
CHICAGO - Ultrasound is not needed to identify breast cancer in women with fatty tissue, suggests a study presented November 28 at the RSNA annual meeting.
Dr. Grace Lin from Kaiser Permanente in California highlighted results from her team's study, which showed that diagnostic mammography alone has a very low risk of missed cancer in these women.
"The ultrasound we are suggesting omitting is the targeted diagnostic ultrasound when assessing a palpable lump in a patient who has fatty breast density and no corresponding finding on mammogram," Lin told AuntMinnie.com. "Our data has shown that when applying our exclusions, the probability of missing cancer is the same as what has been established in benign categories."
Supplemental ultrasound, along with mammography, is typically used for standard workup for women who present with palpable lumps. However, previous studies suggest that ultrasound's overall yield is unclear.
Lin and colleagues wanted to find out if diagnostic mammography alone is good enough to identify breast cancer. They compared this with a combined mammogram and ultrasound method and studied women with almost entirely fatty breast density who also present with palpable lumps.
They looked at data from 2,066 women with 2,149 distinct palpable lumps who were evaluated with mammography and ultrasound. Of these, 1,486 women (72%) had no mammographic correlate.
The team identified 136 cancers (7%), of which one had no mammographic correlate but was seen by ultrasound. In other words, the cancer risk/negative predictive value for using mammography alone with no correlate for a palpable symptom is 0.07%.
Lin said that while the team recognizes that patients, providers, and radiologists have come to rely on the ultrasound to decrease anxiety, the group also wants to provide evidence supporting the practice of omitting the ultrasound in specific conditions. This way, radiologists who seek to practice evidence-based medicine have the data to support them, she added.
"Ultimately, it is up to each radiologist's discretion regarding what specific value they seek to add with the ultrasound," she told AuntMinnie.com. "In our perspective, the main goal of breast imaging is to diagnose cancer and we want to give the patient and the radiologist the freedom to safely omit the ultrasound without fear of missing cancer."
During the presentation, Lin said that while Kaiser Permanente does not employ this practice, they are awaiting paper publication before presenting results to their administrators.
"There's some discomfort with changing practice. Radiologists are creatures of habit," Lin added. "I think it [paper] is a good starting point."
The research was lead by Dr. Dawn Nwamuo from Kaiser Permanente Northern California. Along with Lin, co-authors on the study include Dr. Shivani Gillon from the University of California, San Francisco and Rochester General Hospital; and Dr. Vignesh Arasu, PhD, from Kaiser Permanente Northern California.



![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=100&q=70&w=100)







![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=112&q=70&w=112)







