Abstract Archives of the RSNA, 2010
James Schafer MD, Presenter: Nothing to Disclose
Marilyn A. Roubidoux MD, Abstract Co-Author: Nothing to Disclose
Mitra Noroozian MD, Abstract Co-Author: Nothing to Disclose
Caroline Plowden Daly MD, Abstract Co-Author: Nothing to Disclose
Lubomir M. Hadjiiski PhD, Abstract Co-Author: Nothing to Disclose
Mark Alan Helvie MD, Abstract Co-Author: Institutional grant, General Electric Company
Consultant, General Electric Company
Jamie Myles PhD, Abstract Co-Author: Nothing to Disclose
Heang-Ping Chan PhD, Abstract Co-Author: Nothing to Disclose
Paul L. Carson PhD, Abstract Co-Author: Research collaboration, General Electric Company
Research collaboration, Sonetics Inc
Research collaboration, ZONARE Medical Systems, Inc
Research collaboration, Light Age, Inc
To assess whether radiologists prefer digital breast tomosynthesis (DBT) or clinical mammogram spot views (SV) for characterizing breast masses.
IRB approval and informed consent obtained. A prototype DBT unit developed with GE Global Research acquired DBT images of patients scheduled for breast biopsy of a mass. DBT was performed in CC, Lateral or MLO projections, reconstructed at 1 mm slice intervals. From 68 patients, 96 case pairs (49 benign, 47 malignant) were arranged using a clinical SV and a DBT image volume of comparable projection. The DBT image volume was digitally cropped to appear similar to the SV dimensions. Two MQSA approved fellowship trained mammographers (blinded to clinical information) separately viewed the case pairs in random order. Reader (A) had prior experience whereas reader (B) had no experience viewing DBT. Viewing the DBT and the SV at the same time for each case pair, preference was measured by the question "Your confidence in characterizing this mass is..." and using a 5-point Likert scale, ‘much better’,‘somewhat better’, ‘equal with both’. These preference data were compared to mammogram density categories, determined from a prior reader of same cohort cases. For analysis, BIRADS 1 and 2 density images were combined as low density subgroup (45% of images)and BIRADS 3 and 4 as high density subgroup (55% of images). Mean size of masses was 18 mm. Statistical analysis with Wilcoxon Signed Rank test. Additional readers in progress.
Of 96 case pairs, readers A and B strongly preferred DBT in 18(19%) and 17(18%), preferred DBT in 40(42%) and 30(31%), equal preference in 26(27%) and 35(36%), preferred SV in 11 (11%) and 13(14%) respectively, and strongly preferred SV in only 1 by each reader (1%) (a fat-density mass).There was no significant difference (p=0.55) in preference ratings between readers. DBT was preferred to SV overall (n=96, p <0.0001), and by subgroups, in CC views (n=52, p<0.0001), MLO/Lateral views (n=44, p=0.0006), high density ( p<0.0001) and low density ( p=0.002) images. On average, DBT was rated equal or preferred to SV in 87% of the case pairs.
DBT was equal or preferred to SV for mass characterization regardless of mammogram projection, breast density, or reader experience with DBT. USPHS CA 91713, CA 95153, CA 120234.
DBT may be as useful or more useful than SV views to evaluate breast masses.
Schafer, J,
Roubidoux, M,
Noroozian, M,
Daly, C,
Hadjiiski, L,
Helvie, M,
Myles, J,
Chan, H,
Carson, P,
Digital Breast Tomosynthesis Compared with Spot Views: Radiologist Preferences. Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL.
http://archive.rsna.org/2010/9011752.html