Virtual Library

Start Your Search

I. Yang



Author of

  • +

    P2.19 - Poster Session 2 - Imaging (ID 180)

    • Event: WCLC 2013
    • Type: Poster Session
    • Track: Imaging, Staging & Screening
    • Presentations: 1
    • +

      P2.19-007 - Pulmonary Nodule Detection by Junior Medical Staff is Improved by Digital Tomosynthesis Compared to Chest X-Ray (ID 1110)

      09:30 - 09:30  |  Author(s): I. Yang

      • Abstract

      Background
      Junior doctors may fail to detect subtle pulmonary pathology on plain chest X-ray (CXR). Digital tomosynthesis (DT) is an emerging radiographic technique that provides multiple coronal chest images at only 2% of the radiation of a standard chest CT. Previous studies have demonstrated that pulmonary nodule detection sensitivity is three times greater with DT compared to CXR. We investigated whether DT can increase nodule detection rates by junior doctors compared to CXR.

      Methods
      Ten volunteer junior doctors (post-graduate years 1-3) at The Prince Charles Hospital in Brisbane, Australia, a secondary and tertiary referral hospital, were recruited to view CXR and DT images of 11 patients. All patients had CXR, DT and CT images acquired within a 30 day period for the evaluation of lung nodules. CT images (Philips Brilliance, Philips Medical Systems, Best, Netherlands), with collimation 0.625 mm and reconstructed slice width 0.9 mm, reported by experienced radiologists, served as the gold standard. DT images, consisting of 60 exposures through a 30° arc, were acquired using the GE Definium 8000 Xray Unit (GE Healthcare, Little Chalfont, United Kingdom), with simultaneous CXR as a scout image. Nine of these patients had at least one nodule >10 mm on CT, with two control patients without nodules. All participants undertook brief training to familiarise them with DT images one week prior to the study. In the study session, participants were showed anonymised CXR and DT images in random order and asked to mark “definite” or “possible” pulmonary nodules electronically. The markings were compared to CT detected “true” nodules. Markings made where there were no true nodules on CT were recorded as false positives. The time taken to view each image was measured. Participants completed a brief survey after viewing the images.

      Results
      Nodule detection sensitivity, represented by the proportion of true nodules marked “definitely” present, was significantly higher using DT than CXR (28/65 [43%] versus 3/70 [4%], χ[2], p<0.001), as was the proportion of nodules marked either “definitely” or “possibly” present (32/65 [49%] versus 13/70 [19%], χ[2], p<0.001). When considering instances where a nodule was marked either “definitely” or “possibly” present, where there was no true nodule on CT, significantly fewer false positives were made, on average, when viewing DT compared to CXR (0.36 versus 1.18 false positives per image, t-test, p<0.001). Although the time taken to view each DT image was statistically significantly longer than for each CXR image (86.9 seconds versus 67.9 seconds, t-test, p<0.01), the absolute difference was small. Ninety percent of participants agreed that they could identify nodules more confidently with DT than CXR.

      Conclusion
      In this study, junior doctors correctly identified more pulmonary nodules using DT compared to CXR and reported fewer false positive results. The time taken to view DT images was slightly longer than for CXR images, but this difference was small. Despite the small sample size, this pilot experiment has shown that DT may potentially improve identification of pulmonary nodules by junior doctors and a larger study is underway.