The European Congress of Radiology, with a giant octopus in the registration area and a collage of aquatic images, concluded in Vienna this past weekend. As expected, one of the central themes was artificial intelligence (AI), but a presentation that caught my eye was delivered by a radiologist from Egypt. Dr Norran Hussein Said of Cairo University (seen below right) argued that automated breast ultrasound (ABUS) was more accurate than digital breast tomosynthesis (DBT) when used as a diagnostic tool to work up positive findings based on screening mammography for women with dense breast tissue.
Said acknowledged that ABUS has its limitations – e.g. detecting microcalcifications and imaging the area behind the nipple – but argued that the benefits far outweighed any shortcomings. She told the audience: “ABUS has the capability of 3D assessment [of dense breasts] and can differentiate solid from cystic lesions.” The data were gathered by Egypt’s National Breast Screening Programme which has ten mobile mammography units that travel around the country. The digital mammograms were uploaded to the central repository in Cairo and patients were recalled if there were any positive findings.
A total of 242 women with dense breasts were enrolled, where the findings included focal asymmetries, masses, distortions and microcalcifications. Each patient was evaluated with both DBT and ABUS, and the images were evaluated by two independent radiologists. The results – above left, © AuntMinnie.com – showed equal sensitivity in both modalities, but higher specificity for ABUS. ABUS also yielded much higher positive predictive values and better accuracy than DBT.
There were 51 biopsy-proven breast cancers among the 242 women, with ABUS and DBT both agreeing on 43 true positives, while each modality had four false negatives. There were just four false positives produced by ABUS, while DBT had 15. A decision to perform a biopsy could be avoided in 187 cases using ABUS, and in 176 cases with DBT. The second half of Said’s presentation was used to present seven clinical case studies in which she demonstrated both the benefits and limitations of the two imaging modalities. She concluded: “In our study, ABUS has shown a higher accuracy than DBT.”
Writing in AuntMinnie.com earlier this week, Stephen Holloway of Signify Research commented that DBT is well-positioned as a successor to full-field digital mammography (FFDM) in European breast screening programmes. While acknowledging that ABUS has not yet been widely adopted, he drew attention to a hybrid system from CapeRay, combining FFDM and ABUS, that is commercially available in Europe. Now that’s a welcome boost for our dual-modality Aceso system!