Evaluation of malignant breast lesions in the diagnostic setting with cone beam breast computed tomography (Breast CT): feasibility study.

Seifert, Posy; Conover, David; Zhang, Yan; Morgan, Renee; Arieno, Andrea; Destounis, Stamatia; Somerville, Patricia; Murphy, Philip F

To investigate the feasibility of noncontrast and contrast-enhanced cone beam breast Computed Tomography (CT) in demonstrating malignant breast lesions in the diagnostic setting. This Institutional Review Board approved, Health Information Portability and Accountability Act compliant, prospective study enrolled BI-RADS four and five patients from 2008 to 2010. Eighty-seven subjects had noncontrast breast CT, 42 had contrast-enhanced breast CT (CE-breast CT) with 70 pathologically confirmed cancer diagnoses. All 70 comprise the study cohort for noncontrast breast CT, and 23 who had CE-breast CT comprise the cohort for CE-breast CT. All had diagnostic work-up. Patient age, breast density, lesion size and characteristics, biopsy method, and core pathology were recorded. A Fisher's exact test was used to detect a difference in detectability. For agreement in size measurement between the imaging modalities, a paired t-test was employed. Reported p-values were based on 2-sided tests. Two one-sided tests were calculated to determine equivalence within ±0.3 cm at a 90% significance level. Noncontrast breast CT identified 67 of 70 malignant lesions, detected by diagnostic work-up. CE-breast CT identified 23 of 23 index malignant lesions and in addition, found three malignant lesions in three cases not previously detected. Noncontrast breast CT demonstrated the index lesion in 67 of 70 cases and CE-breast CT demonstrated the index lesion in all 23 cases. An additional three new malignant lesions not seen with conventional diagnostic work-up were detected. In this preliminary study, breast CT with or without contrast was shown to be accurate at identifying malignant breast lesions in the diagnostic setting.