Author of

• 18974

  +

  P3.02 - Biology/Pathology (ID 620)

  • Event: WCLC 2017
  • Type: Poster Session with Presenters Present
  • Track: Biology/Pathology
  • Presentations: 1
  • Moderators:
  • Coordinates: 10/18/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)

  • 23955

    +

    P3.02-008 - Non-Invasive Diagnosis of Solitary Pulmonary Nodules Using High-Throughput Targeted DNA Methylation Sequencing of Circulating Tumor DNA (ID 9178)

    09:30 - 09:30  |  Author(s): Y. Gao
    • Abstract

    Background:
    It remains a great challenge to differentiate solitary pulmonary nodules which might cause excessive medical care and unnecessary psychological burden. Disadvantages of previous image-based or invasive methods call for better diagnostic tools .
    
    Method:
    Based on several cohorts, we performed methylation profiling by high throughput bisulfite DNA sequencing in tissue samples to learn methylation patterns that differentiate cancerous from benign lesions by in-depth data mining. Then we filtered out methylation patterns exhibiting high background in cfDNA for plasma sample classification. Notably, given the usual low amount of ctDNA in plasma, we developed an ultra-sensitive library preparation method (AnchorIRIS[TM]) to perform targeted bisulfite DNA sequencing from as low as 1 ng of cell free DNA (cfDNA) or 1 mL of plasma.
    
    Result:
    In the training set (n = 230) which includes 129 malignant specimens with different subtypes (invasive adenocarcinoma, MIA, AIS, squamous carcinoma, small cell and large cell lung cancer) as well as 101 benign specimens in different categories (hamartoma, granuloma/tuberculosis, inflammation and infections), we were able to achieve a preliminary sensitivity of 94.3% ± 4.3% for identification of maglignacies, with a preliminary specificity of 93.6% ± 4.9% against all benign specimens. From an independent validation set of 145 plasma samples, this assay obtained a preliminary sensitivity of 78.5% and a preliminary specificity of 83.3% for differentiating patients with malignant tumor (n = 79) from patients with benign lesions and asymptomatic normal individuals (n = 66). Specifically, our assay is demonstrated to be highly sensitive towards early-stage lung cancer detection, with a preliminary sensitivity of 71.1% in 38 patients with stage Ia lung cancer and 87.5% in 16 patients with stage Ib lung cancer. To further confirm the clinical application of this model, an additional validation study was carried out in an independent center. Our assay obtained an overall sensitivity of 89.2% for 37 malignant tissue samples, particularly with a sensitivity of 100% for adenocarcinoma, and an overall specificity of 85.7% for 21 benign tissue samples. For plasma samples, our assay achieved a sensitivity and a specificity of 80% and 66.7% respectively for 20 malignant and 3 benign samples.
    
    Conclusion:
    We have developed a highly sensitive blood based non-invasive diagnostic assay for differentiation of solitary pulmonary nodules, which can aid clinical decisions for patients with a CT scan positive for lung nodules. To the best of our knowledge, this is the first study to examine the diagnostic value of high throughput targeted DNA methylation sequencing for early stage lung cancer.