James Reza F. Fernandez MD, MS, Presenter: Nothing to Disclose

Linda Hovanessian-Larsen MD, Abstract Co-Author: Nothing to Disclose

Brent Julius Liu PhD, Abstract Co-Author: Nothing to Disclose

Compare lesion assessment and confidence levels of 2D Full-field digital mammography (FFDM) with stereoscopic 3-Dimensional digital mammography (3D DM) and evaluate the potential application of this technology for screening and diagnostic mammography.

200 symptomatic and asymptomatic subjects were imaged using 2D and 3D mammography system. Data collection occurred in Europe under guidelines set forth by mammograhy clinics. Each patient underwent a bilateral breast exam with standard 2D FFDM images (RCC, LCC, RMLO, LMLO). With the breast under the same compression, each patient had 4 additional images corresponding to the standard image at an offset angle of 4-degrees. 3D screening mammography has an average glandular dose (AGD) of approximately 1.39 mGy compared to traditional 2D FFDM AGD of 1.52 mGy, which may affect image quality and noise on a per image basis in 3D, but combined together with the display technology could potentially yield better results. Cancer cases were biopsy proven and non-cancer cases were either negative cases or benign (with or without biopsy). 2D FFDM images were displayed in 2D and merged with offset 4-degree corresponding images to present the 3D display. An independent, non-study participant radiologist evaluated each case in 2D and 3D along with reports to categorize and label each case. This was the ground truth for the studies. 1 board certified radiologist and 2 radiologists specialized in women’s imaging read the cases in 2D and re-read the same cases in 3D after a 3 week washout period. Cases were randomized to remove bias. Evaluations were recorded for negative cases and for findings: location, morphology, edge definition, possibility of malignancy and actionability. Data was analyzed and compared across 2D and 3D reading as well as for reader variability. We had submitted the IRB application and are pending its approval.  

Specificity for negative cases improved while reading in 3D. Lesion assessment for masses improved with 3D. Time spent in reading in 3D was on average shorter than in 2D.  Readers were also more confident labeling normals as being normals in 3D compared to 2D. 

Studies indicate that 3D DM has potential to be better than 2D FFDM in lesion assessment, decreases false positives and improves accuracy of diagnosis.

Evaluation of stereoscopic 3D digital mammography on lesion assessment and confidence levels.

Fernandez, J, Hovanessian-Larsen, L, Liu, B, Evaluation of Stereoscopic 3D Digital Mammography on Lesion Assessment and Confidence Levels: A Preliminary Study .  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11004961.html