Virtual Library
Start Your Search
P. Zimmerman
Moderator of
-
+
MO11 - Screening and Epidemiology (ID 131)
- Event: WCLC 2013
- Type: Mini Oral Abstract Session
- Track: Imaging, Staging & Screening
- Presentations: 13
- Moderators:P. Zimmerman, J. Bowden
- Coordinates: 10/28/2013, 16:15 - 17:45, Bayside 201 - 203, Level 2
-
+
MO11.01 - Positron Emission Tomography in lung cancer screening: Six-year results (ID 1816)
16:15 - 16:20 | Author(s): G. Veronesi, L. Travaini, C. Rampinelli, R. Bertolotti, L. Spaggiari, M. Bellomi, A. Pardolesi, G. Paganelli
- Abstract
- Presentation
Background
Lung cancer computed tomography (CT) screening is a controversial topic. Results from the National Lung Screening Trial showed that low dose CT (LD-CT) can reduce lung cancer mortality by at least 20%. However a critical issue is the high rate of indeterminate lung nodules and false positive cases. The aim of the study was to evaluate the diagnostic accuracy of CT/PET for lung cancer diagnosis in the context of lung cancer screening.Methods
Between 2004 and 2005, 5203 asymptomatic high-risk individuals (≥20 pack-years, age ≥50 years) were enrolled to undergo annual LD-CT for 10 years. Nodules ≤5 mm were scheduled for repeat LD-CT a year later. Nodules >5.0≤8.0mm received LD-CT 3-6 months later. Nodules >8.0mm or growing nodules underwent CT-PET. Results from all PET scans performed during the screening workup of COSMOS participants were reviewed. Those performed for suspected collateral disease (pleura, mediastinum, other cancers) were excluded. Outcome was based on the pathological findings for patients who underwent surgery or on findings at follow-up. Outcome was considered negative for those with negative CT findings for at least two years. PET results were visually evaluated by expert radiologists.Results
383 nodules in 351 patients were studied by CT/PET. In 5 occasions, PET evaluation was dubious. 197 nodules turned out to be malignant. Overall sensitivity, specificity and accuracy of CT/PET to distinguish between benign and malignant nodules was 64%, 89% and 76% while it was 82%, 92%, 88% when considering only PET scans performed during baseline screening work-up. Diagnostic performance was high (sensitivity 87%, specificity 73%) for nodules larger than 15 mm in their maximum diameter, reaching 98% sensitivity for solid nodules larger than 15 mm. The diagnostic performance of the test was significantly lower for nodules investigated at annual repeat CT compared to baseline CT (p<0.0001, sensitivity ranging from 30 to 71%) and for non-solid compared to solid nodules (p=0.0001; sensitivity 22% for non-solid versus 79% for solid nodules).Strata PETs VN FN FP VP Sensitivity Specificity Accuracy p-value accuraccy Overall 378 164 70 21 123 64% 89% 76% Diam <10mm 145 75 40 4 26 39% 95% 70% Diam 10-15mm 104 51 19 4 30 61% 93% 78% Diam >15mm 117 30 10 11 66 87% 73% 82% 0.06 Solid 263 123 26 16 98 79% 88% 84% Non solid 48 15 25 1 7 22% 94% 46% Partially solid 65 25 19 3 18 49% 89% 66% Non/Part solid 114 41 44 4 25 36% 91% 58% <0.0001 Age <60 147 64 31 14 38 55% 82% 69% Age >60 231 100 39 7 85 69% 93% 80% 0.02 Inferior lobe 139 69 29 8 33 53% 90% 73% Superior lobe 208 85 33 9 81 71% 90% 80% Other location 30 10 8 3 9 53% 77% 63% 0.09 Left lobe 148 64 27 8 49 64% 89% 76% Right lobe 206 90 35 10 71 67% 90% 78% 0.70 Baseline 148 81 11 7 49 82% 92% 88% 2[nd] year 57 17 11 3 26 70% 85% 75% 3[rd] year 62 19 28 3 12 30% 86% 50% 4[th] year 44 10 10 4 20 67% 71% 68% 5[th] year 47 31 6 4 6 50% 89% 79% 6[th] year 20 6 4 0 10 71% 100% 80% <0.0001 Male 264 113 44 12 95 68% 90% 79% Female 114 51 26 9 28 52% 85% 69% 0.05 Non/part solid <15mm 72 27 34 2 9 21% 93% 50% Non/part solid >15mm 37 10 9 2 16 64% 83% 70% Solid <15mm 177 99 25 6 47 65% 94% 82% Solid >15mm 80 20 1 9 50 98% 69% 88% <0.0001 Conclusion
CT/PET is a highly sensitive test for the differential diagnosis of screening-detected cancer, in particular at baseline CT, for solid nodules and those with a diameter larger than 15mm. Sensitivity of CT/PET for sub-solid nodules is very low suggesting that other diagnostic tests, such as volume doubling time, should be used.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.02 - Validation of electronic nose exhaled breath VOC profile in discriminating between subjects with early stage lung cancer and healthy ever smokers. (ID 2574)
16:20 - 16:25 | Author(s): A.G. Dent, R. Bowman, P. Zimmerman, I.A. Yang, K.M. Fong
- Abstract
- Presentation
Background
Early diagnosis of lung cancer is associated with a better survival. The measurement of volatile organic compounds (VOCs) in exhaled breath using an electronic nose may prove to be a novel, effective and simple technique for screening and diagnosing lung cancer. The aim was to test the validity of the VOC profile in discriminating subjects with early stage lung cancer (I and II) (ESLC) from healthy ever smokers (HS).Methods
243 subjects: 54 ESLC and 189 HS provided a breath sample after tidally breathing through an inspiratory port filter for 5 minutes. It was analysed using a 32 sensor Cyranose 320 (Smiths Detection). Subjects were divided into training (n=159) and independent test set (n=84) groups. Canonical discrimination analyses were performed to determine significance of difference between subject groups and calculate cross validated accuracy (CVV) of the groups using leave one out classification method. Area under the curve (AUC) of Receiver Operating Characteristic Curves were also determined (SPSS V17.0).Results
Validation of the training VOC profile model using an independent test group showed 79% accuracy (p=0.001) in distinguishing ESLC (n=20) from HS (n=64). (AUC 0.933). There was no significant difference in age, lung function and smoking history between the training and test groups.Conclusion
Exhaled breath VOC profile model to discriminate between ESLC subjects and HS was validated in an independent group with a high accuracy. A clinically high sensitivity of the VOC profile model to discriminate between ESLC and HS can be achieved by selecting an appropriate cut point. The cyranose has potential to be a clinically useful diagnostic and screening tool for early stage lung cancer.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.03 - The natural history and screen-detectability of lung cancer: estimates from NLST and the PLCO (ID 2383)
16:25 - 16:30 | Author(s): K. Ten Haaf, H.J. De Koning, J. Van Rosmalen
- Abstract
- Presentation
Background
Implementing effective lung cancer screening programs requires extensive knowledge on the natural history of lung cancer and the sensitivity of the proposed screening modality. Data from the National Lung Cancer Screening Trial (NLST), the Prostate, Lung, Colon and Ovarian Cancer Study (PLCO) and Surveillance Epidemiology and End Result (SEER) data from 2004-2008 are used to investigate the screening sensitivity (by stage and histological type) of Computed Tomography (CT) and chest radiography (CXR) and the mean preclinical sojourn time (MPST) of lung cancer (by gender, stage and histological type).Methods
The MISCAN-Lung model was used to reproduce the lung cancer incidence by method of detection (clinical or screen-detected), gender, histology and stage in both trials, by calibrating screening sensitivity and natural history parameters.Results
Major differences in sensitivity between CT and CXR are estimated for the less advanced stages, for example the sensitivity for stage IA adenocarcinoma is estimated to be 56.63% for CT compared to 16.91% for CXR. The model suggests that the MPST varies by histological type and gender. The largest difference between genders was estimated for adenocarcinoma for which the MPST in pre-clinical stages IA to IV was estimated to be 4.48 years for men compared to 6.01 years for women.Conclusion
This study provides detailed insights in the natural history of lung cancer and the differences in the effectiveness of screening between the NLST and PLCO trials. This knowledge may help to determine effective lung cancer screening programs.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.04 - Lung Cancer risk mediated through susceptibility to COPD: preliminary results from a sub-analysis of the NLST (ID 1540)
16:30 - 16:35 | Author(s): R.J. Hopkins, R.P. Young, F. Duan, X. Deng, C. Chiles, G.D. Gamble, D. Aberle
- Abstract
- Presentation
Background
Introduction: Based on a 20% reduction in lung cancer deaths in the CT screening arm of the National Lung Screening Trial (NLST), yearly CT screening for lung cancer is now widely recommended. Recently, a multivariate risk model for lung cancer (LC), be used to identify smokers at greatest risk, has been proposed to better select smokers for CT ([#]Tammemagi et al. JNCI 2011;103:1058).This risk model includes age smoking history, history of COPD, BMI, recent CXR and educational level. This model has been validated in the NLST where recent CXR and educational level was omitted due to the unavailability of this data. In this sub-study of the NLST, we examine the role of the known clinical risk factors for lung cancer and whether they mediate risk for lung cancer via risk for COPD. To test this we compared risk factors between LC cases and controls after stratifying cancer-free screening participants (controls) by spirometry-defined COPD.Methods
Using a sub-group of the NLST (drawn for validation of a gene-based risk stratification tool), we compared known risk factors for LC (recently validated in the PLCO[#] study) in 345 screen-detected lung cancer cases and 1482 randomly selected cancer free screening controls stratified by COPD (pre-bronchodilator GOLD 1-4). These variables included age, pack years, family history of LC (FHx), self-reported COPD and BMI.Results
When the LC risk variables were compared between LC cases (N=345) and the cancer-free screening controls, stratified from baseline data into controls with COPD (N=489) and without COPD (n=993) (see Table1), we found the following; age and pack years but not FHx were significantly higher in those with COPD and LC compared to cancer-free controls with no COPD, self reported COPD was 2 fold higher in COPD controls and LC cases compared to controls without COPD, and BMI was significantly lower in the LC cases and those with COPD compared to cancer-free controls with no COPD. Figure 1Conclusion
These data from the NLST suggest the risk of LC is strongly mediated by variables underlying risk of COPD (age, pack years and BMI). [#]Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO)Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.05 - DISCUSSANT (ID 3993)
16:35 - 16:45 | Author(s): M. Oudkerk
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.06 - Minority Status as a Strong Predictor of Health Disparities: A Global Perspective (ID 3296)
16:45 - 16:50 | Author(s): D.C. Li, S.E. Cox, A. Hulbert, J. Harris, N. Wang, A. Memon, A.J. Alberg, M.V. Brock
- Abstract
- Presentation
Background
The existence of healthcare disparities by race and ethnicity is well documented, often attributed to a lack of economic and educational parity among groups, and manifested by unequal healthcare access and delivery. Inexplicably however, in many countries the wealthiest and most educated populations have the worst health outcomes. We hypothesize that minority status within a country rather than race, ethnicity, socioeconomics or educational level is more closely associated with poor health outcomes globally and functions independently of other variables.Methods
Minority and majority populations in 34 countries/territories were evaluated for smoking prevalence and age-adjusted lung cancer incidence rates. A global, systematic review of over 1000 sources of epidemiological data was performed using rigorous screening criteria including only national data in original form (national cancer registry, census, or health surveys), from an official government agency, or from peer-reviewed publications. Relative risks (RR) of smoking and lung cancer were computed for all minority groups with the majority population as the referent null. Relative wealth based on median per capita income and relative schooling based on educational attainment were also calculated. Minority groups were only included if so defined by both the Minority Rights Group International and the CIA World Fact Book.Results
Data were collected from approximately 60% of the global population including all six WHO world regions and every populated continent. The RR of smoking for at least one minority group was greater than that of the referent majority in every country or territory analyzed. The RR of lung cancer for at least one minority group was greater than that of the referent majority in all but three countries/territories. These results were remarkably consistent and durable with RR ≥ 1.0 for smoking prevalence and lung cancer incidence in nearly all countries despite minority status being defined differently in many nations whether by race, ethnicity, religion, language, indigenous affiliation, or immigrant status. These results were further corroborated by age-specific lung cancer incidence for selected countries/territories. Racial and socioeconomic status differentials were insufficient explanations for these observations. In the U.S., for example, blacks are generally less wealthy and educated than whites and have higher smoking prevalence and lung cancer incidence rates. However, in neighboring Bermuda and distant South Africa where blacks are the majority, whites have both higher smoking prevalence and lung cancer incidence rates despite being far wealthier and educated. This relationship of increased smoking and lung cancer rates in wealthier, more educated minorities is replicated in nine of the 34 countries in this study.Conclusion
Our results show an empirical relationship between minority status and both increased smoking prevalence and lung cancer incidence rates in minority populations globally. This suggests that minority status may be a potent, behavioral driver leading to elevated health risks in minority populations around the world. Moreover, minority status seems to be independent of traditional socioeconomic variables, and alone may be a powerful predictor of disparate health outcomes in many diverse nations, distinctive societies, and unique cultures on a global scale.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.07 - Exploring Disparities in Health Care Outcomes in Minority Populations with Small (SCLC) and Non-small cell lung cancers (NSCLC). (ID 1237)
16:50 - 16:55 | Author(s): L.E. Raez, C. Sareli, R. Mudad, F. Tarrazzi, E. Velis, S. Sundararaman, M. Block
- Abstract
- Presentation
Background
There are population-specific differences in the presence of lung cancer, these are related with: health outcomes, quality of health care and access to health care services that exist across racial and ethnic groups. Disparities represent a lack of efficiency within the health care system and therefore account for unnecessary costs. Few publications are available about disparities in African-American (AA) populations with lung cancer and very few about Hispanic (H) populations with this disease.Methods
We reviewed registry data on 2,255 pts with non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) treated during last 10 years (2002-2011) at Memorial Health Care System. The main objective of the study was to evaluate differences in lung cancer survival according to ethnicity. Chi-square was used to compare distribution of tumor stage. Survival curves were compared using log-rank test for each of the tumor stages. Adjusted hazard ratios (AHR) and 95% confidence intervals (95% CI) were reported based on the results of a multivariate Cox regression model for overall survival (OS) with adjustment for gender, age at diagnosis, race, stage and health insuranceResults
A total of 1940 pts (86%) had the diagnosis of NSCLC, the rest were SCLC. There were 1170 (52%) females. Non-Hispanic whites (NHW) were 1,791 pts (79%), Hispanics (H) 266 (12%) and African-American (AA) 149 (7%). Fifty eight percent of patients had stage III/IV at diagnosis. 2054 of the pts were insured (91%). There was a significant difference in age at diagnosis among H (66.5 y), AA (64.4y) and WNH (69.5y). The probability of being diagnosed at a late stage (IIIB/IV) was two times higher among AA compared to NHW (OR= 1.77, p<0.05) or H (OR) = 1.67, p <0.05). There was no survival difference between NSCLC and SCLC (19m vs. 16m). Females with NSCLC lived significantly longer than males (Adjusted Hazard ratio (AHR= 1.14 p<0.01). The same was true for SCLC (AHR = 1.43 p < 0.01). Significant predictors for worse survival for patients with NSCLC were: older age at diagnosis (AHR = 1.01, p<0.001), male gender (AHR=1.12, p<0.05) and late stage at diagnosis (AHR=2.27, p<0.001). Insurance and ethnicity were not significant predictors of survival.Conclusion
There are significant disparities in presentation and outcomes among minority patients (AA and H) with lung cancer. We will further evaluate if other social or genetic factors can explain these disparities.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.08 - The importance of a Regimen of Screening to Maximize Early Diagnosis and Treatment of Lung Cancer (ID 879)
16:55 - 17:00 | Author(s): C.I. Henschke, R. Yip, D.F. Yankelevitz
- Abstract
- Presentation
Background
Writing on behalf of the I-ELCAP Investigators. The goal of CT screening is to maximize lung cancer cure rates by early diagnosis and treatment of lung cancer. To achieve this, a regimen of screening is important, particularly for small nodules. To better understand the importance of a regimen, we compared two large databases of screen-diagnosed lung cancers, the International Early Lung Cancer Action Program (I-ELCAP) and the National Lung Screening Trial (NLST) CT arm as the former had a specified diagnostic workup algorithm while the latter did not mandate any specific approach.Methods
We compared all lung cancers including small-cell and carcinoids, that were diagnosed under screening, that is, either screen-diagnosed because of a positive result of the CT screening or symptom-prompted after a negative CT screening. We compared the stage and size distribution of the screen-diagnosed cancers in the International Early Lung Cancer Action Program (I-ELCAP) from 1993 to 2011 and those in the NLST-CT arm from 2002 to 2006. In I-ELCAP, the screenings were performed according to a common protocol in which the diagnostic workup for a participant was defined by a specified protocol which has been continually updated whereas in the NLST, “… no specific diagnostic evaluation approach was mandated.”Results
In I-ELCAP, a total of 799 patients were diagnosed under screening of which 11 (1.4%) were interim-diagnoses; 8 prior to the first annual repeat screening and 3 between annual rounds, leaving 788 screen-diagnosed patients. In the NLST CT arm, 1060 patients were diagnosed with lung cancer, but only 692 were diagnosed under screening. There were 18 interim-diagnosed cases before the first annual repeat screening and 26 between the 2 annual repeat screenings, a total of 44 (6.4%) interim-diagnosed cases for this cohort. This left 649 screen-diagnosed patients. The frequency of clinical Stage I was 82% (95% CI: 79%-84%) vs. 69% (95% CI: 65%-72%) in I-ELCAP and NLST, respectively. Average tumor size (95% confidence interval) was 17.3 mm (16.6-18.1) vs. 23.1 mm (21.7 -24.4), respectively. Surgical resection was performed in 86% (676/788) and 77% (497/649) of the screen-diagnosed patients, respectively. The frequency of pathologic Stage I (clinical, if not resected) was 73% (95% CI: 70%-76%) vs. 63% (95% CI: 59%-67%).Conclusion
Stage I disease, both clinical and pathologic, was significantly higher in I-ELCAP than NLST. The tumor size was significantly lower in I-ELCAP than NLST, all strongly suggestive of the importance of a specified regimen of screening. This is further substantiated by the reported 71% for pathologic Stage I (clinical, if not resected) by the NELSON study which is close to that reported by I-ELCAP as the NELSON also followed a well-defined regimen of screening.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.09 - CT Screening for Lung Cancer: Definition of Positive Test Result in the National Lung Screening Trial CT cohort compared with I-ELCAP (ID 872)
17:00 - 17:05 | Author(s): R. Yip, D.F. Yankelevitz, C.I. Henschke
- Abstract
- Presentation
Background
Low-dose CT screening for lung cancer can reduce mortality among high-risk people but to reduce unnecessary evaluations with attendant risks, alternative thresholds for defining positive result and cancer diagnoses needs to be further understood. The purpose of the study is to assess the frequency of positive results and potential delays in diagnosis in the baseline round of screening using more restrictive thresholds.Methods
Among the participants who were randomly assigned to the CT arm of the National Lung Screening Trial (NLST) cohort, we identified the frequency of solid and part-solid pulmonary nodules and the rate of lung cancer diagnoses using a 5.0, 6.0, 7.0. 8.0 and 9.0 mm threshold for the largest noncalcified nodule identified in the baseline CT scan. we compared these results with those previously published for the I-ELCAP cohort.Results
The frequency of positive results in the baseline round, using the definition of positive result (any parenchymal, solid or part-solid, noncalcified nodule > 5.0 mm), was 15.9% (4,104/25,814). Using alternative threshold values of 6.0, 7.0, 8.0 and 9.0 mm, the frequencies (95% CI) of positive results were 10.5% (10.2, 10.9), 7.2% (6.9, 7.5), 5.3% (5.0, 5.6) , and 4.2% (3.9, 4.4), respectively. Use of these alternative definitions would have reduced the workup by 33.8%, 54.7%, 66.6%, and 73.8%, respectively. Concomitantly, proportion of lung cancer diagnoses made within first 12 months would be delayed for 0.9%, 2.6%, 6.1%, and 10.0% of the patients, respectively. These results are similar to those found in I-ELCAP.Conclusion
These results are similar to those found in I-ELCAP and suggest that even in the higher-risk participants enrolled in the NLST, higher threshold values can be used. This reduction in the positive result rate compared to the 28% positive result rate reported in the NLST, which used a 4mm threshold, is considerable.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.10 - DISCUSSANT (ID 3994)
17:05 - 17:15 | Author(s): R. Bowman
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.11 - IL-11 and CCL-1: Novel protein diagnostic biomarkers of lung adenocarcinoma in bronchoalveolar lavage (BAL) (ID 2876)
17:15 - 17:20 | Author(s): M. Pastor, A. Nogal, S. Molina-Pinelo, A. Quintanal, B. Romero, M.J. De Miguel, R. Melendez, J.L. Lopez-Campos, J. Corral, R. Garcia Carbonero, A. Carnero, L. Paz-Ares
- Abstract
- Presentation
Background
Lung cancer (LC) and chronic obstructive pulmonary disease (COPD) are smoking related diseases and the presence of COPD increases itself the risk of developing lung cancer, probably due to underlying inflammation. LC is typically detected at late stages of the disease and carries a dismal prognosis. There is an unmet need for useful early detection methods of lung cancer in high risk subjects, such as smokers.Methods
The expression of inflammatory proteins was studied in bronchoalveolar lavage samples (BAL) by antibody arrays in a prospective discovery cohort of 60 patients with the following inclusion criteria: age > 40 years, diagnostic broncoscopy due to hemoptysis or pulmonary nodule, smokers or ex-smokers pack of more than 30 pack-years, divided into four groups (control, LC, COPD, LC & COPD). Relevant biomarkers were validated by western blot. Additional validation was carried out by ELISA in two independent controlled cohorts of 139 (control, LC, COPD, LC & COPD) and 160 patients (control, all LC histological subtypes).Results
CCL-1 and IL-11 were selectively expressed in samples of adenocarcinoma patients, with or without COPD (p<0•001) in the discovery cohort. These proteins exhibited a remarkable diagnostic performance for lung adenocarcinoma in an independent cohort of 139 patients. ROC curves showed that the optimum diagnostic cutoff value for IL-11 was 42 pg/mL (area under curve [AUC] 0•93 [95% CI 0•896-0•975], sensitivity 90%, specificity 86%), and for CCL-1 was 39•5 pg/ml (0•83 [95% CI 0•749-0•902], sensitivity 83%, specificity 74%). Further validation of the ELISA biomarkers at the mentioned cutoffs was performed in an additional cohort of 160 patients (20 controls, 66 LC, 74 LC & COPD). There was a significant correlation between BAL levels of IL-11 and CCL-1 (r2= 0•76, p<0•001), and the use of both biomarkers increased the diagnostic accuracy to 96,1% in the two validation cohorts. Appropriate diagnostic performance was observed for all subgroups regardless of stage at diagnosis, involvement of bronchial tract, pack-years smoked, and number of cells in BAL.Diagnostic performance of IL-11 and CCL-1 in the first validation cohort (N=139) Adenocarcinoma vs all patients AUC (95%IC) Sensivity (95%IC) Specificity (95%IC) PPV (95%IC) NPV (95%IC) Positive LR Negative LR IL-11 0.93 (0.896-0.975) 90.2% (79%-95.7%) 88.7% (80.6%-93.5%) 80.7% (68.7%-88.9%) 94.5% (87.8%-97.6%) 7.95 (4.53-13.98) 0.11 (0.05-0.26) CCL-1 0.83 (0.749-0.902) 80% (66.4%-87.7%) 74.1% (63.9%-82.2%) 72.1% (59.2%-73.4%) 86.3% (76.6%-92.4%) 3.02 (2.05-4.47) 0.29 (0.17-0.52) IL11 and CCL-1 71.2% (57.7%-81.7%) 94.4% (88.4%-97.4%) 86% (72.7%-93.4%) 87.2% (79.9%-92.1%) 12.8 (5.77-28.41) 0.31 (0.20-0.47) IL-11 and/or CCL-1 94.3% (84.6%-98.1%) 74.1% (65.1%-81.4%) 64.1% (53%-73.9%) 96.4% (89.9%-98.8%) 3.64 (2.63-5.04) 0.08 (0.03-0.23) Diagnostic performance of IL-11 and CCL-1 in the second validation cohort (N=160) Adenocarcinoma vs all patients AUC (95%IC) Sensivity (95%IC) Specificity (95%IC) PPV (95%IC) NPV (95%IC) Positive LR Negative LR IL-11 0.95 (0.92-0.98) 90.6% (79.7%-95.9%9 83% (86.8%-87.7%) 60.8% (49.7%-70.8%) 96.8% (92.7%-98.6%) 5.32 (3.81-7.41) 0.11 (0.05-0.26) CCL-1 0.91 (0.87-0.96) 91.7% (80.4%-96.7%) 77.5% (71.0%-82.9%) 51.2% (40.8%-61.4%) 97.3% (93.3%-99%) 4.08 (3.09-5.04) 0.11 (0.04-0.28) IL11 and CCL-1 71.2% (57.7%-81.7%) 96.3% (92.5%-98.2%) 84.1% (70.6%-92.1%) 92.3% (87.7%-95.3%) 19.1 (9-41.13) 0.3 (0.19-0.46) IL-11 and/or CCL-1 92.3% 82.6%-98.1%) 84% (78%-88.5%) 62.5% (51.5%-72.3%) 98.1% (94.6%-99.4%) 5.88 (4.21-8.22) 0.07 (0.02-0.20) Conclusion
IL-11 and CCL-1 are highly specific biomarkers with great accuracy for the diagnosis of lung adenocarcinoma in BAL specimens. Further study of these proteins as markers for early diagnosis and screening in plasma and other biological materials is warranted.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.13 - Low selenium serum level is a good preselection factor for patients invited for low dose chest CT lung cancer screening (ID 823)
17:20 - 17:25 | Author(s): T. Grodzki, J. Wójcik, A. Jakubowska, N. Wójcik, M. Muszynska, B. Kubisa, K. Jaworska-Bieniek, J. Pieróg, S. Gupta, J. Alchimowicz, T. Gromowski, H. Janowski, G. Sukiennicki, M. Bielewicz, J. Lubinski
- Abstract
- Presentation
Background
Selenium is known as a chemoprotective anti-cancer agent. The impact of low selenium serum level on the frequency of early lung cancer (LC) detection rate in patients participating in a low dose chest CT (LDCT) screening program is assessed.Methods
A program of early LC detection was conducted in a single city of 400000 inhabitants from 2008 till 2011 (Protocol I). Enrollment criteria included both sexes aged 55-65 years with a history of 20 pack-years of tobacco smoking. All detected lesions were followed up in accordance with IELCAP protocols. A new pilot study (Protocol II) launched in 2012 was based on preselection of the participants by measure of their selenium serum level. Only individuals with a low selenium level (< 75 microgramms/ml) were invited for CT scans. Other enrollment criteria were the same as in Protocol I. All cases requiring surgery were referred to a single local thoracic surgery department. The following data were analyzed: number of all detected lesions as well as LC detected in both protocols.Results
Protocol I: 15020 patients were screened by LDCT. 6240 pulmonary lesions were detected with the majority (59%) smaller than 5mm. 182 patients (2.9% of all detected lesions) were referred for thoracic surgery. 119 primary LC were diagnosed and treated. Protocol II: 2440 patients have had selenium serum level measured. 720 of them were screened by LDCT. 210 lesions were detected with 49% smaller than 5mm. 14 patients (6.8%) were referred for thoracic surgery. 9 primary LC were diagnosed and treated surgically. Protocol II was more specific because number of all detected lesions was significantly smaller than in Protocol I (p < 0.0001, OR: 0.55; CI: 0.46 – 0.64) but the LC detection rate in Protocol II was more than twice as high as in Protocol I (p = 0.0226, OR: 2.38; CI: 1.19 – 4.76).Conclusion
The detection rate of LC in the program of early detection based on LDCT is higher if the protocol is supported by preselection of high risk tobacco - smoking patients based on a low level of selenium in serum.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
MO11.14 - DISCUSSANT (ID 3995)
17:25 - 17:35 | Author(s): S. Lam
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29 - Cancer Control & Epidemiology IV (ID 132)
- Event: WCLC 2013
- Type: Oral Abstract Session
- Track: Prevention & Epidemiology
- Presentations: 8
- Moderators:P. Zimmerman, N. Kurimoto
- Coordinates: 10/30/2013, 10:30 - 12:00, Bayside 103, Level 1
-
+
- Abstract
- Presentation
Background
The risk of developing lung cancer increases with age and is highly associated with exposure to tobacco smoking. This study aimed to understand public awareness about the risk of developing lung cancer in New South Wales (NSW), Australia and to investigate attitudes and beliefs about lung cancer which could potentially impact health seeking behaviours regarding symptoms of lung cancer. The risk of developing lung cancer increases with age and is highly associated with exposure to tobacco smoking. This study aimed to understand public awareness about the risk of developing lung cancer in New South Wales (NSW), Australia and to investigate attitudes and beliefs about lung cancer which could potentially impact health seeking behaviours regarding symptoms of lung cancer.Methods
A mixed methods triangulation approach was used, with a sequential-parallel design which incorporated 16 qualitative focus groups (n=125) to explore themes. Participants were aged 45+ from metropolitan and regional centres and grouped according to smoking status (current, former and never smokers). Data were analysed by content and thematically. A cross sectional quantitative telephone survey (n=1,000) followed to determine overall awareness and generalisability of findings amongst adults >45 years across NSW.Results
The qualitative research highlighted that symptoms of haemoptysis, dyspnoea and an unusual or persistent cough were well recognised symptoms of lung cancer however participants were more likely to assume these symptoms were related to other health problems. Haemoptysis was the only symptom which created a sense of urgency to seek immediate medical attention. A ‘wait and see’ attitude towards any concerning symptom was prevalent across groups, only severe/ long term persistent symptoms would induce action. Smokers and former heavy smokers were more likely to say they would delay seeing their doctor because of perceived stigma associated with smoking. Older participants were more likely to rely on previous experiences of symptoms to govern their health seeking behaviour. Perceived susceptibility and understanding of causes of lung cancer differed by smoking status; smokers more likely to down-play the risk of smoking or attempt to offset their risk through lifestyle choices. Former smokers were more likely to perceive their risk comparative with a never smoker. The quantitative research findings suggested that unprompted awareness of lung cancer related symptoms was high with 88.5% able to correctly identify ≥1 symptoms however symptoms were more typical of late stage cancer. Age(<65 years), sex(female) and high socio-economic status were associated with higher recognition of symptoms (p<0.05). The majority of the survey respondents identified smoking as a cause of lung cancer (90.6%, 95%Confidence Interval (CI) 88.4-92.8) however fewer recognised the risk from second-hand smoke (25.6%,95%CI22.3-28.9). Ever-smokers were less likely to recognise the risk of smoking (odds ratio 0.7 95%CI0.5-0.9).Conclusion
These findings provide evidence that while awareness of lung cancer symptoms and causative factors is reasonably understood in the community, perceived susceptibility is low (particularly among current and former smokers). A lack of urgency in seeking medical attention for symptoms considered not severe, together with other smoking-related barriers, may lead to further delays in diagnosis and missed opportunity for surgical treatment.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29.02 - Impact of a national Lung Cancer Awareness Campaign in Merseyside (ID 2768)
10:40 - 10:50 | Author(s): J. Maguire, V. Kelly, A. McIver, C. Smyth, M. Ledson, M. Walshaw, J. Littler
- Abstract
- Presentation
Background
In May 2012 the United Kingdom Department of Health initiated a centrally funded National Lung Cancer Awareness Campaign. This campaign was influenced by encouraging results from local pilot studies, and featured news items and public health awareness adverts in national and local media, including nationwide television and radio exposure, accompanied by celebrity involvement and endorsements. The cost of this initiative has not been disclosed. Liverpool has the highest lung cancer incidence and mortality in England, and our Lung Cancer Unit, providing a service for patients in South Liverpool, is the largest in the Cancer Network and the largest in the countryMethods
To assess the impact of the national campaign we have reviewed referrals to our Rapid Access Lung Cancer Service from General Practice via the two week suspected cancer pathway, during the months of April to December 2012; we have also reviewed referrals during the same months in 2011 and 2010. For these time periods we have compared the total number of referrals, the number and proportion of referrals judged as unsuitable for immediate investigation with CT scan and bronchoscopy, the number and stage distribution of cancers diagnosed and the proportion of patients undergoing potentially curative surgery.Results
In the period May to December 2012 a total of 323 patients were referred to our Rapid Access Service; of these, 140 (43%) did not have any features suggestive of a possible lung cancer - these patients were reviewed in a routine chest clinic. A total of 103 patients (32% of referrals in this time period) were diagnosed as having lung cancer. Of these 34 (33%) had stage I or II disease on CT staging and 30 patients (29%) underwent potentially curative surgery. The corresponding figures for 2011 are: total referrals 283, 93 (33%) sent to routine chest clinic, 88 (31%) diagnosed lung cancers, 17 (20%) stage I and II and 16 (18%) potentially curative operations. In May to December 2010 we received 274 referrals. 75 (27%) were referred on to routine chest review. 106 patients (39%) were diagnosed as having lung cancer. 23 (21%) were stage I or II and 24 (23%) underwent potentially curative surgery.Conclusion
In the 8 months following the initiation of a National Lung Cancer Awareness Campaign, the largest lung cancer unit in England recorded a 14% increase in patients referred from General Practice with suspected lung cancer. There was a 17% increase in diagnosed lung cancers, and the percentage of patients undergoing potentially curative surgery for stage I and II disease increased from 18% to 29%. Compared to the same time period in 2010, an additional 6 patients underwent surgery for stage I and II disease in 2012, an increase of 6% in the surgical resection rate. We have observed a beneficial effect from the National Lung Cancer Awareness Campaign, with significant increases in our lung cancer diagnosis rates and the number of patients undergoing curative surgery.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29.03 - DISCUSSANT (ID 3996)
10:50 - 11:05 | Author(s): M.D. Peake
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29.04 - A new medical device for in vivo isolation of circulating tumor cells in non-small cell lung cancer (NSCLC) patients and immunofluorescence identification of ALK (ID 2659)
11:05 - 11:15 | Author(s): L. Gasiorowski, K. Haubold, S. Herold, B. Nowack, K. Pantel, T. Krahn, G. Dworacki, K. Lücke, W. Dyszkiewicz
- Abstract
- Presentation
Background
Circulating tumor cells (CTCs) in the bloodstream of lung cancer patients provide a source for early detection, prognosis, and therapy monitoring. CTCs are currently mostly isolated in vitro from small volumes of patient blood samples. In order to increase the sensitivity of the CTC detection in the peripheral blood GILUPI has developed a novel in vivo method, the CellCollector, which enables the capture of CTCs from the patient´s blood stream with a higher sensitivity and efficacy than existing methods. Enumeration and characterization of those CTCs will serve to improve and monitor clinical cancer early detection and treatment. The aim of this study was to assess the Detektor CANCER01 (DC01) for in vivo isolation of CTCs directly from the blood of NSCLC patients, and to compare it to the CellSearch® method.Methods
The device was inserted in a cubital vein through a standard cannula for thirty minutes. The interaction of target CTCs with the DC01 was mediated by an antibody directed against the epithelial cell adhesion molecule (EpCAM). To confirm the CTCs binding to the wire, the immunohistochemical staining against EpCAM and/ or Cytokeratins as well as CD45 for negative cell selection and nuclei counterstaining was performed. There were enumeration data available for 34 stage I-IIIB NSCLC patients and 8 non cancer patients. For 34 patients, samples were also tested with the CellSearch® method.Results
In this study, we successfully isolated EpCAM-positive tumor cells in the peripheral blood originating from NSCLC patients. We obtained in vivo CTC detection rate of 94% in 32 of 34 NSCLC patients with a median (range) of 13 (0-300) CTCs. In 2 of 34 samples (5.8 %), CTCs were detected by the CellSearch® method. In all matched pairs, the DC01detected the same number or more CTCs compared to the CellSearch® method. The sensitivity was similar for early and late stage NSCLC patients. In the non-cancer patients, no CTCs were detected (100 % specificity). ALK gene rearrangements in NSCLC patients are an indication for targeted therapy with crizotinib. Therefore the staining for CTCs on the CellCollector was enhanced to identify the anaplastic lymphoma kinase (ALK). It could be identified by capturing cells and immunohistochemical staining.Conclusion
Due to a detection rate of over 90% this new device might overcome present limitations in the enrichment of CTCs. This proof of concept study may have important clinical implications, as the implementation of the device into clinical practice may improve early detection, prognosis and therapy monitoring of NSCLC patients. The performed IHC for ALK expression on samples using a novel combination of a new ALK antibody with the high detection rate of the CellCollector offers an alternative to FISH or IHC on tumor tissue. This new technology also allows, as the captured tumor cells are ready using immunofluorescence approaches or qPCR, the possibility of establishing more personalized treatment regiments.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29.05 - Prognostic Impact of the Tumor Size Eliminating the Ground Glass Opacity Component: Modified Clinical T Descriptors of the TNM Classification of Lung Cancer (ID 751)
11:15 - 11:25 | Author(s): S. Nakamura, T. Fukui, T. Taniguchi, N. Usami, K. Kawaguchi, F. Ishiguro, A. Hirakawa, K. Yokoi
- Abstract
- Presentation
Background
The presence of ground glass opacity (GGO) on high-resolution computed tomography (HRCT) is well known to be pathologically closely associated with adenocarcinoma in situ. Recently, measuring the tumor diameter including areas of GGO on HRCT has been reported to possibly overestimate the T status. The purpose of this study was to evaluate the significance of the tumor size measured eliminating the area of GGO on HRCT as a prognostic factor and to propose a refined TNM classification based on modified T descriptors.Methods
Four hundred and seventy-five patients with clinical T1a-T2bN0M0 non-small cell lung cancer underwent surgical resection. All tumors were reclassified based on the diameter measured eliminating the GGO area on HRCT according to the 7th TNM classification of lung cancer. We defined this new classification as modified T descriptors categorized into four groups: mTis+T1a, mT1b, mT2a and mT2b. The overall survival rates of the patients in the current and modified staging groups were evaluated.Results
The 5-year survival rates were 88% and 82% in the patients with T1a and T1b tumors and 90% and 75% in the patients with mTis+T1a and mT1b tumors, respectively. The differences in the survival of the patients classified using mTis+T1a and the other modified T descriptors were more clearly separated statistically than those of the patients classified using the current T1a and other T descriptors.Conclusion
The use of clinically modified T descriptors of the tumor size measured eliminating the GGO component on HRCT may more clearly classify the prognoses of patients with early lung cancer.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
- Abstract
- Presentation
Background
Whole-body metabolic tumor volume (MTV~WB~) is a prognostic index independent of TNM staging, age, gender, performance status and treatment in non-small cell lung cancer (NSCLC). The current TNM staging system is also a prognostic index. However, it lacks advanced quantitative volumetric measurement in the 7[th] edition NSCLC TNM staging system. Integrating quantitative MTV~WB~ measurement into the current NSCLC staging system may make the staging system more quantitative and probably more prognostic. This study’s aim was to determine the utility of a FDG-PET volumetric prognostic (PVP) staging system based on univariate Cox regression models as a new prognostic index in NSCLC.Methods
328 consecutive patients (156 males,172 females) with histologically proven NSCLC and median age of 68.3 years were identified for this retrospective analysis. The PET metabolic tumor volume in the whole body (MTV~WB~) was measured using a semi-automated 3D contouring program on baseline FDG PET/CT scans. The prognostic power for survival status of the PVP staging system was evaluated using the Kaplan–Meier method, Cox regression models, and compared with those of TNM staging system and ln(MTV~WB~) using C-statistic index (Gönen and Heller’s K concordance statistic).Results
There were 46 cases with stage IA, 43 stage IB, 19 stage IIA, 18 stage IIB, 52 stage IIIA, 39 stage IIIB and 111 stage IV NSCLC. At the end of this investigation, 249 patients had died (88.4%). Median follow-up of 79 lost follow-up and known surviving patients was 58 months. On univariate survival analysis the HR of ln(MTV~WB~) was 1.56, and the HRs of TNM stages 2.25, 1.96, 2.89, 4.24, 4.93 and 7.63 for TNM stages IB, IIA, IIB, IIIA, IIIB and IV in reference to stage IA, respectively. These HRs were used for computing the PVP stage of each patient using following equation; PVP stage = [Hazard ratio of ln(MTV~WB~) × ln(MTV~WB~)] × [Hazard ratios of TNM stage]. There was a statistically significant association of better overall survival with lower PVP stage on both univariate [HR of 1.033 (95%CI =1.027 to1.039)] and multivariate [HR of 1.03 (95%CI =1.02 to1.04)] survival analyses. The range of the PVP stage were 0.06 to 87.57. For comparison with 7[th] edition of TNM stage and ln(MTV~WB~), the PVP stage and ln(MTV~WB~) were divided into 7 groups with similar numbers of patients. The C-statistic value of PVP staging system (mean±SE = 0.70±0.014) was statistically significantly higher than those of TNM stage (0.67 ±0.015, p=0.002) and ln(MTV~WB~) (0.67±0.015, p=0.02) on univariate survival analysis. In multivariate Cox regression models adjusted by patient’s age, gender, treatment types, tumor histology and tumor SUV~max~, the C-statistic value of PVP staging system (mean±SE=0.73±0.013) was also statistically significantly higher than that of TNM stage (0.71 ±0.015, p=0.036).Conclusion
The PVP staging system from FDG-PET/CT has better prognostic ability and may prove to be a useful prognostic index in NSCLC. It can be treated as a complement to the TNM staging system.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29.07 - The clinical outcome of non-small cell lung cancer patients with adjacent lobe invasion: proposal for optimal classification according to the status of interlobar pleura in invasion point (ID 1168)
11:35 - 11:45 | Author(s): T. Hishida, Y. Ohtaki, J. Yoshida, G. Ishii, K. Aokage, K. Shimizu, I. Takeyoshi, K. Nagai
- Abstract
- Presentation
Background
In the 7th TNM classification, non-small cell lung cancer (NSCLC) with adjacent lobe invasion (ALI) is classified as T2a regardless of whether across the complete or incomplete fissure. However, no validation analysis has been conducted on this classification. The aim of this study was to analyze the survival of NSCLC patients with ALI with emphasis on the interlobar fissure status at invasion point.Methods
We retrospectively evaluated 2097 consecutive patients with surgically resected NSCLC from 1993 through 2006. Of these, 90 (4.3%) patients had tumors with ALI. Interlobar fissure status of tumors with ALI was examined by using elastic stain. We classified ALI into 2 types: direct ALI beyond incomplete interlobar fissure (no visceral pleurae separating two lobes; ALI-D) and ALI across complete fissure (two lobes separated by 2 visceral pleurae; ALI-A), and compared the clinicopathological features and survival between the groups.Results
The patients with ALI without any other criteria higher than T2b category (n = 60) demonstrated intermediate survival between T2a and T2b tumors (5-year overall survival [OS]: T2a, 61.0%; ALI, 59.6%; T2b, 49.2%). Distinct survival difference was observed between the patients with ALI-A (n = 46) and ALI-D (n = 14) (5-year OS: ALI-D, 85.7%; ALI-A, 52.0%; p = 0.01). The survival of patients with ALI-A was not statistically different from that of patients with T2b tumors, regardless of tumor size (p = 0.85). Conversely, the survival of the patients with ALI-D did not statistically differ from those with T1a or T1b tumors (p = 0.77 and 0.42, respectively).Figure 1Figure 2Conclusion
Interlobar fissure status clearly affected survival of the patients with ALI. ALI should be examined by elastic stains and only ALI-A should be classified as true ALI. We propose that ALI-A tumors ≤ 5 cm should be assigned to T2b but ALI-D tumors do not require adjustment of T descriptor.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
O29.08 - DISCUSSANT (ID 3997)
11:45 - 12:00 | Author(s): M. Reck
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
Author of
-
+
MO11 - Screening and Epidemiology (ID 131)
- Event: WCLC 2013
- Type: Mini Oral Abstract Session
- Track: Imaging, Staging & Screening
- Presentations: 1
- Moderators:P. Zimmerman, J. Bowden
- Coordinates: 10/28/2013, 16:15 - 17:45, Bayside 201 - 203, Level 2
-
+
MO11.02 - Validation of electronic nose exhaled breath VOC profile in discriminating between subjects with early stage lung cancer and healthy ever smokers. (ID 2574)
16:20 - 16:25 | Author(s): P. Zimmerman
- Abstract
- Presentation
Background
Early diagnosis of lung cancer is associated with a better survival. The measurement of volatile organic compounds (VOCs) in exhaled breath using an electronic nose may prove to be a novel, effective and simple technique for screening and diagnosing lung cancer. The aim was to test the validity of the VOC profile in discriminating subjects with early stage lung cancer (I and II) (ESLC) from healthy ever smokers (HS).Methods
243 subjects: 54 ESLC and 189 HS provided a breath sample after tidally breathing through an inspiratory port filter for 5 minutes. It was analysed using a 32 sensor Cyranose 320 (Smiths Detection). Subjects were divided into training (n=159) and independent test set (n=84) groups. Canonical discrimination analyses were performed to determine significance of difference between subject groups and calculate cross validated accuracy (CVV) of the groups using leave one out classification method. Area under the curve (AUC) of Receiver Operating Characteristic Curves were also determined (SPSS V17.0).Results
Validation of the training VOC profile model using an independent test group showed 79% accuracy (p=0.001) in distinguishing ESLC (n=20) from HS (n=64). (AUC 0.933). There was no significant difference in age, lung function and smoking history between the training and test groups.Conclusion
Exhaled breath VOC profile model to discriminate between ESLC subjects and HS was validated in an independent group with a high accuracy. A clinically high sensitivity of the VOC profile model to discriminate between ESLC and HS can be achieved by selecting an appropriate cut point. The cyranose has potential to be a clinically useful diagnostic and screening tool for early stage lung cancer.Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
