes from segmentectomy as it is applied more widely. When employed on a broad scale, morbidity and survival rates should be monitored. As segmentectomy is performed more frequently, patients may experience improved postoperative quality of life while maintaining the same oncologic benefit.

Gentrification and Air Quality in a Large Urban County in the United States.

Introduction: Increases in industrialization and manufacturing have led to worsening pollution in some components of air quality. In addition, gentrification is occurring in large cities throughout the world. As these socioeconomic and demographic changes occur, there have been no studies examining the association of gentrification with air quality. To investigate this association, we studied the trends of gentrification, changes in racial distribution and changes in air quality in each zip code of a large urban county over a 40-year period. Methods: We conducted a retrospective longitudinal study over 40 years in Wayne County, Michigan using socioeconomic and demographic data from the National Historical Geographic Information System (NHGIS) and air quality data from the United States Environmental Protection Agency (EPA). To assess gentrification, longitudinal analyses were performed to examine median household income, percentage with a college education, median housing value, median gross rent and employment level. The racial distribution was evaluated in each zip code during the time period. Gentrification was studied in relation to air quality using nonparametric 2-sample Wilcon-Mann-Whitney tests and Binomial Generalized Linear Regression models. Results: Although air quality improved overall over the 40-year period, there was a lesser rate of improvement in gentrified areas. Furthermore, gentrification was strongly associated with racial distribution. The most substantial gentrification occurred from 2010 to 2020, in which a specific cluster of adjacent zip codes in downtown Detroit experienced intense gentrification and a drop in the percentage of African-American residents. Conclusions: Gentrified areas seem to have a less pronounced improvement in air quality over time. This reduction in air quality improvement is likely associated with demolitions and the construction of new buildings, such as sporting arenas and accompanying traffic density. Gentrification is also strongly associated with an increase in non-minority residents in an area. Although previous definitions of gentrification in the literature have not included racial distribution, we suggest that future definitions should include this metric given the strong association. Minority residents who are displaced as a result of gentrification do not experience the improvements in housing quality, accessibility to healthy foods and other associations of gentrification.

Treatment patterns and outcomes in patients with Pancoast tumors: a national cancer database analysis.

Background: Pancoast tumors represent 5% of non-small cell lung cancers. Complete surgical resection and no lymph node involvement are important positive prognostic factors. Previous literature has identified neoadjuvant chemoradiation treatment, followed by surgical resection, as the standard of care. But many institutions choose upfront surgery. Our goal was to identify the treatment patterns and outcomes in patients with node-negative Pancoast tumors using the National Cancer Database (NCDB). Methods: The NCDB was queried from 2004 through 2017 to identify all patients who had undergone surgery for a Pancoast tumor. Treatment patterns, including the percentage of patients who received neoadjuvant treatment, were recorded. Logistic regression and survival analyses were used to determine outcomes based on different treatment patterns. Secondary analyses were performed on the cohort who received upfront surgery. Results: A total of 2,910 patients were included in the study. Overall 30- and 90-day mortality were 3% and 7% respectively. Only 25% (717/2,910) of the group received neoadjuvant chemoradiation treatment prior to surgery. Patients who received neoadjuvant chemoradiation treatment experienced significantly improved 90-day survival (P<0.01) and overall survival (P<0.01). When analyzing the cohort who received upfront surgery, there was a statistically significant difference in survival based on adjuvant treatment pattern (P<0.01). Patients in this group who received adjuvant chemoradiation had the best survival, whereas patients who received adjuvant radiation only or no treatment had the worst outcomes. Conclusions: Patients with Pancoast tumors receive neoadjuvant chemoradiation treatment in only a quarter of cases nationally. Patients who received neoadjuvant chemoradiation treatment had improved survival compared to patients who had upfront surgery. Similarly, when surgery is performed first, adjuvant chemoradiation treatment improved survival compared to other adjuvant strategies. These results suggest underutilization of neoadjuvant treatment for patients with node-negative Pancoast tumors. Future studies with a more clearly defined cohort are needed to assess the treatment patterns being utilized on patients with node-negative Pancoast tumors. It will be beneficial to see whether neoadjuvant treatment for Pancoast tumors has increased in recent years.

The relationship of social determinants and distress in newly diagnosed cancer patients.

Patients with a new cancer diagnosis can experience distress when diagnosed. There are disparities in treatment of cancer patients based on social determinants, but minimal research exists on the relationship of those social determinants and distress after a new cancer diagnosis. Our goals were to determine the social determinants associated with distress after a new cancer diagnosis and determine the relationship of distress with outcome. Patients with a new cancer diagnosis at one institution from January 2019 to December 2020 were analyzed. Patients were given the National Comprehensive Cancer Network (NCCN) distress thermometer during their first visit. Demographics, tumor characteristics, clinical variables and survival were recorded. Patients were also asked to share specific factors that led to distress, including: (1) financial, (2) transportation, (3) childcare and (4) religious. A total of 916 patients returned distress thermometers. Mean age was 59.1 years. Females comprised 71.3 (653/916) percent of the cohort. On Dunn's multiple comparison, the following factors were associated with increased distress level: female (p < 0.01), ages 27 to 45 (p < 0.01), uninsured (p < 0.01) and unemployed (p < 0.01). Patients with higher distress scores also experienced worse overall survival (p < 0.05). Females, young patients, uninsured patients and unemployed patients experience more distress after a new cancer diagnosis. Increased distress is independently associated with worse overall survival. Social determinants can be used to predict which patients may require focused interventions to reduce distress after a new cancer diagnosis.

Risk factors for bronchopleural fistula after right pneumonectomy: does eliminating the stump diverticulum provide protection?

PURPOSE: Bronchopleural fistula (BPF) remains an important source of morbidity and mortality after right pneumonectomy. We reviewed our 18-year institutional experience with right pneumonectomy to identify risks factors for BPF. METHODS: From 1992 to 2010, a total of 145 patients who underwent right pneumonectomy were identified from an institutional database. Median age was 56 years. Most patients (66.2%) underwent surgery for non-small cell lung cancer. Sixty-seven patients (46.2%) received either chemotherapy or radiotherapy before surgery. Medical records were reviewed for 14 variables potentially predictive for BPF, including two airway closure techniques (standard bronchial closure and carinal closure). Variables predictive of BPF by univariate analysis were entered into a logistic regression model. RESULTS: The overall mortality rate was 13.1% (n=19), with 15.9 and 10.5% mortality in the bronchial closure and carinal closure groups, respectively (P=0.33). The overall BPF rate was 7.6% (n=11), with a 3.9% (3 of 76) rate in the carinal closure group compared to 11.6% (8 of 69) in the bronchial closure group (P=0.08). Seven of eight bronchial closure patients who developed BPF required operative repair. Only one of three patients who developed BPF after carinal closure did not spontaneously heal after open drainage. Multivariate analysis identified preoperative radiation dose (P=0.042) and bronchial closure (P=0.041) as independent risk factors for BPF, while the length of postoperative ventilation before development of BPF approached significance (P=0.057). CONCLUSIONS: In our experience, higher preoperative radiation doses are a risk factor for BPF after right pneumonectomy, while carinal closure exerts a protective effect.

Applicant perception of virtual interviews in cardiothoracic surgery: A Thoracic Education Cooperative Group Study.

OBJECTIVES: Cardiothoracic programs used virtual interviews exclusively this year. As programs consider using virtual interviews permanently, our goal was to evaluate the experience of applicants with virtual interviews. METHODS: All 2020-2021 traditional cardiothoracic fellowship applicants received an anonymous electronic survey after the Match process ended. The survey assessed the number of interviews, strengths, and inadequacies of virtual interviews and factors that affected rank decision. RESULTS: Forty-three percent of applicants responded (60/139). The average number of interviews was 16.0. Eighty percent (48/60) of respondents successfully matched. Eighty-seven percent (52/60) of respondents had a favorable experience with virtual interviews, and 97% (58/60) found them to be convenient. However, only 50% (30/60) were able to evaluate a program fully. Respondents who matched were more likely to have a favorable experience (P = .02), but not more likely to be able to evaluate a program fully (P = .35). The most valued aspect was the informal meet and greet session with fellows (4.2 of 5). The least valued aspect was the program's social media site (2.0 of 5). The factors most frequently used to decide ranking were case numbers by 92% (55/60) and culture/personality by 82% (49/60). CONCLUSIONS: Virtual interviews were perceived more favorably compared with last year, but half of applicants were still unable to evaluate a program fully. Fellow interactions were the most popular aspect of virtual interviews. As programs consider using virtual interviews permanently, more exposure to current trainees and a more robust social media/online presence will improve favorability.

Microbiota Detection Patterns Correlate With Presence and Severity of Barrett's Esophagus.

Background: The microbiome has been increasingly associated with different disease processes, but its role in esophagus is largely unknown. Our goal was to determine the associations of the esophageal microbiota with Barrett's esophagus. Methods: A total of 74 patients were included in this prospective study, including 34 patients with Barrett's esophagus and 40 patients without Barrett's esophagus. Esophageal swabs were obtained from the uvula, and mucosal biopsies were obtained from the proximal esophagus and distal esophagus in each patient. The microbiome of each sample was assessed using a customized Esophageal Microbiome qPCR array (EMB). For each clinical sample, we completed a detection/non-detection analysis for each organism in the EMB. The limit of detection (LOD) for each target was established by analysis of plasmid dilutions. Results: Average age was 60.2 years. There were significantly different microbial detection patterns in patients with Barrett's esophagus compared to the control population. There were a greater number of organisms which had different likelihoods of detection in the distal esophagus, compared to the proximal esophagus or uvula. In addition, as the length of the Barrett's column increased, multiple organisms were less likely to be detected. This decreased likelihood occurred only in the distal esophagus. Beside Barrett's esophagus, no other demographic factors were associated with differences in detection patterns. Conclusions: Microbial community structures differ between patients with and without Barrett's esophagus. Certain organisms are less likely to be detected as the severity of Barrett's esophagus worsens. These results suggest that particular organisms may have a protective effect against the development of Barrett's esophagus.

Trends in lung cancer screening in the United States, 2016-2017.

BACKGROUND: Lung cancer is the most common cancer killer annually, yet the overall rate of eligible patients who undergo screening with low-dose computed tomography (LDCT) is low. Our goal was to determine factors which were associated with the probability of obtaining lung cancer screening. METHODS: The Clinformatics Data Mart (CDM) database, a national commercial health insurance database with over 18,000,000 enrollees, was queried to determine the rate of LDCT screening and factors which influenced receiving LDCT screening. All enrollees between the ages of 55 and 77 from 2016 to 2017 were included. Demographics, history of tobacco exposure and state smoking statistics were recorded. RESULTS: All 8,350,197 enrollees aged 55-77 were included in the study. Among enrollees, the rate of screening increased throughout 2016 and early 2017 and then appeared to stabilize. In the second half of 2017 the LDCT rate was approximately 6 per 1,000 enrollees per year, and was increasing at a slope of 0.1 additional LDCT per 1,000 enrollees per year. There was marked geographic variation, with 5-fold differences in LDCT rates between different regions. There was no correlation between smoking rate and LDCT rate at the state level (r=0.02; P=0.87). Enrollees aged 65-69 were most likely to be screened (OR =1.53; 1.45-1.61) compared to enrollees aged 55-59. CONCLUSIONS: The rate of LDCT screening is increasing very slowly with time, and most eligible patients are not screened. Further studies are needed to determine the reasons for low screening rates, and the marked geographic variation.

Correlation between air quality and lung cancer incidence: A county by county analysis.

BACKGROUND: Poor air quality has previously been associated with lung cancer development, but the risks associated with regional differences in air quality are poorly understood. We investigated the association of air quality indices with development of lung cancer in all Texas counties. METHODS: Lung cancer incidence, air quality indicators (particulate matter <2.5 µm, radon levels, oil well density), and known risk factors were obtained using data from the Texas Commission on Environmental Quality and the Texas Cancer Registry. Linear regression models were constructed to correlate air quality indicators with lung cancer incidence and advanced stage at diagnosis (stage III or IV), while controlling for other patient characteristics. RESULTS: Lung cancer incidence ranged from 27.6 to 103.4 cases per 100,000 people. In the study, 2.5 µm was associated with increased lung cancer incidence (ß = 4.38, P < .0001), but not radon levels (ß = -2.70, P = .41). Air quality indicators were not significantly associated with an advanced cancer diagnosis. CONCLUSION: There are wide differences in the incidence of lung cancer across Texas. These differences seem to be related to air quality. Identifying high-risk areas may help to guide strategies such as implementation of targeted lung cancer screening programs.

Microbiota of the Oropharynx and Endoscope Compared to the Esophagus.

The role of the microflora in the development of esophageal disease is still largely unknown and is being investigated in more detail. Our goal was to determine how the microbiota levels of endoscope and uvular swabs compared to the levels of tissue biopsies along various points of the esophagus. 17 patients with Barrett's esophagus agreed to participate in the study. Biopsies of esophageal mucosa were taken from the (1) proximal esophagus, (2) mid-esophagus, (3) distal esophagus, and (4) Barrett's esophagus. Swabs were also taken from the uvula and the endoscope. Throughout the esophagus, 17 bacterial genera were detected from the samples. The microflora pattern obtained from the uvula and endoscopic swabs did not correlate well with mucosal biopsies along any aspect of the esophagus. There were statistically significant differences in the levels and proportions of bacteria found when comparing the uvula swab to the esophageal biopsies and when comparing the endoscope swab to the esophageal biopsies. Obtaining a simple swab of the uvula or endoscope itself appears to be a poor substitute for tissue biopsy of esophageal mucosa when evaluating microflora patterns. When performing microflora studies of the esophagus, mucosal biopsies should be used for analysis.

[18F]Fluorodeoxyglucose uptake by positron emission tomography for diagnosis of suspected lung cancer: impact of verification bias.

BACKGROUND: Verification bias occurs when test findings influence the decision to perform a gold standard test. It adversely influences diagnostic test accuracy by inflating sensitivity and deflating specificity. We studied the impact of verification bias on the estimated accuracy of a test commonly used in suspected lung cancer. METHODS: We studied 534 consecutive patients referred for [(18)F]fluorodeoxyglucose uptake by positron emission tomography (PET). Primary outcomes were tissue diagnoses of cancer and of mediastinal lymph node metastases. A secondary outcome was 3-year mortality. We accounted for verification bias using 2 validated methods. RESULTS: The gold standard test, namely tissue acquisition, was performed in 419 patients (78%); mediastinal lymph node sampling occurred in 301 (56%). While the 410 patients with PET-diagnosed stage I cancer or higher were more likely than patients with negative PET scan findings to undergo tissue diagnosis testing (92% vs 34%) (P<.001), there was no association between PET findings and performance of mediastinal sampling. Without accounting for verification bias, the sensitivity and specificity of PET for diagnosis of cancer were 0.95 (95% confidence interval [CI], 0.92-0.97) and 0.31 (95% CI, 0.21-0.42), respectively. After adjustment, sensitivity fell to 0.85 (95% CI, 0.81-0.89), while specificity increased to 0.51 (95% CI, 0.40-0.60). For diagnosis of mediastinal disease, verification bias had slight effects on test accuracy. There were 224 deaths, with a strong gradient between PET stage and death (P<.001). CONCLUSION: The diagnostic accuracy of PET for assessment of suspected lung cancer is substantially affected by verification bias.