Eman Shaheen, Presenter: Nothing to Disclose

Chantal Van Ongeval MD, Abstract Co-Author: Nothing to Disclose

Federica Zanca PhD, Abstract Co-Author: Nothing to Disclose

Lesley Cockmartin, Abstract Co-Author: Nothing to Disclose

David Dance PhD, Abstract Co-Author: Nothing to Disclose

Kenneth C. Young PhD, Abstract Co-Author: Nothing to Disclose

Hilde Bosmans PhD, Abstract Co-Author: Co-founder, Qaelum NV Research Grant, Siemens AG

The detection and characterization of microcalcifications play an important role in the early diagnosis of breast cancer. Morphology and number of microcalcifications are two important determinative factors of the differential diagnosis between benign and malignant calcifications. Therefore, we focused on describing microcalcification clusters based on these parameters in a comparative study between 2D full field digital mammography (FFDM) and breast tomosynthesis (BT).

Image data was collected from 46 patients. Fifty microcalcification clusters were simulated into raw projection images of 2D and BT at the same insertion positions for the same patient. The projections were then processed (2D) or reconstructed (BT). The simulated clusters were validated for realistic appearance by radiologists in a previous study where no significant difference was found between real and simulated clusters. In this study, six radiologists evaluated the clusters using the following parameters: morphology in terms of the Le Gal classification determining the dominant type, and the number of calcifications counted in the cluster. The 2D and BT images were read in separate sessions. The observed agreement and a corresponding p-value were reported for the dominant Le Gal type for every observer. The Wilcoxon rank test and linear regression were applied to compare the number of microcalcifications. P-value < 0.05 indicates significant difference.

For the dominant Le Gal, the observed agreements ranged from 0.34 to 0.84 and p-values from 0.076 to 0.611 indicating no statistically significant difference between 2D and BT for all observers. The Wilcoxon test showed significant differences, regarding the number of calcifications in 2D compared to BT, for all observers (p-values<0.005). The percentage of increased number of microcalcifications counted by the radiologists in 2D compared to BT ranged between 31% and 47%.

Our results show that no significant difference was found comparing 2D FFDM and BT for the morphological description based on the Le Gal classification, but 2D outperformed BT in the number of microcalcifications. Further development in BT will address this.

No systematic change in Le Gal classification was found between 2D and BT but the number of counted calcifications was lower in BT suggesting future clinical studies examine microcalcifications in BT.

Shaheen, E, Van Ongeval, C, Zanca, F, Cockmartin, L, Dance, D, Young, K, Bosmans, H, The Comparison between 2D Digital Mammography and Digital Breast Tomosynthesis for Morphological Assessment of Microcalcification Clusters: A Simulation Study.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13012606.html