OC120: Contribution of three‐dimensional ultrasound to two‐dimensional ultrasound analysis of benign and malignant breast disease

We studied the benefits of adding three-dimensional (3D) to two-dimensional (2D) ultrasound to diagnose malignant and benign breast disease. This prospective study included 121 patients with clinical or mammographic breast anomalies who underwent 2D and then 3D breast ultrasound centred on the 2D anomaly. Patients with cystic masses were excluded because they are well described with 2D ultrasound. 2D and 3D images were classified into three categories - benign, atypical, and malignant. A single operator performed all ultrasounds. All patients underwent breast biopsies; pathology results, classified into four types, were compared with ultrasound results. The sensitivity, specificity, and predictive values of both ultrasound techniques for diagnosing malignant diseases were calculated, as was the kappa coefficient of concordance between the ultrasound and pathology results. Fisher's exact test was used to compare 2D and 3D ultrasound assessment categories according to pathology results. 3D ultrasound provided the same sensitivity as 2D ultrasound (0.96) and better specificity (1 versus 0.74) for diagnosis of malignant disease. The kappa coefficient of concordance between 2D ultrasound and pathology findings was 0.53 versus 0.97 between 3D and pathology findings. Three identifying criteria in the coronal plane (rim, type of convergences and architectural distortion) were especially helpful for image analysis. 3D ultrasound also provided significantly better diagnostic information for fibrocystic breast disease: 100% (25/25) of cases were characterized on 3D ultrasound, and only 40% (10/25) on 2D ultrasound (P<0.001). 3D ultrasound, although not more accurate than 2D ultrasound for diagnosing breast cancer, improves accuracy of fibrocystic breast disease diagnosis by reducing false-positive results.