Margarita Louise Zuley MD, Presenter: Institutional Grant support, Hologic, Inc Scientific Advisory Board, Hologic, Inc

Ben Guo PhD, Abstract Co-Author: Nothing to Disclose

Marie Adele Ganott MD, Abstract Co-Author: Nothing to Disclose

Andriy I. Bandos PhD, Abstract Co-Author: Nothing to Disclose

Victor J. Catullo MD, Abstract Co-Author: Nothing to Disclose

Amy Hwei-Mei Lu MD, Abstract Co-Author: Nothing to Disclose

Amy Elizabeth Kelly MD, Abstract Co-Author: Nothing to Disclose

Maria Lena Anello DO, Abstract Co-Author: Nothing to Disclose

Gordon Scott Abrams MD, Abstract Co-Author: Nothing to Disclose

Denise Chough MD, Abstract Co-Author: Nothing to Disclose

To compare lesion visibility and diagnostic accuracy of cone beam computed tomography (CBCT) and tomosynthesis (DBT) to MRI and digital mammography (FFDM)

IRB approval was obtained. From 04/16/2009 to 06/21/2011,, 178 mass lesions in 151 consecutively consenting women underwent FFDM, DBT, CBCT and contrast enhanced MRI prior to percutaneous biopsy. 97 CBCTs were unenhanced (NC-CBCT) and 81 had contrast (CE-CBCT). DBTstudies were unenhanced. Histopathology established truth. A nonparticipating radiologist marked each lesion location. A retrospective fully crossed, balanced reader study was performed with 7 MQSA qualified academic breast radiologists who recorded lesion visibility in each mode and if visible provided a probability of malignancy (POM) score on a 100 point scale. For each mode, ROC curves were obtained by a vertical average of the reader specific curves. Statistical analyses accounting for correlation and random reader effects were performed using the MRMC analysis (DBM MRMC, v.3.0) for area under the ROC curve (AUC) and using the generalized linear mixed model (proc glimmix, SAS, v.9.3) for visibility.  

100 benign and 78 malignant masses were included. Average size was 19.7 mm (median 14mm, range 4-100mm). Percentage of visible lesions differed (88% FFDM, 91% DBT, 82% CBCT [81% NC-CBCT sub-set, 84% CE-CBCT sub-set] and 93% MRI). For visualization, MRI was significantly better than CBCT (p<0.01) but not DBT (p=0.36). FFDM and DBT were not significantly different (p=0.18), but DBT was superior to CBCT (p<0.01). AUCs also varied with modality (MRI 0.88, CBCT 0.75, CE-CBCT 0.79, DBT 0.84, FFDM 0.81). MRI was significantly better than CBCT (p<0.01), but not DBT (p=0.24). FFDM and DBT were not significantly different (p=0.31), but DBT was superior to CBCT (p<0.01).

For masses MRI has the highest accuracy and visibility and was significantly better than CBCT but not DBT. CBCT accuracy and visibility improve with use of contrast but further improvements are necessary for use as an alternative to MRI, FFDM or DBT.  

Tomosynthesis may possibly be a viable alternative to MRI for breast mass evaluation.

Zuley, M, Guo, B, Ganott, M, Bandos, A, Catullo, V, Lu, A, Kelly, A, Anello, M, Abrams, G, Chough, D, Comparison of Visibility and Diagnostic Accuracy of Cone Beam Computed Tomography, Tomosynthesis, MRI and Digital Mammography for Breast Masses.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13022530.html