September 2021
Volume 21, Issue 9
Open Access
Vision Sciences Society Annual Meeting Abstract  |   September 2021
How does a 2D preview help a 3D search? – An eye tracking study of Digital Breast Tomosynthesis
Author Affiliations & Notes
  • Chia-Chien Wu
    Harvard Medical School
    Brigham and Women's Hospital
  • Nicholas D'ardenne
    University of Pittsburgh
  • Robert Nishikawa
    University of Pittsburgh
  • Jeremy Wolfe
    Harvard Medical School
    Brigham and Women's Hospital
  • Footnotes
    Acknowledgements  NIH-NCI grant CA207490
Journal of Vision September 2021, Vol.21, 2155. doi:https://doi.org/10.1167/jov.21.9.2155
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Chia-Chien Wu, Nicholas D'ardenne, Robert Nishikawa, Jeremy Wolfe; How does a 2D preview help a 3D search? – An eye tracking study of Digital Breast Tomosynthesis. Journal of Vision 2021;21(9):2155. https://doi.org/10.1167/jov.21.9.2155.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Digital Breast Tomosynthesis (DBT) is a technology that renders the 3D volume of the breast as a stack of images that a radiologist can scroll through. DBT has been shown to produce higher hit rates and lower false alarm rates than traditional 2D mammogram (breast x-ray). However, it comes with an additional time cost of about 50 – 100% per case. In the current study, we investigated whether providing a 2D preview before reading DBT can help the radiologist read DBT more efficiently. Two versions of a 2D breast x-ray were tested: 1) standard mammogram (MMG), 2) synthetic “C-View” image – a 2D view synthesized from the 3D DBT data. Results were compared to a no-preview control. Radiologists were allowed to view preview as long as they wanted but they could not revisit the preview once they started to read DBT. Radiologists were asked to mark all potential lesions in DBT by mouse click and were asked whether they would recall the patient for further examination. The results show that providing either mammogram or C-view preview decreased the reading time in DBT, but this benefit was completely offset by the time spent on the preview. Providing either MMG or C-View preview reduced the false alarm rates, but C-view preview also reduced hit rates. Eye movements data show that when there was a preview, radiologists tended to fixate on a lesion faster and cover the less of the breast area than in the no preview condition. Radiologists also made smaller saccades to initiate search after seeing the preview. Our results suggest that providing this type of preview may not save time in reading DBT, though it does add different information which can change search strategy and effectively guide attention.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×