Click here to


Are you sure ?

Yes, do it No, cancel

Predicting Malignant Mass in Digital Breast Tomosynthesis Using a Multi-Objective Radiomics Model

G Qin1,2*, Z Zhou1 , Y Xu3 , J Ma1,2 , Q Zhang1 ,D Nguyen1 ,J Wang1 , L Zhou3 , W Chen2 , S Jiang1 , (1)Medical Artificial Intelligence and Automation Laboratory, Department of Radiation Oncology, UT Southwestern Medical Center, Dallas, Texas, 75235, (2)Department of Radiology, Nanfang Hospital, Southern Medical University, Guangzhou, China, 510515, (3)Institute of Medical Instrument, School of Biomedical Engineering, Southern Medical University, Guangzhou, China, 510515


(Monday, 7/30/2018) 3:45 PM - 4:15 PM

Room: Exhibit Hall | Forum 6

Purpose: The purpose of this study is to apply a multi-objective radiomics model for predicting the malignancy of the mass-like lesions in Digital Breast Tomosynthesis (DBT).

Methods: DBT is a newly developed three-dimensional (3D) imaging modality, which holds the potential to improve the accuracy of mammography through reducing the tissue overlap. Since the increasing of ability to detect mass-like lesions, more challenging is faced when distinguishing benign from malignant mass. In this study, a total of 963 cases with diagnosed mass-like lesions were retrospectively used for model training and testing. Patients with Infiltrating Ductal Carcinoma (IDC), Ductal Carcinoma in Situ (DCIS), Invasive Lobular Carcinoma (ILC), Adenofibroma, Cystic Hyperplasia, Cyst and Hyperplasia were included. The follow up time was about 24 months and malignant cases were confirmed by biopsy or surgical pathology. DBT images were contoured and reviewed by 3 radiologists each with more than 5 years’ experience in breast diagnosis. (try to modify the above sentences to be different from the other abstract) 497 Image features including intensity features, textural features and geometric features were extracted and selected based on the multi-objective model. Support vector machine (SVM) with radial basis function kernel was used for building the predictive model.

Results: In the predictiion model, accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were used as the evaluation criteria. The multi-objective radiomics model shows a diagnostic accuracy of 79.1%, AUC of 85.2%, sensitivity of 77.3% and specificity of 77.9%.

Conclusion: This study demonstrated the feasibility of using multi-objective radiomics model on DBT images to predict the malignancy of the mass-like lesions. The model can handle a large number of mass-like lesions in the breast cancer screening, and allow for improving the accuracy of diagnosis and increasing the radiologists' work efficiency.

Funding Support, Disclosures, and Conflict of Interest: Supported by Science and Technology Planning Project of Guangdong Province, China(No.2016ZC0058) and Medical Scientific Research Foundation of Guangdong Province, China(No. A2017496).


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email