Automated identification of lung nodules

A system that can automatically detect nodules within lung images may assist expert radiologists in interpreting the abnormal patterns as nodules in 2D CT lung images. A system is presented that can automatically identify nodules of various sizes within lung images. The pattern classification method is employed to develop the proposed system. A random forest ensemble classifier is formed consisting of many weak learners that can grow decision trees. The forest selects the decision that has the most votes. The developed system consists of two random forest classifiers connected in a series fashion. A subset of CT lung images from the LIDC database is employed. It consists of 5721 images to train and test the system. There are 411 images that contained expert- radiologists identified nodules. Training sets consisting of nodule, non-nodule, and false-detection patterns are constructed. A collection of test images are also built. The first classifier is developed to detect all nodules. The second classifier is developed to eliminate the false detections produced by the first classifier. According to the experimental results, a true positive rate of 100%, and false positive rate of 1.4 per lung image are achieved.

Lung nodules detection by ensemble classification

A method is presented that achieves lung nodule detection by classification of nodule and non-nodule patterns. It is based on random forests which are ensemble learners that grow classification trees. Each tree produces a classification decision, and an integrated output is calculated. The performance of the developed method is compared against that of the support vector machine and the decision tree methods. Three experiments are performed using lung scans of 32 patients including thousands of images within which nodule locations are marked by expert radiologists. The classification errors and execution times are presented and discussed. The lowest classification error (2.4%) has been produced by the developed method.

Automated detection of lung nodules in computed tomography images: a review

Lung nodules refer to a range of lung abnormalities the detection of which can facilitate early treatment for lung patients. Lung nodules can be detected by radiologists through examining lung images. Automated detection systems that locate nodules of various sizes within lung images can assist radiologists in their decision making. This paper presents a study of the existing methods on automated lung nodule detection. It introduces a generic structure for lung nodule detection that can be used to represent and describe the existing methods. The structure consists of a number of components including: acquisition, pre-processing, lung segmentation, nodule detection, and false positives reduction. The paper describes the algorithms used to realise each component in different systems. It also provides a comparison of the performance of the existing approaches.

A random forest for lung nodule identification

A method is presented for identification of lung nodules. It includes three stages: image acquisition, background removal, and nodule detection. The first stage improves image quality. The second stage extracts long lobe regions. The third stage detects lung nodules. The method is based on the random forest learner. Training set contains nodule, non-nodule, and false-positive patterns. Test set contains randomly selected images. The developed method is compared against the support vector machine. True-positives of 100% and 85.9%, and false-positives of 1.27 and 1.33 per image were achieved by the developed method and the support vector machine, respectively.

New system to help detect lung cancer

The article tells about the development of an intelligent system that can improve early detection of lung tissue abnormalities.

Hybrid classification of pulmonary nodules

Automated classification of lung nodules is challenging because of the variation in shape and size of lung nodules, as well as their associated differences in their images. Ensemble based learners have demonstrated the potentialof good performance. Random forests are employed for pulmonary nodule classification where each tree in the forest produces a classification decision, and an integrated output is calculated. A classification aided by clustering approach is proposed to improve the lung nodule classification performance. Three experiments are performed using the LIDC lung image database of 32 cases. The classification performance and execution times are presented and discussed.

Cancer detective

The artile describes development of an automated system for detection of lung nodules in CT images.

Pulmonary nodule classification aided by clustering

Lung nodules can be detected through examining CT scans. An automated lung nodule classification system is presented in this paper. The system employs random forests as it base classifier. A unique architecture for classification-aided-by-clustering is presented. Four experiments are conducted to study the performance of the developed system. 5721 CT lung image slices from the LIDC database are employed in the experiments. According to the experimental results, the highest sensitivity of 97.92%, and specificty of 96.28% are achieved by the system. The results demonstrate that the system has improved the performances of its tested counterparts.

Automatic detection of lung nodule from computer tomography images

Lung nodule refers to lung tissue abnormalities that may become cancerous. An automated system that detects nodules of common sizes within lung images is developed. It consists of acquisition, pre-processing, background removal, nodule detection, and false positives reduction. The system can assist expert radiologists in their decision making.

Random forest based lung nodule classification aided by clustering

An automated lung nodule detection system can help spot lung abnormalities in CT lung images. Lung nodule detection can be achieved using template-based, segmentation-based, and classification-based methods. The existing systems that include a classification component in their structures have demonstrated better performances than their counterparts. Ensemble learners combine decisions of multiple classifiers to form an integrated output. To improve the performance of automated lung nodule detection, an ensemble classification aided by clustering (CAC) method is proposed. The method takes advantage of the random forest algorithm and offers a structure for a hybrid random forest based lung nodule classification aided by clustering. Several experiments are carried out involving the proposed method as well as two other existing methods. The parameters of the classifiers are varied to identify the best performing classifiers. The experiments are conducted using lung scans of 32 patients including 5721 images within which nodule locations are marked by expert radiologists. Overall, the best sensitivity of 98.33% and specificity of 97.11% have been recorded for proposed system. Also, a high receiver operating characteristic (ROC) A_z of 0.9786 has been achieved.

Vascular attack by 5,6-dimethylxanthenone-4-acetic acid combined with B7.1 (CD80)-mediated immunotherapy overcomes immune resistance and leads to the eradication of large tumors and multiple tumor foci

The promise of cancer immunotherapy is that it will not only eradicate primary tumors but will generate systemic antitumor immunity capable of destroying distant metastases. A major problem that must first be surmounted relates to the immune resistance of large tumors. Here we reveal that immune resistance can be overcome by combining immunotherapy with a concerted attack on the tumor vasculature. The functionally related antitumor drugs 5,6-dimethylxanthenone-4-acetic acid (DMXAA) and flavone acetic acid (FAA), which cause tumor vasculature collapse and tumor necrosis, were used to attack the tumor vasculature, whereas the T-cell costimulator B7.1 (CD80), which costimulates T-cell proliferation via the CD28 pathway, was used to stimulate antitumor immunity. The injection of cDNA (60–180 µg) encoding B7.1 into large EL-4 tumors (0.8 cm in diameter) established in C57BL/6 mice, followed 24 h later by i.p. administration of either DMXAA (25 mg/kg) or FAA (300 mg/kg), resulted in complete tumor eradication within 2–6 weeks. In contrast, monotherapies were ineffective. Both vascular attack and B7.1 immunotherapy led to up-regulation of heat shock protein 70 on stressed and dying tumor cells, potentially augmenting immunotherapy. Remarkably, large tumors took on the appearance of a wound that rapidly ameliorated, leaving perfectly healed skin. Combined therapy was mediated by CD8+ T cells and natural killer cells, accompanied by heightened and prolonged antitumor cytolytic activity (P < 0.001), and by a marked increase in tumor cell apoptosis. Cured animals completely rejected a challenge of 1 x 107 parental EL-4 tumor cells but not a challenge of 1 x 104 Lewis lung carcinoma cells, demonstrating that antitumor immunity was tumor specific. Adoptive transfer of 2 x 108 splenocytes from treated mice into recipients bearing established (0.8 cm in diameter) tumors resulted in rapid and complete tumor rejection within 3 weeks. Although DMXAA and B7.1 monotherapies are complicated by a narrow range of effective doses, combined therapy was less dosage dependent. Thus, a broad range of amounts of B7.1 cDNA were effective in combination with 25 mg/kg DMXAA. In contrast, DMXAA, which has a very narrow range of high active doses, was effective at a low dose (18 mg/kg) when administered with a large amount (180 µg) of B7.1 cDNA. Importantly, combinational therapy generated heightened antitumor immunity, such that gene transfer of B7.1 into one tumor, followed by systemic DMXAA treatment, led to the complete rejection of multiple untreated tumor nodules established in the opposing flank. These findings have important implications for the future direction and utility of cancer immunotherapies aimed at harnessing patients’ immune responses to their own tumors.

Validity of the assessment of quality of life (AQoL) utility instrument in patients with operable and inoperable lung cancer

There have been few longitudinal studies of quality of life in patients with all stages of lung cancer, particularly those that have included measures of utility. The purpose of this study was to examine the psychometric properties of the Assessment of Quality of Life instrument (AQoL) in patients with lung cancer. The AQoL is a health-related quality of life questionnaire and provides a descriptive system for a multi-attribute utility instrument (MAU), so that scores can be used in cost-utility evaluations. In the present study the reliability (internal consistency) of the AQoL was examined and the concurrent validity was assessed using the Medical Outcomes 36-item Short Form Health Survey (SF-36) as the comparator instrument. The sensitivity to different health states of the AQoL and the responsiveness to change over time was also examined. A prospective, non-experimental cohort study was undertaken. Ninety-two participants with all stages of lung cancer were recruited from a tertiary multi-disciplinary lung cancer clinic. Ninety participants had non-small cell lung cancer (NSCLC) and two had limited stage small cell lung cancer. The AQOL and SF-36 surveys were administered concurrently at baseline. In patients with NSCLC the surveys were then repeated 3 and 6 months later. Correlations between the baseline AQoL summary scales and SF-36 summary scales support the divergent and convergent validity of the AQoL. Reliability was also found to be sufficient (Cronbach's Alpha = 0.76). In addition, in patients with inoperable NSCLC, baseline AQoL scores were found to be predictive of survival at 6 months in Cox proportional hazards multivariate analysis. However, the physical components summary score of the SF-36 was more sensitive to differences in health states between patients with different stages of NSCLC at 6 months of follow-up and more responsive to change over time in both operable and inoperable patients with NSCLC than the AQoL. The findings support the construct validity and reliability of the AQoL in this population. However, there remains some uncertainty about whether the AQoL has sufficient sensitivity to different health states in this population. Further studies using other MAU instruments may determine whether alternative instruments are more sensitive to different health states in individuals with lung cancer.

Discordance between lung function of Chinese university students of 20-year-old established norms

Objective: We examined the validity of the 20-year-old established Asian norms for pulmonary function in a contemporary cohort of Hong Kong Chinese university students. Design and participants: Pulmonary function testing was conducted in university students (n = 805). Setting: A university campus in Hong Kong. Measurements and results: Parameters recorded included gender, age, height, weight, standard lung function variables (ie, FEV1, FVC, and peak expiratory flow rate [PEFR]), and exhaled carbon monoxide (CO) level. Subjects completed a questionnaire on pulmonary health, smoking history, and their dietary and exercise habits within 3 months of the study. Data were compared with the established norms for lung function for Chinese persons from Hong Kong. On average, subjects were taller than those reported in the original cohort, on whom the established norms are based; however, FEV1, FVC, and PEFR were lower. As predicted, the exhaled CO level was higher in smokers. Those who exercised regularly had a higher FEV1 and FVC, and reported fewer respiratory complaints. Conclusions: Our findings support the idea that lung function norms not only differ across ethnic groups, but that they may be susceptible to change over a single generation within an ethnic group living in the same geographic region. Assuming the equivalence of our testing methods and those on which established norms are based, our findings shed further insight into the dynamic nature of lung function, and have implications regarding the definition of normal pulmonary function and its variance over the short term.

Cost-effectiveness analysis of screening for lung cancer with low dose spiral CT (computed tomography) in the Australian setting

Introduction:
Low dose spiral computed tomography (CT) is a sensitive screening tool for lung cancer that is currently being evaluated in both non-randomised studies and randomised controlled trials.
Methods:
We conducted a quantitative decision analysis using a Markov model to determine whether, in the Australian setting, offering spiral CT screening for lung cancer to high risk individuals would be cost-effective compared with current practice. This exploratory analysis was undertaken predominantly from the perspective of the government as third-party funder. In the base-case analysis, the costs and health outcomes (life-years saved and quality-adjusted life years) were calculated in a hypothetical cohort of 10,000 male current smokers for two alternatives: (1) screen for lung cancer with annual CT for 5 years starting at age 60 year and treat those diagnosed with cancer or (2) no screening and treat only those who present with symptomatic cancer.
Results:
For male smokers aged 60–64 years, with an annual incidence of lung cancer of 552 per 100,000, the incremental cost-effectiveness ratio was $57,325 per life-year saved and $105,090 per QALY saved. For females aged 60–64 years with the same annual incidence of lung cancer, the cost-effectiveness ratio was $51,001 per life-year saved and $88,583 per QALY saved. The model was used to examine the relationship between efficacy in terms of the expected reduction in lung cancer mortality at 7 years and cost-effectiveness. In the base-case analysis lung cancer mortality was reduced by 27% and all cause mortality by 2.1%. Changes in the estimated proportion of stage I cancers detected by screening had the greatest impact on the efficacy of the intervention and the cost-effectiveness. The results were also sensitive to assumptions about the test performance characteristics of CT scanning, the proportion of lung cancer cases overdiagnosed by screening, intervention rates for benign disease, the discount rate, the cost of CT, the quality of life in individuals with early stage screen-detected cancer and disutility associated with false positive diagnoses. Given current knowledge and practice, even under favourable assumptions, reductions in lung cancer mortality of less than 20% are unlikely to be cost-effective, using a value of $50,000 per life-year saved as the threshold to define a “cost-effective” intervention.
Conclusion:
The most feasible scenario under which CT screening for lung cancer could be cost-effective would be if very high-risk individuals are targeted and screening is either highly effective or CT screening costs fall substantially.

From lung images to lung models : a review

Automated 3D lung modeling involves analyzing 2D lung images and reconstructing a realistic 3D model of the lung. This paper presents a review of the existing works on automatic formation of 3D lung models from 2D lung images. A common framework for 3D lung modeling is proposed. It consists of eight components: image acquisition, image pre-processing, image segmentation, boundary creation, image recognition, image registration, 3D surface reconstruction, and 3D rendering and visualization. The algorithms used by the existing systems to implement these components are also reviewed.