Updates from RSNA

December 9, 2024 by JoAnn Pushkin

SBI Announcements
Congratulations to DBI’s MAB member Dr Paula Gordon for the SBI Gold Medal recognition for her outstanding contributions to breast imaging, and to JoAnn Pushkin for being named 2025 SBI Honorary Fellow!

Abstracts
Prevalence screening results from the randomized UK BRAID trial were presented by Dr. Fiona Gilbert at RSNA 2024.  More than 9000 women with dense breasts and negative 2D mammogram aged 50-70 years were randomized to one of four screening groups (over 2000 women in each group): supplemental screening with abbreviated MRI (Ab-MRI), automated whole breast ultrasound (ABUS), contrast-enhanced mammography (CEM), or standard of care (no supplemental screening).

• Cancer detection rates and recall rates were higher for CEM and Ab-MRI than ABUS, but all modalities significantly increased cancer detection¹
• Among those women invited to supplemental screening, participation rates were highest among women invited to have ABUS²
• All cancers detected were small, but cancers seen with contrast methods were smaller³
• All cancers detected by ABUS were invasive, as were 82% and 86% detected by CEM and Ab-MRI, respectively

 Supplemental screening in women with dense breasts results in earlier detection, with Ab-MRI and CEM finding more cancers than ABUS with higher recall rates.

 ¹ Cancer detection rates and recall rates for Ab-MRI, ABUS, and CEM were 17.4/1000, 4.2/1000, and 19.2/1000, respectively and 9.7%, 4.0%, and 9.7%, respectively.

² Approximately 95.5% of invited women completed ABUS vs. 91.8% Ab-MRI and 91.1% CEM.

³ Median invasive cancer size for Ab-MRI, ABUS, and CEM respectively was 10 mm (IQR: 8-15 mm), 22 mm (IQR: 14 – 35 mm) and 11 mm (7 – 15 mm).

 Results of another study presented at RSNA 2024 by Annie Ng, Ph.D., from Kheiron Medical Technologies (now Deep Health), London, UK, found that a 2D-mammographic AI system identified more cancers than a human reader in dense breasts. The retrospective study compared sensitivity of the commercially available Mia AI system to the first human reader in double reading among 306,839 cases from over 236,739 women between 2017-2021. 

• Sensitivity of AI was significantly greater than the human reader in women with dense breasts, whereas there was no statistical difference among women with non-dense breasts¹
• These results were maintained when analyzed by age in women aged 49-55 years or 56-71 years, respectively

 AI as a second reader has the potential to increase cancer detection in women with dense breasts compared to human readers alone.