Rapid intraoperative digital specimen tomosynthesis

• Published in: Journal of clinical oncology Vol. 33; no. 28_suppl; p. 59
• Main Authors: Kaufman, Cary Steven, Behrndt, Valerie, Hill, Laurie, Caro, Rebecca, Nix, Sid, Zacharias, Karen, Evans, Erik, Mahon, Carol, Ness, Karen, Schnell, Nancy
• Format: Journal Article
• Language: English
• Published: 01-10-2015
• ISSN: 0732-183X; 1527-7755
• DOI: 10.1200/jco.2015.33.28_suppl.59

Abstract only 59 Background: About 75% of newly diagnosed breast cancers are not palpable and require image localization to remove the target lesion. Digital specimen mammography devices were developed to identify lumpectomy targets in the operating room. Despite the availability of two-dimensional digital specimen mammography (2D), the re-excision rate for lumpectomy remains significant. Specimen tomosynthesis (3D) may provide a more detailed image than standard 2D with consecutive image slices of the lumpectomy specimens for immediate review. Methods: A consecutive series of 28 breast cancer patients underwent intraoperative specimen imaging with both 2D and 3D imaging. Data recorded for each specimen on each device included 1) accuracy of identification of target lesion, 2) time required to produce comparable images, 3) ease of forwarding images via PACS to radiology, 4) predicted closest margin according to each specimen imaging device compared with final pathologic measured margin, and 5) ease of use by the surgeon and/or nurse. Results: The central focus of all 28 lesions were accurately identified with both 2D and 3D. After a short learning curve, details on the tomosynthesis images were seen not clearly shown on the 2D unit including some spiculated masses and architectural distortions. The location of closest margin was more specific with the 3D device due to “Z-axis” measurements which obtain the vertical distance of the target within the specimen. The spatial relationship of marginal calcifications or marking clips at the edge of the specimen provided by the 3D “Z-axis” was superior to the 2D orthogonal views. Although the time taken to obtain the 3D image was 74 seconds longer than a single image with the 2D device (106 vs. 32 seconds), this was equal to the time taken to obtain the two orthogonal images using the 2D device. Ease of use was equal for both 2D and 3D. Too few patients with positive margins were found to assess a difference in intraoperative positive margin prediction. Conclusions: Our initial experience with 3D tomosynthesis of lumpectomy specimens demonstrate clear images and increased information available for the breast surgeon. Research is planned to further examine the added value of high resolution tomosynthesis in breast surgery.