(Sunday, 7/12/2020) [Eastern Time (GMT-4)]
Room: AAPM ePoster Library
Purpose: An image reconstruction method was developed to facilitate Biopsy Guidance using Real-Time Digital Breast Tomosynthesis.
Method: We created a Defrise Phantom (105x105x50 mm), emulating a low-frequency periodic pattern of adipose and glandular breast tissue. The insertion of a highly attenuating metallic needle was simulated transversely and centrally through the phantom. 60 x-ray projections were simulated with a circular x-ray source motion above the phantom. One revolution was completed every 15 projections, making a total of 4 revolutions. The plane of the needle was then reconstructed in a commercial software. Each reconstruction utilized a batch of 15 spatially assorted projections. A reconstruction method was performed to create time-dependent tomosynthesis image reconstructions, wherein; the phantom background was reconstructed using the assortment of 15 projections while the reconstruction image of the needle was generated through its immediately preceding 15th projection alone. This ensured a real-time imaging for the needle while still maintaining an accurate rendering of its position relative to the phantom background. The background and the needle reconstructions were segmented and superimposed to create a real-time image of the needle being inserted into the phantom.
Results: We found that position of the end of the needle, as calculated from the segmented reconstruction technique, matched the known position in the phantom accurately. Compared with the conventional unsegmented reconstruction, this method improved the accuracy by 86.9%.
Conclusion: We are able to demonstrate the feasibility and accuracy of a tomosynthesis technique that can offer real time guidance for biopsy or surgical procedures. We expect our method to be superior to conventional 3D Tomosynthesis methods as it can ensure a more accurate excision of the target tissue from the breast, less chance of missing the target, quicker procedure durations, lower radiation dose as well as lesser patient discomfort.
Funding Support, Disclosures, and Conflict of Interest: W81XWH-18-1-0082 - Department of Defense Breast Cancer Research Program, IRSA 1016451 -the Burroughs Welcome Fund, 1R01CA196528 - National Institute of Health, IIR13264610 - Susan G. Komen. Equipment support- Analogic Inc., Barco NV, Real Time Tomography. Andrew Maidment is a shareholder of RTT and a member of the scientific advisory board.
Keywords
Breast, Reconstruction, Tomosynthesis
Taxonomy
IM- Breast X-Ray Imaging: Digital Breast Tomosynthesis (DBT)
Contact Email
