Virtual Library

Start Your Search

A.C. Gower

Author of

  • +

    ORAL 39 - Potential Biomarkers for CT Screening (ID 149)

    • Event: WCLC 2015
    • Type: Oral Session
    • Track: Screening and Early Detection
    • Presentations: 1
    • +

      ORAL39.07 - A Bronchial Genomic Classifier Measured in Airway Epithelial Cells Improves Diagnostic Sensitivity of Bronchoscopy for Lung Cancer (ID 2215)

      16:45 - 18:15  |  Author(s): A.C. Gower

      • Abstract
      • Presentation
      • Slides

      Bronchoscopy is often used for the diagnosis of lung cancer however its sensitivity is imperfect, especially for small and peripheral lesions. Adjunctive methods to improve the sensitivity of cancer detection would reduce the need for more invasive follow-up procedures when bronchoscopy is non-diagnostic. It has previously been shown that gene expression of cytologically-normal bronchial airway epithelial cells is altered in smokers with lung cancer. In this study we evaluated the performance of a bronchial genomic classifier to predict malignancy in an independent cohort of suspect lung cancer patients.

      A bronchial genomic classifier consisting of the expression of 23 genes measured in the airway epithelium was evaluated in a previously published, independent cohort (n=163) of current and former undergoing bronchoscopy for suspect lung cancer. In cases where bronchoscopy was non-diagnostic for malignancy, the performance of the classifier was evaluated using ROC-AUC, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

      In the test set, bronchoscopy led to a diagnosis in 40 of 78 patients with cancer (sensitivity=51%, 95% CI 40-63%). The combination of the classifier with bronchoscopy improved the sensitivity to 96% (95% CI 89-99%; p <0.001); see Table. The prediction accuracy of the classifier was similar in lesions <3cm, as well as across cancer stage and histology. Among the 123 patients with a non-diagnostic bronchoscopy, 38 were ultimately diagnosed with lung cancer (prevalence of 31%). In this group of patients, the classifier had an AUC of 0.81 (95% CI, 0.73-0.88), accurately identifying 35 of the 38 lung cancer patients (sensitivity=92%; 95% CI, 78-98%), and 45 of 85 patients with benign lesions (specificity=53%; 95% CI, 42-63%). Of the 48 patients with a negative classifier result, 45 were diagnosed with benign lesions (NPV=94%, 95% CI 83-99%).

      Table. Performance of bronchoscopy, classifier, and the combined procedures in the test set
      Category Bronchoscopy Classifier[a] Combined
      Total, N 163 123 163
      Lung Cancer, N 78 38 78
      Benign Lesion, N 85 85 85
      Sens. (95% CI) 51% (40-62%) 92% (78-98%) 96% (89-99%)
      Spec. (95% CI) 100% (95-100%) 53% (42-63%) 53% (42-63%)
      NPV (95% CI) 69% (60-77%) 94% (83-99%) 94% (83-98%)
      PPV (95% CI) 100% (90-100%) 47% (36-58%) 65% (56-73%)
      a) The performance of the classifier was evaluated for patients in whom bronchoscopy did not result in a finding of lung cancer (n=123).

      A gene expression classifier measured in bronchial epithelial cells is able to accurately identify those at low risk for lung cancer in patients who have undergone bronchoscopy with non-diagnostic results. Due to the high sensitivity and NPV of the classifier, it could potentially inform clinical decisions regarding the need for further invasive testing for lung cancer in patients whose bronchoscopy is non diagnostic.

      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.