Virtual Library

Start Your Search

Rozemarijn Vliegenthart



Author of

  • +

    OA06 - Refining Lung Cancer Screening (ID 131)

    • Event: WCLC 2019
    • Type: Oral Session
    • Track: Screening and Early Detection
    • Presentations: 1
    • Now Available
    • +

      OA06.05 - Evaluation of a Deep Learning-Based Automatic Classifier for the Classification of Perifissural Nodules (Now Available) (ID 1928)

      11:00 - 12:30  |  Author(s): Rozemarijn Vliegenthart

      • Abstract
      • Presentation
      • Slides

      Background

      Perifissural nodules (PFNs) comprise approximately 20% of screening-detected nodules and are almost certainly benign. Automatic PFN classification could therefore reduce the number of follow-up procedures required for nodule work-up. Prior work has shown some success in AI classification with limited datasets. Here we evaluate the performance of a new deep convolutional neural network (CNN) for PFN classification, trained on a dataset of nodules retrospectively collected from multiple European centers, including validation on an independent reader-study dataset.

      Method

      Data (1103 Patients, 1557 unique nodules and 3320 nodule images) were collected from three centers in the UK and the Netherlands. Each nodule was categorized into subtypes, including “PFN”, by on-site radiologists. Labels were reviewed centrally, overseen by a single clinician to ensure consistency between sites.

      A CNN classifier was trained to produce a score that classifies nodules as (typical) PFN or not, using five-fold cross validation. The PFN classifier was developed by “transfer learning” from an existing benign-vs-malignant AI trained on the US National Lung Screening Trial.

      To compare the CNN with human performance, independent validation was performed on a separate dataset of 158 benign patients (196 nodules/nodule images) from two of the sites. Three readers (two radiologists and a radiology resident) were asked to label each nodule as typical PFN, atypical PFN, or non-PFN. To match the AI training procedure, only the typical-PFN labels were used in the reader study, and compared to atypical/non-PFN classified nodules.

      Model performance was evaluated by area under the ROC curve (AUC). For the independent validation, Cohen’s kappa was used to measure both the model’s agreement with reader consensus (at least 2 in agreement) and inter-reader agreement. For Cohen’s kappa calculations the CNN score was binarized using a threshold determined from the internal validation data.

      Result

      The mean cross-validated AUC on the internal dataset was 92% (95% CI = 90.6–92.9). For the independent dataset, the classifier labelled 61/196 (31%) as typical PFNs, and reader consensus gave 45/196 (23%). Versus reader consensus, the AUC of the CNN on the reader-study dataset was 96% (95% CI 93.3–98.4). Both the classifier–reader agreement [(k=0.74) 90%] and the inter-reader agreement [(k=0.64–0.79) 88%-92%] were substantial.

      Conclusion

      The performance of the PFN classifier is similar to that of radiologists and is within the inter-reader variability of radiologists. This demonstrates the potential utility of CNN-based systems for automatic PFN classification.

      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.

      Only Active 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 or select "Add to Cart" and proceed to checkout.

  • +

    P1.11 - Screening and Early Detection (ID 177)

    • Event: WCLC 2019
    • Type: Poster Viewing in the Exhibit Hall
    • Track: Screening and Early Detection
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/08/2019, 09:45 - 18:00, Exhibit Hall
    • +

      P1.11-27 - Computed Tomography Screening for Early Lung Cancer, COPD and Cardiovascular Disease in Shanghai: Rationale and Design of a Population-Based Comparative Study (ID 1863)

      09:45 - 18:00  |  Author(s): Rozemarijn Vliegenthart

      • Abstract
      • Slides

      Background

      Volume-based management for lung nodules is associated with a lower rate of unnecessary referral for further work up as compared to diameter-based management in European population. Screening for chronic obstructive pulmonary disease (COPD) and cardiovascular disease (CVD), in addition to lung cancer, may significantly increase the benefits of lung cancer low-dose computed tomography (CT) screening. While this is unclear in Chinese population. The aim of this study is to assess the diagnostic performance of volume-based lung nodule management for lung cancer CT screening as compared to diameter-based management, and to improve the effectiveness of CT screening for COPD and CVD based on quantitative measurement of CT imaging biomarkers in a Chinese screening setting.

      Method

      A comparative population-based study is ongoing, that will include 10,000 asymptomatic participants between 40 and 74 years old from Shanghai urban population in China.

      Participants will be randomized into the intervention and control groups and will undergo a low-dose chest CT scan at baseline and one year after baseline. NELCIN-B3 protocol will be applied in the intervention group. It recommends management of detected solid and part-solid lung nodules based on the volume and volume doubling time (VDT) of a lung nodule. The imaging biomarkers for COPD and CVD, such as emphysema score, bronchial wall thickness from inspiratory and expiratory chest CT scan, and coronary calcium score from ECG-triggered cardiac CT scan will be evaluated. In addition data on laboratory parameters and lung function test will be collected. The participants in the control group will be managed according to the standard hospital protocol based on visual assessment of the CT images. It recommends management of detected lung nodules based on the diameter according to the NCCN Clinical Practice Guideline in Oncology for Lung Cancer Screening. Epidemiological data (eg., risk factors) will be collected through questionnaires for all participants. Four years after the initial assessment the incidence of the three diseases will be evaluated. The design is shown in Figure 1.
      figure1.png

      Result

      The unnecessary referral rate will be compared between the NELCIN-B3 and standard protocol for early detected lung nodules management. The effectiveness of quantitative measurement of CT imaging biomarkers for early detection of lung cancer, COPD and CVD will be evaluated.

      Conclusion

      We expect that the quantitative assessment of the CT imaging biomarkers will reduce the number of unnecessary referrals for early detected lung nodules and improve the early detection of COPD and CVD in Chinese urban populations.

      Only Active 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 or select "Add to Cart" and proceed to checkout.

  • +

    P2.10 - Prevention and Tobacco Control (ID 176)

    • Event: WCLC 2019
    • Type: Poster Viewing in the Exhibit Hall
    • Track: Prevention and Tobacco Control
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/09/2019, 10:15 - 18:15, Exhibit Hall
    • +

      P2.10-16 - Lung Cancer Occurrence Attributable to Passive Smoking Among Never Smokers in China: A Systematic Review and Meta-Analysis (ID 1946)

      10:15 - 18:15  |  Author(s): Rozemarijn Vliegenthart

      • Abstract
      • Slides

      Background

      Quantifying lung cancer occurrence due to passive smoking is a necessary step for policy makers. The aim of this study is to estimate the proportion of lung cancer cases attributable to passive smoking among never smokers in China.

      Method

      A systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. We comprehensively searched six databases up to July 2019 for original observational studies in both English and Chinese languages. Studies that reported relative risks (RR) or odds ratios (OR) for lung cancer occurrence associated with passive smoking in Chinese never smokers were included. For each selected publication, two reviewers assessed publications in English and Chinese independently and assessed the quality of included studies using the Newcastle-Ottawa Scale (NOS). Any disagreements encountered were settled through a consensus. The population attributable fraction (PAF) was calculated using the combined proportion of lung cancer cases exposed to passive smoking and the pooled OR yielded from meta-analysis under the assumption of homogeneity.

      Result

      Thirty-one studies (all designed as case-control) were identified, comprising 9,614 cases and 13,093 controls. The overall proportion of lung cancer cases among never smokers attributable to passive smoking was estimated at 20.5% (95% CI: 16.0% - 24.7%), based on the proportion of lung cancer cases exposed to passive smoking (61.6%) and the pooled OR for passive smoking and lung cancer risk of 1.50 (95% CI: 1.35-1.76). ). The PAF was 15.5% (95%CI: 9.3%-21.0%) based on population-based studies and was 22.7% (95%CI: 16.6%-28.0%) based on hospital-based studies. The subgroup analysis (Table 1) showed that the PAF was similar in non-smoking men (20.9%) and women (21.3%). The proportion of lung cancer cases attributable to household passive smoking was much higher than workplace passive smoking (19.2% vs 10.5%).

      Table 1 Population attributable fraction (PAF) of lung cancer caused by passive smoking among never smokers in subgroups

      subgroup

      No. of studies

      NOS score

      cases

      cases_

      exposed

      cases-exposed(%)

      Pooled OR

      95%CI

      I2

      P

      PAF

      95%CI

      Study year

      before 2000

      13

      6.23

      2600

      1639

      63.04%

      1.70

      1.43-2.03

      48.40%

      0.025

      25.96%

      18.96% - 31.99%

      after 2000

      20

      5.57

      7000

      4348

      62.11%

      1.50

      1.31-1.72

      67.60%

      <0.001

      20.70%

      14.70% - 26.00%

      Gender

      men

      9

      6.22

      809

      473

      58.47%

      1.55

      1.10-2.19

      62.00%

      0.007

      20.75%

      5.32% - 31.77%

      women

      26

      5.77

      7248

      4803

      66.27%

      1.49

      1.34-1.66

      47.80%

      0.004

      21.79%

      16.81% - 26.35%

      Region

      mainland

      23

      5.83

      6468

      3925

      60.68%

      1.55

      1.35-1.79

      67.60%

      <0.001

      21.53%

      15.73% - 26.78%

      non-mainland*

      10

      5.80

      3132

      2062

      65.84%

      1.61

      1.36-1.90

      49.00%

      0.039

      24.94%

      17.43% - 31.19%

      Exposure age

      childhood/adulthood

      5

      6.40

      1219

      972

      79.74%

      1.50

      1.14-1.96

      52.70%

      0.078

      26.58%

      9.79% - 39.06%

      childhood

      4

      6.50

      654

      315

      48.17%

      1.50

      1.08-2.10

      43.30%

      0.152

      16.06%

      3.57% - 25.23%

      adulthood

      5

      6.60

      835

      517

      61.92%

      1.65

      1.05-2.58

      69.70%

      0.010

      24.39%

      2.95% - 37.92%

      Cancer type

      all types

      28

      5.89

      7642

      4759

      62.27%

      1.62

      1.44-1.82

      59.80%

      <0.001

      23.83%

      19.03% - 28.06%

      adenocarcinoma

      10

      6.10

      2485

      1627

      65.47%

      1.52

      1.20-1.91

      67.80%

      0.001

      22.40%

      10.91% - 31.19%

      squamous cell carcinoma

      3

      6.67

      101

      57

      56.44%

      1.36

      0.80-2.32

      0.00%

      0.400

      14.94%

      -14.11% - 32.11%

      Publication language

      EN

      22

      5.87

      7082

      4633

      65.42%

      1.40

      1.27-1.54

      35.40%

      0.049

      18.69%

      13.91% - 22.94%

      CN

      11

      5.72

      2518

      1354

      53.77%

      1.99

      1.63-2.42

      55.90%

      0.012

      26.75%

      20.78% - 31.55%

      Source of passive smoking

      household/workplace

      19

      5.73

      5183

      3668

      70.77%

      1.70

      1.46-1.99

      72.20%

      <0.001

      29.14%

      22.30% - 35.21%

      household

      7

      6.14

      1170

      595

      50.85%

      1.67

      1.32-2.10

      41.40%

      0.115

      20.40%

      12.33% - 26.64%

      workplace

      5

      6.40

      1178

      245

      20.80%

      2.01

      1.62-2.50

      0.00%

      0.514

      10.45%

      7.96% - 12.48%

      Study setting

      hospital-based

      24

      5.60

      7428

      4767

      64.18%

      1.67

      1.47-1.9

      68.50%

      <0.001

      25.75%

      20.52% - 30.40%

      population-based

      9

      6.44

      2172

      1220

      56.17%

      1.31

      1.14-1.51

      0.00%

      0.464

      13.29%

      6.90% - 18.97%

      Conclusion

      Around 20% of lung cancer cases in never smokers, both men and women, are potentially attributable to passive smoking in China. These lung cancer cases in never smokers might be potentially prevented by eliminating exposure to passive smoking, in particular with regards to household passive smoking.

      Only Active 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 or select "Add to Cart" and proceed to checkout.