Search

Lung cancer continues to be a major public health problem, and more patients die from this disease than any other cancer. The vast majority of patients present with advanced stage disease when therapeutic options are limited and the overall 5 year survival rate remains approximately 15%. Screening with low dose helical computed tomography (CT)has been suggested for early detection, although the effect on mortality is currently under investigation. As part of the National Lung Cancer Screening Trial, a specimen biorepository including blood, sputum, and urine were collected serially for the primary purpose of validating early detection lung cancer biomarkers. In addition tumor samples have been obtained from patients diagnosed with lung cancer to be included in a tissue microarray. This commentary describes the rationale, composition, intent, and availability of specimen in the biorepository.

Rationale and ObjectivesStudies that evaluate the lung-nodule-detection performance of radiologists or computerized methods depend on an initial inventory of the nodules within the thoracic images (the “truth”). The purpose of this study was to analyze (1) variability in the “truth” defined by different combinations of experienced thoracic radiologists and (2) variability in the performance of other experienced thoracic radiologists based on these definitions of “truth” in the context of lung nodule detection on computed tomography (CT) scans.Materials and MethodsTwenty-five thoracic CT scans were reviewed by four thoracic radiologists, who independently marked lesions they considered to be nodules ≥ 3 mm in maximum diameter. Panel “truth” sets of nodules then were derived from the nodules marked by different combinations of two and three of these four radiologists. The nodule-detection performance of the other radiologists was evaluated based on these panel “truth” sets.ResultsThe number of “true” nodules in the different panel “truth” sets ranged from 15–89 (mean: 49.8±25.6). The mean radiologist nodule-detection sensitivities across radiologists and panel “truth” sets for different panel “truth” conditions ranged from 51.0–83.2%; mean false-positive rates ranged from 0.33–1.39 per case.ConclusionSubstantial variability exists across radiologists in the task of lung nodule identification in CT scans. The definition of “truth” on which lung nodule detection studies are based must be carefully considered, since even experienced thoracic radiologists may not perform well when measured against the “truth” established by other experienced thoracic radiologists.Keywords: lung nodule, computed tomography (CT), thoracic imaging, inter-observer variability, computer-aided diagnosis (CAD)

Summary Background Annual low-dose CT screening for lung cancer has been recommended for high-risk individuals, but the necessity of yearly low-dose CT in all eligible individuals is uncertain. This study examined rates of lung cancer in National Lung Screening Trial (NLST) participants who had a negative prevalence (initial) low-dose CT screen to explore whether less frequent screening could be justified in some lower-risk subpopulations. Methods We did a retrospective cohort analysis of data from the NLST, a randomised, multicentre screening trial comparing three annual low-dose CT assessments with three annual chest radiographs for the early detection of lung cancer in high-risk, eligible individuals (aged 55–74 years with at least a 30 pack-year history of cigarette smoking, and, if a former smoker, had quit within the past 15 years), recruited from US medical centres between Aug 5, 2002, and April 26, 2004. Participants were followed up for up to 5 years after their last annual screen. For the purposes of this analysis, our cohort consisted of all NLST participants who had received a low-dose CT prevalence (T0) screen. We determined the frequency, stage, histology, study year of diagnosis, and incidence of lung cancer, as well as overall and lung cancer-specific mortality, and whether lung cancers were detected as a result of screening or within 1 year of a negative screen. We also estimated the effect on mortality if the first annual (T1) screen in participants with a negative T0 screen had not been done. The NLST is registered with ClinicalTrials.gov, number NCT00047385. Findings Our cohort consisted of 26 231 participants assigned to the low-dose CT screening group who had undergone their T0 screen. The 19 066 participants with a negative T0 screen had a lower incidence of lung cancer than did all 26 231 T0-screened participants (371·88 [95% CI 337·97–408·26] per 100 000 person-years vs 661·23 [622·07–702·21]) and had lower lung cancer-related mortality (185·82 [95% CI 162·17–211·93] per 100 000 person-years vs 277·20 [252·28–303·90]). The yield of lung cancer at the T1 screen among participants with a negative T0 screen was 0·34% (62 screen-detected cancers out of 18 121 screened participants), compared with a yield at the T0 screen among all T0-screened participants of 1·0% (267 of 26 231). We estimated that if the T1 screen had not been done in the T0 negative group, at most, an additional 28 participants in the T0 negative group would have died from lung cancer (a rise in mortality from 185·82 [95% CI 162·17–211·93] per 100 000 person-years to 212·14 [186·80–239·96]) over the course of the trial. Interpretation Participants with a negative low-dose CT prevalence screen had a lower incidence of lung cancer and lung cancer-specific mortality than did all participants who underwent a prevalence screen. Because overly frequent screening has associated harms, increasing the interval between screens in participants with a negative low-dose CT prevalence screen might be warranted. Funding None.

Lung cancer screening identifies cancers with heterogeneous behaviors. Some lung cancers will be identified among patients who had prior negative CT screens and upon follow-up scans develop a de novo nodule that was determined to be cancerous. Other lung cancers will be identified among patients who had one or more prior stable positive scans that were not determined to be lung cancer (indeterminate pulmonary nodules), but in follow-up scans was diagnosed with an incidence lung cancer. Using data from the CT arm of the National Lung Screening Trial, this analysis investigated differences in patient characteristics and survival endpoints between prevalence-, interval-, and screen-detected lung cancers, characterized based on sequence of screening results. Lung cancers immediately following a positive baseline (T0), and prior to the T1 screen, formed the prevalence cohort. Interval cancers were diagnosed following a negative screen at any time point prior to the next screening round. Two cohorts of screen-detected lung cancers (SDLC) were identified that had a baseline positive screen that was that was not determined to be lung cancer (i.e., an indeterminate pulmonary nodule), but in follow-up scans was diagnosed with an incidence lung cancer 12 (SDLC1) or 24 (SDLC2) months later. Two other incidence cohorts had screen-detected lung cancers that had baseline negative screen and upon follow-up scans developed a de novo nodule determined to be cancerous at 12 (SDLC3) or 24 (SDLC4) months later. Differences in patient characteristics, progression-free survival (PFS), and overall survival (OS) were assessed. The lung cancer-specific death rate was higher for SDLC3/SDLC4 compared to SDLC1/SDLC2 lung cancers (136.6/1,000 person-years vs. 71.3/1,000 person-years, P < 0.001). Moreover, PFS and OS were significantly lower for SDLC3/SDLC4 compared to SDLC1/SDLC2 (P < 0.004; P < 0.002, respectively). The findings were consistent when stratified by stage and histology. Multivariable Cox proportional models revealed that the SDLC3/SDLC4 case groups were associated with significantly poorer PFS (HR = 1.89; 95% CI 1.31-2.74) and OS (HR = 1.80; 95% CI 1.21-2.67) compared to SDLC1/SDLC2 lung cancers (HR = 1.00). Lung cancer patients who develop a de novo nodule that determined to be cancerous (i.e., at least one negative CT screen prior to cancer diagnosis) had poorer survival outcomes compared to patients who had at least one positive screen prior to cancer diagnosis. As such, the observation that de novo screen-detected are associated with poorer survival could be attributed to faster growing, more aggressive cancers that arose from a lung environment previously lacking focal abnormalities.

Rationale and Objectives The goal was to investigate the effects of choosing between different metrics in estimating the size of pulmonary nodules as a factor both of nodule characterization and of performance of computer aided detection systems, because the latter are always qualified with respect to a given size range of nodules. Materials and Methods This study used 265 whole-lung CT scans documented by the Lung Image Database Consortium (LIDC) using their protocol for nodule evaluation. Each inspected lesion was reviewed independently by four experienced radiologists who provided boundary markings for nodules larger than 3 mm. Four size metrics, based on the boundary markings, were considered: a unidimensional and two bidimensional measures on a single image slice and a volumetric measurement based on all the image slices. The radiologist boundaries were processed and those with four markings were analyzed to characterize the interradiologist variation, while those with at least one marking were used to examine the difference between the metrics. Results The processing of the annotations found 127 nodules marked by all of the four radiologists and an extended set of 518 nodules each having at least one observation with three-dimensional sizes ranging from 2.03 to 29.4 mm (average 7.05 mm, median 5.71 mm). A very high interobserver variation was observed for all these metrics: 95% of estimated standard deviations were in the following ranges for the three-dimensional, unidimensional, and two bidimensional size metrics, respectively (in mm): 0.49–1.25, 0.67–2.55, 0.78–2.11, and 0.96–2.69. Also, a very large difference among the metrics was observed: 0.95 probability-coverage region widths for the volume estimation conditional on unidimensional, and the two bidimensional size measurements of 10 mm were 7.32, 7.72, and 6.29 mm, respectively. Conclusions The selection of data subsets for performance evaluation is highly impacted by the size metric choice. The LIDC plans to include a single size measure for each nodule in its database. This metric is not intended as a gold standard for nodule size; rather, it is intended to facilitate the selection of unique repeatable size limited nodule subsets.

Annual computed tomography (CT) is now widely recommended for lung cancer screening in the United States, although concerns remain regarding the potential harms, including those from overdiagnosis.To examine the effect of airflow limitation on overdiagnosis by comparing lung cancer incidence, histology, and stage shift in a subgroup of the National Lung Screening Trial (NLST).In an NLST subgroup (n = 18,714), screening participants were randomized to annual computed tomography (CT, n = 9,357) or chest radiograph (n = 9,357) screening and monitored for a mean of 6.1 years. After baseline prebronchodilator spirometry, to identify the presence of airflow limitation, 18,475 subjects (99%) were assigned as having chronic obstructive pulmonary disease (COPD) or no COPD. Lung cancer prevalence, incidence, histology, and stage shift were compared after stratification by COPD.For screening participants with spirometric COPD (n = 6,436), there was a twofold increase in lung cancer incidence (incident rate ratio, 2.15; P0.001) and, when compared according to screening arm, no excess lung cancers and comparable histology. Compared with chest radiography, there was also a trend favoring reduced late-stage and increased early-stage cancers in the CT arm (P = 0.054). For those with normal baseline spirometry (n = 12,039), we found an excess of lung cancers during screening in the CT arm, almost exclusively early-stage adenocarcinoma-related cancers (histology shift and overdiagnosis). After correction for these excess cancers, stage shift was marginal (P = 0.077).In the CT arm of the NLST-ACRIN (American College of Radiology Imaging Network) cohort, COPD status was associated with a doubling of lung cancer incidence, no apparent overdiagnosis, and a more favorable stage shift.

Lung cancer screening with low-dose computed tomography (LDCT) has been recommended, based primarily on the results of the NLST (National Lung Screening Trial). The American College of Radiology recently released Lung-RADS, a classification system for LDCT lung cancer screening.To retrospectively apply the Lung-RADS criteria to the NLST.Secondary analysis of a group from a randomized trial.33 U.S. screening centers.Participants were randomly assigned to the LDCT group of the NLST, were aged 55 to 74 years, had at least a 30-pack-year history of smoking, and were current smokers or had quit within the past 15 years.3 annual LDCT lung cancer screenings.Lung-RADS classifications for LDCT screenings. Lung-RADS categories 1 to 2 constitute negative screening results, and categories 3 to 4 constitute positive results.Of 26 722 LDCT group participants, 26 455 received a baseline screening; 48 671 screenings were done after baseline. At baseline, the false-positive result rate (1 minus the specificity rate) for Lung-RADS was 12.8% (95% CI, 12.4% to 13.2%) versus 26.6% (CI, 26.1% to 27.1%) for the NLST; after baseline, the false-positive result rate was 5.3% (CI, 5.1% to 5.5%) for Lung-RADS versus 21.8% (CI, 21.4% to 22.2%) for the NLST. Baseline sensitivity was 84.9% (CI, 80.8% to 89.0%) for Lung-RADS versus 93.5% (CI, 90.7% to 96.3%) for the NLST, and sensitivity after baseline was 78.6% (CI, 74.6% to 82.6%) for Lung-RADS versus 93.8% (CI, 91.4% to 96.1%) for the NLST.Lung-RADS criteria were applied retrospectively.Lung-RADS may substantially reduce the false-positive result rate; however, sensitivity is also decreased. The effect of using Lung-RADS criteria in clinical practice must be carefully studied.National Institutes of Health.

Bronchioloalveolar carcinoma (BAC) is a previously uncommon subset of adenocarcinoma with unique epidemiology, pathology, radiographic presentation, clinical features, and natural history compared with other non-small cell lung cancer (NSCLC) subtypes. Classically, BAC demonstrates a relatively slow growth pattern and indolent clinical course. However, in a subset of patients, rapid growth and death from bilateral diffuse consolidative disease occurs within months of diagnosis or recurrence. Recent data suggest that the incidence of BAC is increasing, notably in younger nonsmoking women. The initial radiographic presentation of BAC varies considerably, from single ground glass opacities (GGOs) or nodules of mixed ground glass and solid attenuation to diffuse consolidative or bilateral multinodular disease. The rising incidence of BAC is also reflected in recent lung cancer screening studies employing helical computed tomography (CT), where the differential diagnosis of GGOs includes not only BAC and overt adenocarcinoma, but inflammatory disease, focal fibrosis, and atypical adenomatous hyperplasia. Because advanced-stage BAC presents as measurable mass lesions in fewer than 50% of cases, determination of radiographic response to therapy by standard criteria is often difficult. Here, we review current data regarding the radiographic imaging of BAC: its radiographic presentations in asymptomatic early-stage and in advanced-stage disease, the functional imaging characteristics of BAC, and challenges of response assessment, including evolving opportunities for computer-assisted image analysis.

RATIONALE AND OBJECTIVES: Integral to the mission of the National Institutes of Health-sponsored Lung Imaging Database Consortium is the accurate definition of the spatial location of pulmonary nodules. Because the majority of small lung nodules are not resected, a reference standard from histopathology is generally unavailable. Thus assessing the source of variability in defining the spatial location of lung nodules by expert radiologists using different software tools as an alternative form of truth is necessary.MATERIALS AND METHODS: The relative differences in performance of six radiologists each applying three annotation methods to the task of defining the spatial extent of 23 different lung nodules were evaluated. The variability of radiologists' spatial definitions for a nodule was measured using both volumes and probability maps (p-map). Results were analyzed using a linear mixed-effects model that included nested random effects.RESULTS: Across the combination of all nodules, volume and p-map model parameters were found to be significant at P CONCLUSION: Radiologists represent the major source of variance as compared with drawing tools independent of drawing metric used. Although the random noise component is larger for the p-map analysis than for volume estimation, the p-map analysis appears to have more power to detect differences in radiologist-method combinations. The standard deviation of the volume measurement task appears to be proportional to nodule volume.

Purpose To determine whether conventional nodule densitometry or analysis based on contrast enhancement maps of indeterminate lung nodules imaged with contrast-enhanced CT can distinguish benign from malignant lung nodules.

Background— We sought to characterize the distinctive pulmonary vascular abnormalities seen on chest radiographs and computed tomography (CT) scans in Eisenmenger syndrome. Methods and Results— Thoracic CT scans, chest radiographs, and clinical data were reviewed for 24 Eisenmenger syndrome patients subdivided into those with interatrial (pretricuspid) versus interventricular or great arterial (posttricuspid) communications and in 14 acyanotic patients with pulmonary arterial hypertension (PAH) and no congenital heart disease. CT scans were scored blindly by 2 thoracic radiologists for the presence and severity of small, tortuous, intrapulmonary vessels, termed “neovascularity,” lobular ground-glass opacification, and systemic perihilar and intercostal vessels. Histopathologic lung sections from 5 patients with Eisenmenger syndrome and from 3 patients with acyanotic PAH were reviewed. Associations between clinical and imaging features were tested by ANOVA and χ 2 tests. Kendall’s rank-order coefficient and the Kruskal-Wallis test were used to test for significant differences in imaging features between Eisenmenger syndrome and acyanotic PAH. Neovascularity on chest radiographs was more common in Eisenmenger syndrome than acyanotic PAH, but differences were not significant. On CT, neovascularity, lobular ground-glass opacification, and hilar and intercostal systemic collaterals were more prevalent in Eisenmenger syndrome, with severity greater in posttricuspid communications. Three previously undescribed vascular lesions were identified histologically in Eisenmenger syndrome: malformed, dilated, muscular arteries within alveolar septa; congested capillaries within alveolar walls; and congested capillaries within the walls of medium-size, muscular pulmonary arteries. These lesions may correspond to the distinctive vascular abnormalities observed on chest radiographs and CT scans. Conclusions— Distinctive vascular lesions on chest radiographs and CT scans in Eisenmenger syndrome appear to be correlated histologically with collateral vessels that develop more extensively with posttricuspid communications.

Rationale and Objectives To investigate the utility of a computer-aided diagnosis (CAD) in the task of differentiating malignant nodules from benign nodules based on quantitative features extracted from volumetric thin section CT image data acquired before and after the injection of contrast media. Materials and Methods 35 volumetric CT datasets of solitary pulmonary nodules (SPN) with proven diagnoses (19 malignant/16 benign) were contoured by a thoracic radiologist. All studies had at least a baseline series obtained without contrast media and at least one series following an intravenous contrast injection at 45, 90, 180, and 360 seconds. Two separate contours were created for each nodule: one including only the solid portion and another including the ground-glass component, if any, of the nodule. For each contour 31 features were calculated that measured the attenuation, shape, and enhancement of the nodule due to the injection of contrast. These features were input into a feature selection step and three different classifiers to determine if the diagnosis could be predicted from the resulting feature vector. In addition, observer input was introduced to two of the classifiers as an a priori probability of malignancy and the resulting performance was compared. Training and testing was conducted in a resubstitution and leave-one-out fashion and performance was evaluated using ROC analysis. Results In a leave-one-out testing methodology, the classifiers achieved areas under the ROC curves (A Z ) that ranged from 0.69 to 0.92. A classifier based on logistic regression performed the best with an A Z of 0.92 while a classifier based on quadratic discriminant analysis performed the poorest (A Z , 0.69). The A Z increased when using a priori observer input in most cases reaching a maximum of 0.95. Conclusion Based on this initial work with a limited number of nodules in our dataset, it appears that CAD using volumetric and contrast-enhanced data has the potential to assist radiologists in the task of differentiating solitary pulmonary nodules and in the management of these patients. Further studies with an increased number of patients are required to validate these results.

The American College of Radiology Imaging Network (ACRIN) is a cooperative group funded by the National Cancer Institute and dedicated to developing and conducting clinical trials of diagnostic imaging and image-guided treatment technologies. ACRIN's six disease site committees are responsible for developing scientific strategies and resultant trials within the framework of ACRIN's five key hypotheses: (a) Screening and early detection with imaging can reduce cancer-specific mortality. (b) Less invasive image-guided therapeutic methods can reduce the mortality and morbidity associated with treating cancer. (c) Molecular-based physiologic and functional imaging can improve the diagnosis and staging of cancer, thus improving treatment. (d) Functional imaging can portray the effectiveness of treatment earlier and more accurately, thus reducing mortality and improving the likelihood of a cure. (e) Informatics and other "smart systems" can improve the evaluation of patients with cancer, thus leading to better and more effective treatments. This article details ACRIN's research strategy according to disease site through the year 2007.

Rationale and Objectives The objective is to study the incremental effects of using a computer-aided lung nodule detection (CAD) system on the performance of a large pool of observers. Materials and Methods A set of eight thin-section computed tomographic data sets with limited longitudinal coverage, containing a total of 22 lung nodules, was analyzed by using the automated nodule detection system. When applied to all eight cases, the CAD system alone achieved a detection rate of 86.4%, with 2.64 false-positive results per case. This study included 202 observers at a national radiology meeting: 39 thoracic radiologists, 95 nonthoracic radiologists, and 68 nonradiologists. Each participant read from one to eight cases in random order, first without and then with CAD system output available. Observer performance in nodule detection was measured before and after CAD was made available. Differences in performance of groups of observers before and after CAD were tabulated by mean, median, and SD in detection rate and number of false-positive results and tested by using nonparametric methods. Results In an analysis involving only the first randomly selected case read by all 202 participants, there were statistically significant increases in nodule detection rates and numbers of false-positive results for all types of observers. There was a significant difference in detection rates between radiologists and nonradiologists before CAD, but after CAD, there was no significant difference in detection rates between these observer types. In a second analysis involving 13 participants who read all eight cases, mean detection rates were 64.0% before CAD and 81.9% after CAD. Mean numbers of false-positive results were 0.144 per case before CAD and 0.173 after CAD. Conclusion In a large observer study, use of a CAD system for nodule detection resulted in an incremental increase in detection rate, but also led to an increase in number of false-positive results. Also, CAD appears to be an equalizer of detection rates between observers of different levels of experience.

Rationale and objectives We sought to investigate the utility of a computer-aided diagnosis in the task of differentiating malignant nodules from benign nodules based on single thin-section computed tomography image data. Materials and methods Eighty-one thin-section computed tomography data sets of solitary pulmonary nodules with proven diagnoses (48 malignant and 33 benign) were contoured manually on a single representative slice by a thoracic radiologist (>10 years of experience). Two separate contours were created for each nodule, one including only the solid portion of the nodule and one including any ground-glass components. For each contour 75 features were calculated that measured the attenuation, shape, and texture of the nodule. These features were than input into a feature selection step and four different classifiers to determine if the diagnosis could be predicted from the feature vector. Training and testing was conducted in a resubstitution and leave-one-out fashion and performance was evaluated using ROC techniques. Results In a leave-one-out testing methodology the classifiers resulted with areas under the ROC curve (A Z ) that ranged from 0.68 to 0.92. When evaluating with resubstitution the A Z ranged from 0.93 to 1.00. Conclusion Computer-aided diagnosis has the potential to assist radiologists in the task of differentiating solitary pulmonary nodules and in the management of these patients.

Rationale and objectives To investigate the performance of observers with different levels of experience in distinguishing between benign and malignant solitary pulmonary nodules (SPN) on CT, and to determine the effects on interpretation of three different conditions: image data alone, the addition of clinical data, and the addition of output from a computer-aided diagnosis (CAD) system. Materials and methods 28 thin-section CT datasets of SPNs with proven diagnoses (15 malignant and 13 benign) were used to measure observer performance. Readers were categorized according to their experience and read the cases in random order. For each case readers were asked to assign a level of confidence on a scale from 0.0–1.0 (0.0 benign, 1.0 malignant) for the diagnosis of the nodule. Each reader scored the cases based on review of image data alone (phase 1), then with limited clinical data (phase 2), and finally with CAD output (phase 3). To assess performance, multiple reader multiple case (MRMC) receiver operating characteristic (ROC) analysis was used. Results 2 thoracic radiologists, 1 thoracic radiology fellow, 2 nonthoracic radiologists, and 3 radiology residents read the cases. The average area under the ROC curve for all readers (A z ) at each stage was 0.68, 0.75, and 0.81, for image data alone, with clinical data, and with CAD output respectively. The difference in performance between phases (2 and 3) and (1 and 3) was significantly different ( P = 0.018 and P = 0.020). However, the difference between phases (1 and 2) was not significantly different ( P = 0.155). Conclusion Diagnostic performance increased significantly with the addition of CAD output. With further validation CAD output may play a significant role in SPN management.

To stimulate the advancement of computer-aided diagnostic (CAD) research for lung nodules in thoracic computed tomography (CT), the National Cancer Institute launched a cooperative effort known as the Lung Image Database Consortium (LIDC). The LIDC is composed of five academic institutions from across the United States that are working together to develop an image database that will serve as an international research resource for the development, training, and evaluation of CAD methods in the detection of lung nodules on CT scans. Prior to the collection of CT images and associated patient data, the LIDC has been engaged in a consensus process to identify, address, and resolve a host of challenging technical and clinical issues to provide a solid foundation for a scientifically robust database. These issues include the establishment of (a) a governing mission statement, (b) criteria to determine whether a CT scan is eligible for inclusion in the database, (c) an appropriate definition of the term qualifying nodule, (d) an appropriate definition of "truth" requirements, (e) a process model through which the database will be populated, and (f) a statistical framework to guide the application of assessment methods by users of the database. Through a consensus process in which careful planning and proper consideration of fundamental issues have been emphasized, the LIDC database is expected to provide a powerful resource for the medical imaging research community. This article is intended to share with the community the breadth and depth of these key issues.

Fifty-four solitary pulmonary nodules with diagnosis established with pathological or 2 years of radiographic followup were imaged using CT with a high-resolution protocol without contrast. A thoracic radiologist manually outlined each nodule on a single slice that they chose as most representative of it. Nineteen two-dimensional attenuation, size, and shape features were calculated based on this slice and were input into a OneR and logistic regression classifier. Classifier performance was measured using resubstitution, leave-one-out, and a split training set (42 cases) and test set (12 cases). When testing the classifier using resubstitution OneR achieved 90.9% sensitivity, 71.4% specificity, and 83.3% accuracy while the logistic regression classifier achieved 81.8% sensitivity, 85.7% specificity, and 83.3% accuracy. The OneR classifier in a leave-one-out testing methodology was able to achieve 90.9% sensitivity, 66.7% specificity, and 81.4% accuracy compared to 63.6% sensitivity, 52.4% specificity, and 59.3% accuracy of the logistic regression classifier. When separate training/testing sets were made OneR resulted in 85.7% sensitivity, 60% specificity, and 75% accuracy while logistic regression resulted in 71.4% sensitivity, 40% specificity, and 58.3% accuracy. Thus a simple classifier like OneR can perform better than a more complex classifier.

Rationale and Objectives. A streamlined process of care supported by technology and imaging may be effective in managing the overall healthcare process and costs. This study examined the effect of an imaging-based electronic process of care on costs and rates of hospitalization, emergency room (ER) visits, specialist diagnostic referrals, and patient satisfaction. Materials and Methods. A healthcare process was implemented for an employer group, highlighting improved patient access to primary care plus routine use of imaging and teleconsultation with diagnostic specialists. An electronic infrastructure supported patient access to physicians and communication among healthcare providers. The employer group, a self-insured company, manages a healthcare plan for its employees and their dependents: 4,072 employees were enrolled in the test group, and 7,639 in the control group. Outcome measures for expenses and frequency of hospitalizations, ER visits, traditional specialist referrals, primary care visits, and imaging utilization rates were measured using claims data over 1 year. Homogeneity tests of proportions were performed with a chi-square statistic, mean differences were tested by two-sample t -tests. Patient satisfaction with access to healthcare was gauged using results from an independent firm. Results. Overall per member/per month costs post-implementation were lower in the enrolled population ($126 vs $160), even though occurrence of chronic/expensive diseases was higher in the enrolled group (18.8% vs 12.2%). Lower per member/per month costs were seen for inpatient ($33.29 vs $35.59); specialist referrals ($21.36 vs $26.84); and ER visits ($3.68 vs $5.22). Moreover, the utilization rate for hospital admissions, ER visits, and traditional specialist referrals were significantly lower in the enrolled group, although primary care and imaging utilization were higher. Comparison to similar employer groups showed that the company’s costs were lower than national averages ($119.24 vs $146.32), indicating that the observed result was not attributable to normalization effects. Patient satisfaction with access to healthcare ranked in the top 21st percentile. Conclusion. A streamlined healthcare process supported by technology resulted in higher patient satisfaction and cost savings despite improved access to primary care and higher utilization of imaging.

An automated system was developed for detecting lung micronodules on thin-section computed tomographic images and was applied to data from 15 subjects with 77 lung nodules. The automated system, without user interaction, achieved a sensitivity of 100% for nodules (>3 mm in diameter) and 70% for micronodules (

Rationale and Objectives The purpose of this study was to determine the electronic requirements for supporting evidence-based radiology in today's medical environment. Materials and Methods A software engineering technique, use case modeling, was performed for several clinical settings to determine the use of imaging and its role in evidence-based practice, with particular attention to issues relating to data access and the usage of clinical information. From this basic understanding, the analysis was extended to encompass evidence-based radiologic research and teaching. Results The analysis showed that a system supporting evidence-based radiology must (a) provide a single point of access to multiple clinical data sources so that patient data can be readily used and incorporated into comprehensive radiologic consults and (b) provide quick access to external evidence in the way of similar patient cases and published medical literature, thus supporting evidence-based practice. Conclusion Information infrastructures that aim to support evidence-based radiology not only must address issues related to the integration of clinical data from heterogeneous databases, but must facilitate access and filtering of patient data in order to improve radiologic consultation.

After long debate about the worth of screening for lung cancer, and even about the merits of doing a randomized trial to address the issue, initial results from the first large-scale randomized controlled trial ever to show a reduction in lung cancer mortality associated with screening were announced to the public last November 4.1 The story behind this historical first is informative, and the editor of the Journal of Medical Screening played a role in that story. But more of that later.

Computed tomography (CT) screening for lung cancer has been associated with a high frequency of false positive results because of the high prevalence of indeterminate but usually benign small pulmonary nodules. The acceptability of reducing false-positive rates and diagnostic evaluations by increasing the nodule size threshold for a positive screen depends on the projected balance between benefits and risks.We examined data from the National Lung Screening Trial (NLST) to estimate screening CT performance and outcomes for scans with nodules above the 4mm NLST threshold used to classify a CT screen as positive. Outcomes assessed included screening results, subsequent diagnostic tests performed, lung cancer histology and stage distribution, and lung cancer mortality. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for the different nodule size thresholds. All statistical tests were two-sided.In 64% of positive screens (11598/18141), the largest nodule was 7 mm or less in greatest transverse diameter. By increasing the threshold, the percentages of lung cancer diagnoses that would have been missed or delayed and false positives that would have been avoided progressively increased, for example from 1.0% and 15.8% at a 5 mm threshold to 10.5% and 65.8% at an 8 mm threshold, respectively. The projected reductions in postscreening follow-up CT scans and invasive procedures also increased as the threshold was raised. Differences across nodules sizes for lung cancer histology and stage distribution were small but statistically significant. There were no differences across nodule sizes in survival or mortality.Raising the nodule size threshold for a positive screen would substantially reduce false-positive CT screenings and medical resource utilization with a variable impact on screening outcomes.

PET/CT with the glucose analog (18)F-FDG has several potential applications for monitoring tumor response to therapy in patients with non-small cell lung cancer (NSCLC). A prerequisite for many of these applications is detailed knowledge of the repeatability of quantitative parameters derived from (18)F-FDG PET/CT studies.The repeatability of the (18)F-FDG signal was evaluated in 2 prospective multicenter trials. Patients with advanced NSCLC (tumor stage III-IV) underwent two (18)F-FDG PET/CT studies while not receiving therapy. Tumor (18)F-FDG uptake was quantified by measurement of the maximum standardized uptake value within a lesion (SUVmax) and the average SUV within a small volume of interest around the site of maximum uptake (SUVpeak). Analysis was performed for the lesion in the chest with the highest (18)F-FDG uptake and a size of at least 2 cm (target lesion) as well as for up to 6 additional lesions per patient. Repeatability was assessed by Bland-Altman plots and calculation of 95% repeatability coefficients (RCs) of the log-transformed SUV differences.Test-retest repeatability was assessed in 74 patients (34 from the ACRIN 6678 trial and 40 from the Merck MK-0646-008 trial). SUVpeak was 11.57 ± 7.89 g/mL for the ACRIN trial and 6.89 ± 3.02 for the Merck trial. The lower and upper RCs were -28% (95% confidence interval [CI], -35% to -23%) and +39% (95% CI, 31% to 54%) in the ACRIN trial, indicating that a decrease of SUVpeak by more than 28% or an increase by more than 39% has a probability of less than 2.5%. The corresponding RCs from the Merck trial were -35% (95% CI, -42% to -29%) and +53% (95% CI, 41% to 72%). Repeatability was similar for SUVmax of the target lesion, averaged SUVmax, and averaged SUVpeak of up to 6 lesions per patient.The variability of repeated measurements of tumor (18)F-FDG uptake in patients with NSCLC is somewhat larger than previously reported in smaller single-center studies but comparable to that of gastrointestinal malignancies in a previous multicenter trial. The variability of measurements supports the definitions of tumor response according to PET Response Criteria in Solid Tumors.

A retrospective medical record review was performed to study the differences in clinical risk profiles and the relationships between test results versus management for suspected pulmonary thromboembolism (TE) in patients undergoing either radionuclide ventilation perfusion (V/Q) scans or pulmonary computed tomographic angiography (CTA), as the initial test. Data of 138 consecutive V/Q patients were compared with that of 149 consecutive CTA patients during equivalent 6-month intervals before and after the introduction of CTA. Information on risk factors, signs and symptoms, all diagnostic test results, and the relationships between the test results and ultimate physician management were collected and analyzed. V/Q results predicted physician management in all patients with high probability scans and 91% with normal to low probability scans. There were 35 patients with indeterminate V/Q scans--43% of these patients were managed without any other diagnostic test. CTA results predicted management in all patients with positive studies and 99% of patients with negative studies. In contrast to the V/Q cohort, only seven CTA studies were inconclusive--additional diagnostic tests determined management in all but one case. Compared with V/Q, CTA has fewer indeterminate results, is more directly reflective of management, and reduces the number of patients managed with inconclusive data.

Summary: This consensus statement by the Society of Thoracic Radiology is a summary of the current understanding of low dose computed tomography (CT) for screening for lung cancer. Lung cancer is the most common fatal malignancy in the industrialized world. Unlike the next three most common cancers, screening for lung cancer is not currently recommended by cancer organizations. Improvements in CT technology make lung screening feasible. Early prevalence data indicate that about two-thirds of lung cancers that are detected by CT screening are at an early stage. Other data support the postulate that patients with lung cancers detected at this early stage have better rates of survival. Whether this will translate into an improved disease specific mortality is yet to be demonstrated. The suggested technical protocols, selection criteria, and method of handling the numerous benign nodules that are detected are discussed. It is the consensus of this committee that mass screening for lung cancer with CT is not currently advocated. Suitable subjects who wish to participate should be encouraged to do so in controlled trials, so that the value of CT screening can be ascertained as soon as possible.

The Scientific Assembly and Annual Meeting of the Radiological Society of North America includes many exciting scientific sessions and educational exhibits. None is greeted with more anticipation than the Sunday afternoon Image Interpretation Session. Again this year, the session will include 10 unknown cases, two each in the classic radiologic categories of genitourinary, gastrointestinal, chest, musculoskeletal, and neurologic imaging. The cases were chosen to challenge the panelists and to provide a stimulating learning experience for the attendees. Representative images are reprinted here along with pertinent clinical case histories. The panelists have been asked to describe the imaging abnormalities and the pertinent negative findings in each case. Each panelist will construct a differential diagnosis and enumerate the points in favor or against each diagnosis before committing to a most likely final diagnosis. Additional tests or treatment options may be suggested, as emphasis is placed on the pract...

The purpose of this research is to characterize solitary pulmonary nodules as benign or malignant based on quantitative measures extracted from high resolution CT (HRCT) images. High resolution CT images of 31 patients with solitary pulmonary nodules and definitive diagnoses were obtained. The diagnoses of these 31 cases (14 benign and 17 malignant) were determined from either radiologic follow-up or pathological specimens. Software tools were developed to perform the classification task. On the HRCT images, solitary nodules were identified using semiautomated contouring techniques. From the resulting contours, several quantitative measures were extracted related to each nodule's size, shape, attenuation, distribution of attenuation, and texture. A stepwise discriminant analysis was performed to determine which combination of measures were best able to discriminate between the benign and malignant nodules. A linear discriminant analysis was then performed using selected features to evaluate the ability of these features to predict the classification for each nodule. A jackknifed procedure was performed to provide a less biased estimate of the linear discriminator's performance. The preliminary discriminant analysis identified two different texture measures--correlation and difference entropy--as the top features in discriminating between benign and malignant nodules. The linear discriminant analysis using these features correctly classified 28/31 cases (90.3%) of the training set. A less biased estimate, using jackknifed training and testing, yielded the same results (90.3% correct). The preliminary results of this approach are very promising in characterizing solitary nodules using quantitative measures extracted from HRCT images. Future work involves including contrast enhancement and three-dimensional measures extracted from volumetric CT scans, as well as the use of several pattern classifiers.

In the United States, Lung Cancer is responsible for more cancer deaths than the next four cancers combined. In addition, the 5 year survival rate for lung cancer patients has not improved over the past 40 to 50 years. To combat this deadly disease, in 2002 the National Cancer Institute launched a very large Randomized Control Trial called the National Lung Screening Trial (NLST). This trial would randomize subjects who had substantial risk of lung cancer (due to age and smoking history) into either a Chest X-ray arm or a low dose CT arm. In November 2010, the National Cancer Institute announced that the NLST had demonstrated 20% fewer lung cancer deaths among those who were screened with low-dose CT than with chest X-ray. In December 2013, the US Preventive Services Task Force recommended the use of Lung Cancer Screening using low dose CT and a little over a year later (Feb. 2015), CMS announced that Medicare would also cover Lung Cancer Screening using low dose CT. Thus private and public insurers are required to provide Lung Cancer Screening programs using CT to the appropriate population(s). The purpose of this Symposium is to inform medical physicists and prepare them to support the implementation of Lung Screening programs. This Symposium will focus on the clinical aspects of lung cancer screening, requirements of a screening registry for systematically capturing and tracking screening patients and results (such as required Medicare data elements) as well as the role of the medical physicist in screening programs, including the development of low dose CT screening protocols. Learning Objectives: 1. To understand the clinical basis and clinical components of a lung cancer screening program, including eligibility criteria and other requirements. 2. To understand the data collection requirements, workflow, and informatics infrastructure needed to support the tracking and reporting components of a screening program. 3. To understand the role of the medical physicist in implementing Lung Cancer Screening protocols for CT, including utilizing resources such as the AAPM Protocols and the ACR Designated Lung Screening Center program. UCLA Department of Radiology has an Institutional research agreement with Siemens Healthcare; Dr. McNitt-Gray has been a recipient of Research Support from Siemens Healthcare in the past. Dr. Aberle has been a Member of Advisory Boards for the LUNGevity Foundation (2011-present) and Siemens Medical Solutions. (2013).

Importance Screening for lung cancer has the potential to reduce mortality, but in addition to detecting aggressive tumors, screening will also detect indolent tumors that otherwise may not cause clinical symptoms. These overdiagnosis cases represent an important potential harm of screening because they incur additional cost, anxiety, and morbidity associated with cancer treatment. Objective To estimate overdiagnosis in the National Lung Screening Trial (NLST). Design, Setting, and Participants We used data from the NLST, a randomized trial comparing screening using low-dose computed tomography (LDCT) vs chest radiography (CXR) among 53 452 persons at high risk for lung cancer observed for 6.4 years, to estimate the excess number of lung cancers in the LDCT arm of the NLST compared with the CXR arm. Main Outcomes and Measures We calculated 2 measures of overdiagnosis: the probability that a lung cancer detected by screening with LDCT is an overdiagnosis ( PS ), defined as the excess lung cancers detected by LDCT divided by all lung cancers detected by screening in the LDCT arm; and the number of cases that were considered overdiagnosis relative to the number of persons needed to screen to prevent 1 death from lung cancer. Results During follow-up, 1089 lung cancers were reported in the LDCT arm and 969 in the CXR arm of the NLST. The probability is 18.5% (95% CI, 5.4%-30.6%) that any lung cancer detected by screening with LDCT was an overdiagnosis, 22.5% (95% CI, 9.7%-34.3%) that a non–small cell lung cancer detected by LDCT was an overdiagnosis, and 78.9% (95% CI, 62.2%-93.5%) that a bronchioalveolar lung cancer detected by LDCT was an overdiagnosis. The number of cases of overdiagnosis found among the 320 participants who would need to be screened in the NLST to prevent 1 death from lung cancer was 1.38. Conclusions and Relevance More than 18% of all lung cancers detected by LDCT in the NLST seem to be indolent, and overdiagnosis should be considered when describing the risks of LDCT screening for lung cancer.

Purpose Our goal was to develop a protocol and image-processing methods to quantitate both bronchial and lung attenuation changes in patients imaged with helical high-resolution CT (HRCT). Method Human subjects underwent helical HRCT at two suspended breath-hold conditions, functional residual capacity and residual volume, at baseline and following methacholine-induced bronchoprovocation. A semiautomated contouring program was used to define anatomically like bronchi and axial lung sections from the different physiologic sequences, from which automated measurements of area, shape, and attenuation were made. Because the gray level threshold for contouring directly affects the measured area of an anatomic structure, two types of evaluation studies were performed. These included in vivo measurements using baseline parameters of human subjects as the standard of reference and in vitro measurements of a CT phantom designed to simulate the air-soft tissue interfaces of bronchi. Results Phantom tests showed that the minimum difference between actual and measured areas of holes occurred at a threshold of -500 HU. The smallest diameter holes were most sensitive to changes in threshold value. However, although absolute area measurements of both simulated and human bronchi varied with threshold level, the percent changes in airway areas between baseline and bronchoprovocation sequences were relatively stable at any given threshold. Conclusion These image-processing tools provide reproducible measurements of area as well as attenuation characteristics of pulmonary structures and may offer insights into the practical use of functional imaging in evaluating conditions of airflow obstruction.

Conventional incremental CT has for many years been useful in the fortuitous diagnosis of pulmonary thromboembolic disease, allowing for visualization of both the central occluding thrombus and the pleuroparenchymal sequelae. Unfortunately, the slow data acquisition times precluded the inclusion of conventional CT in diagnostic algorithms for the diagnosis of this disease. The development and increasing availability of fast scanning techniques, namely helical (spiral) CT and electron-beam CT, now provide a noninvasive means of consistently and accurately demonstrating acute and chronic pulmonary arterial thrombus to the segmental level. CT has the added advantage over ventilation-perfusion scanning and pulmonary angiography of depicting unsuspected intrathoracic disease that may account for the patient's presenting illness.

The use of traditional pulmonary function tests (PFT) to assess airflow obstruction is limited by the fact that these measurements are global and can neither express the potential regional heterogeneity nor localize the distribution of airways involved. High-resolution computed tomography (HRCT) is an established technique for the detailed evaluation of the pulmonary parenchyma. The development of helical scanning techniques enables volume datasets to be acquired through regions of interest in a single breathhold. Acquired during different phases of suspended respiration or under different physiologic conditions, these image sets provide potential insights into the relationships between airway structure and function. Volumetric data in also amenable to advance image processing algorithms enabling more precise characterization of bronchial morphology. This review considers the challenge of assessing cardiopulmonary function due to the complexity of the heart and lungs. The considerable insights that have been gained about airway physiology in animal experimentation are discussed. Studies investigating the role of volume acquisition's Helical CT and Electron Beam CT techniques of evaluating functional small airways function in humans are reviewed in detail and potential methods for analysis of airflow obstruction and detection of airway reactivity are extensively considered.

The national cost of cancer care is projected to reach $173 billion by 2020, increasing from $125 billion in 2010. This steep upward cost trajectory has placed enormous an financial burden on patients, their families, and society as a whole and raised major concern about the ability of the health care system to provide and sustain high-quality cancer care. To better understand the cost drivers of cancer care and explore approaches that will mitigate the problem, the National Cancer Policy Forum of the Institute of Medicine held a workshop entitled “Delivering Affordable Cancer Care in the 21st Century” in October 2012. Workshop participants included bioethicists, health economists, primary care physicians, and medical, surgical, and radiation oncologists, from both academic and community settings. All speakers expressed a sense of urgency about the affordability of cancer care resulting from the future demographic trend as well as the high cost of emerging cancer therapies and rapid diffusion of new technologies in the absence to evidence indicating improved outcomes for patients. This article is our summary of presentations at the workshop that highlighted the overuse and underuse of screening, treatments, and technologies throughout the cancer care continuum in oncology practice in the United States. J Clin Oncol 31:4151-4157. © 2013 by American Society of Clinical Oncology

Rationale and Objectives. We quantified differences in the detection of simulated lung nodules on computed radiographs on the basis of variations in nodule size, local contrast, body habitus (global contrast), and exposure. Methods. A step-wedge phantom was developed to simulate the attenuation ranges of the lung, retrocardiac, and subdiaphragmatic regions of the adult human chest. Additional Lucite wedges were used to simulate two different body thicknesses and to provide variable structural noise. Soft-tissue-equivalent nodules of 3-mm and 5-mm diameter that resulted in 10% differences in attenuation from lung equivalence were embedded in lung-equivalent material. By superimposing the sheets in various positions, 84 unique nodule configurations containing eight nodules per image were exposed on a computed radiography system. Computed radiographs were acquired at two different exposures approximating standard exposure and underexposure. For each resulting phantom image, seven observers scored the presence or absence of a nodule within individual cells of a 5 × 5 grid matrix. Results. True-positive fractions for 3-mm-diameter nodules were very low across all conditions. True-positive fractions for 5-mm-diameter nodules varied from 0.23 to 0.98. Significant differences in the conspicuity of 5-mm nodules depended on differences in phantom thickness and differences in the locations of nodules within lung-, retrocardiac-, or subdiaphragmatic-equivalent regions. Accuracy in detecting nodules was significantly lower at lower exposures when nodules were located in the subdiaphragmatic-equivalent region. Conclusion. On computed radiographs, small nodules (5-mm diameter) can be reliably detected when they are located in areas of high or moderate surrounding local contrast, such as the lung or mediastinal regions. Detection of nodules decreases in regions of lower optical density corresponding to the subdiaphragmatic regions of the chest. The decrease in nodule detectability is greatest under conditions that simulate large body thickness and underexposure.

A multimedia medical database model and prototype is described for supporting a timeline-based presentation of information. The database links image and text data in a way that permits users to look at medical information in a single unified view. Various visualization programs permit the user to view data in various ways, including full image views, graphs, and tables. Our technology is applied for proof-of-concept to two areas: thoracic oncology and thermal tumor ablation therapy of the brain. This effort is part of the multidisciplinary KMeD project in collaboration with medical research and clinical treatment projects at UCLA.

A prototype multimedia medical database has been developed to provide image and textual data for thoracic oncology patients undergoing treatment of advanced malignancies. The database integrates image data from the hospital picture archiving and communication system with textual reports from the radiology information system, alphanumeric data contained in the hospital information system, and other electronic medical data. The database presents information in a timeline format and also contains visualization programs that permit the user to view and annotate radiographic measurements in tabular or graphic form. The database provides an efficient and intuitive display of the changing status of oncology patients. The ability to integrate, manage, and access relevant multimedia information may substantially enhance communication among distributed multidisciplinary health care providers and may ensure greater consistency and completeness of patient-related data.

Now the leading subtype of lung cancer, adenocarcinoma received a new classification in 2011. For tumors categorized previously as bronchioloalveolar carcinoma (BAC), criteria and terminology had not been uniform, so the 2011 classification provided four new terms: (a) adenocarcinoma in situ (AIS), representing histopathologically a small (≤3-cm), noninvasive lepidic growth, which at computed tomography (CT) is usually nonsolid; (b) minimally invasive adenocarcinoma, representing histopathologically a small (≤3-cm) and predominantly lepidic growth that has 5-mm or smaller invasion, which at CT is mainly nonsolid but may have a central solid component of up to approximately 5 mm; (c) lepidic predominant nonmucinous adenocarcinoma, representing histopathologically invasive adenocarcinoma that shows predominantly lepidic nonmucinous growth, which at CT is usually part solid but may be nonsolid or occasionally have cystic components; and (d) invasive mucinous adenocarcinoma, histopathologically showing lepidic growth as its predominant component, which at CT varies widely from solid to mostly solid to part solid to nonsolid and may be single or multiple (when multifocal, it was formerly called multicentric BAC). In addition, new histopathologic subcategories of acinar, papillary, micropapillary, and solid predominant adenocarcinoma are now described, all as nonmucinous, predominantly invasive, may include a small lepidic component, and at CT are usually solid but may include a small nonsolid component. The micropapillary subtype has a poorer prognosis than the other subtypes. In addition, molecular genetic correlations for the subcategories of adenocarcinoma of the lung are now a topic of increasing interest. As the new classification enters common use, further descriptions of related correlations can be anticipated.

Pathology and radiology form the core of cancer diagnosis, yet the workflows of both specialties remain ad hoc and occur in separate "silos," with no direct linkage between their case accessioning and/or reporting systems, even when both departments belong to the same host institution. Because both radiologists' and pathologists' data are essential to making correct diagnoses and appropriate patient management and treatment decisions, this isolation of radiology and pathology workflows can be detrimental to the quality and outcomes of patient care. These detrimental effects underscore the need for pathology and radiology workflow integration and for systems that facilitate the synthesis of all data produced by both specialties. With the enormous technological advances currently occurring in both fields, the opportunity has emerged to develop an integrated diagnostic reporting system that supports both specialties and, therefore, improves the overall quality of patient care.

Intrathoracic calcifications occur in a wide variety of disorders. Although they are usually harmless sequelae of remote processes, calcifications provide important information for establishing the diagnosis or for evaluating the progression of known disease. They may arise in the pulmonary parenchyma, mediastinum, hilar and mediastinal lymph nodes, pleura, chest wall, or any combination of these structures. The cause of the calcification may be determined by means of the location and pattern of the calcifications within the lung parenchyma and knowledge of the associated clinical features. Calcifications in the thorax are frequently manifestations of previous infectious processes. Less often, they may be due to neoplasms, metabolic disorders, occupational exposure, or previous medical therapy. Large intrathoracid calcifications are usually identified on conventional chest radiographs; detection of smaller calcifications may require use of other imaging modalities, such as dual-energy digital radiography, fluoroscopy, radionuclide scanning, computed tomography (CT), and high-resolution CT.

Background Randomized controlled trials (RCTs) evaluating cancer screening modalities usually employ cause-specific mortality as their primary endpoint. Because death certificate cause of death can be inaccurate, RCTs frequently use review committees to assign an underlying cause of death. We describe the National Lung Screening Trial's (NLST's) death review approach, the Endpoint Verification Process (EVP), which strives to minimize errors in assignment of cause of death due to lung cancer. Methods Deaths selected for review include those with a notation of lung cancer on the death certificate and those occurring among participants ever diagnosed with lung cancer. Other criteria that trigger death review include, but are not limited to, death within 6 months of a screen suspicious for lung cancer and death within 60 days of certain diagnostic evaluation procedures associated with a screen suspicious for lung cancer or a lung cancer diagnosis. EVP requires concordance on whether death was due to lung cancer. Deaths are first reviewed by the EVP chair. If concordance is not achieved, the death is next reviewed by an Endpoint Verification Team (EVT) member. If concordance between the chair- and member-assigned cause of death is not achieved, the death is next reviewed by a group of at least three EVT members. Cause of death is assigned at the step in which concordance was achieved, or if necessary, at the team review. Conclusions NLST's EVP is designed to produce a highly accurate count of lung cancer deaths.

Ample studies suggest that the cyclooxygenase-2 (COX-2)/prostaglandin E2 (PGE2) pathway plays a pivotal role in carcinogenesis and that COX-2 inhibition may help prevent lung cancer. Therefore, we conducted a randomized, double-blind, placebo-controlled trial of the COX-2–selective inhibitor celecoxib (400 mg bid for 6 months) in former-smokers (age ≥ 45, ≥ 30 pack-years of smoking, ≥ 1 year of sustained abstinence from smoking). We assessed the impact of celecoxib on cellular and molecular events associated with lung cancer pathogenesis; the primary endpoint was bronchial Ki-67 labeling index (Ki-67 LI) after 6 months of treatment. Of 137 randomized subjects, 101 completed both baseline and 6-month bronchoscopies and were evaluable for the primary endpoint analysis. The beneficial effect on Ki-67 LI was greater in the celecoxib arm (versus placebo) in a mixed-effects analysis (P = 0.0006), and celecoxib significantly decreased Ki-67 LI by an average of 34%, whereas placebo increased Ki-67 LI by an average of 3.8% (P = 0.04; t test). In participants who crossed over to the other study arm at 6 months (all of whom had received 6 months of celecoxib at the end of a 12 months treatment period), the decreases in Ki-67 LI correlated with a reduction and/or resolution of lung nodules on computed tomography. Celecoxib significantly reduced plasma c-reactive protein and interleukin-6 mRNA and protein and increased 15(S)-hydroxy-eicosatetraenoic acid levels in bronchoalveolar lavage (BAL) samples. The baseline ratio of COX-2 to 15-hydroxyprostaglandin dehydrogenase mRNA in BAL cells was a significant predictive marker of Ki-67 response to celecoxib (P = 0.002). Our collective findings support the continued investigation of celecoxib for lung cancer chemoprevention in former smokers at a low risk of cardiovascular disease. Cancer Prev Res; 4(7); 984–93. ©2011 AACR.