Giovanna Mariscotti, Abstract Co-Author: Nothing to Disclose

Manuela Durando, Presenter: Nothing to Disclose

Laura Martincich MD, Abstract Co-Author: Speaker, Bracco Group Consultant, Bayer AG Speaker, ABC Medical Imaging

Enrica Caramia, Abstract Co-Author: Nothing to Disclose

Pier Paolo Campanino, Abstract Co-Author: Nothing to Disclose

Andrea Luparia, Abstract Co-Author: Nothing to Disclose

Laura Bergamasco, Abstract Co-Author: Nothing to Disclose

Paolo Fonio, Abstract Co-Author: Nothing to Disclose

Giovanni Gandini MD, Abstract Co-Author: Nothing to Disclose

To compare the diagnostic performance of DM with that of DM combined with DBT for the detection of ILC.

We conducted a retrospective multi-reader blinded study, including 6 radiologists with experience of breast imaging ranging between 15 and 4 years and DBT experience ranging between 4 years and no experience (inexperienced readers underwent a DBT training session prior to the study). The radiologists interpreted 56 examinations of women (mean age 59.4 years, range 40-78) with 68 newly diagnosed ILCs, proved at definitive histology (tumours mean size 30.1 mm; range 5-95). All women, who signed an informed consent, underwent mammographic bilateral two standard views in Combo mode: DM and DBT images were acquired within a single compression for each projection. The readers, blinded to histology, were asked to primarily detect lesions in 2D images alone, reporting in previously predisposed modules breast density, mammographic signs with BI-RADS category, localization and size of lesions; then, they were asked to detect lesions by viewing DBT images in conjunction with 2D images, reporting the same parameters. Two experienced radiologists, non-participants in the study, compared the results of each reader to pathology. A statistical analysis was performed to evaluate performances and inter-observer agreement.

All 6 readers had a significantly higher sensitivity for detection of ILC by using DM combined with DBT (ranging between 84-91%) than DM alone (77-80%) (p=0.0002). There was no significant difference (p=0.29) in specificity values (DM alone 56.6-92.3%, DM combined with DBT 69.2-92.3%). The diagnostic accuracy was higher considering DM combined with DBT (ranging between 80.3-91.4%) than DM alone (69.2-85.5%) (p=0.0029). Considering breast density, for the readers the relative risk to miss a ILC in dense breasts (30/56) by using DM alone compared to DM with DBT was 2.0 (1.3-3.3) (p=0.0007) (i.e. a 100% increase in the probability of missing the cancer). In the no dense breasts (26/56), there were no significant differences (p=0.34).

The addition of DBT significantly increased sensitivity and diagnostic accuracy of DM in detecting ILCs, especially in dense breasts.

The use of Digital Breast Tomosynthesis in addition to Digital Mammography (DM) improve diagnostic performances of DM and could be helpful in detecting Invasive Lobular Carcinomas.

Mariscotti, G, Durando, M, Martincich, L, Caramia, E, Campanino, P, Luparia, A, Bergamasco, L, Fonio, P, Gandini, G, Comparative Study with Digital Mammography (DM) vs. DM Combined with Digital Breast Tomosynthesis (DBT) for the Detection of Invasive Lobular Carcinoma (ILC) .  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13024159.html