Automated Breast Ultrasound Tied to Fewer False-Positives in Dense Breasts

Automated breast ultrasound (ABUS) detected breast cancer at a rate similar to handheld ultrasound (HHUS) in women with dense breasts while producing fewer abnormal interpretations, fewer false-positive initial assessments, and fewer recommendations for short-interval follow-up, according to an observational study published in the October 2026 issue of the Journal of the American College of Radiology.

The analysis, drawn from more than 25,000 supplemental screening examinations in US community practice, was not designed to establish sensitivity or specificity, and ABUS may not be appropriate for every patient with dense breast tissue.

According to the investigators and an accompanying news release, this is the first study to directly compare the screening performance of the 2 modalities in clinical practice in the United States.

Why does supplemental ultrasound screening matter for patients with dense breasts?

Use of screening ultrasound alongside mammography has expanded in recent years, in part because state laws require that women undergoing screening mammography be notified when they have heterogeneously or extremely dense breasts. The news release noted that federal rules now carry similar notification requirements and explained that dense tissue and lesions both appear white on a mammogram, which can allow abnormalities to go undetected.

Prior research has shown that adding HHUS to mammography detects an additional 1 to 4 cancers per 1,000 examinations among women with a negative screening mammogram, but HHUS has also been associated with a high percentage of false-positive examinations, increased biopsy rates, and a high proportion of benign biopsy findings. HHUS is also known to have limitations related to operator dependency and a lack of reproducibility.

ABUS was approved in 2012 for breast cancer screening as an adjunct to screening mammography and remains an emerging technology. Similar to screening mammography, an ABUS examination can be acquired by a technologist and interpreted later by a radiologist.

How was the study designed?

Investigators used observational data from 3 Breast Cancer Surveillance Consortium (BCSC) registries: the Metro Chicago Breast Cancer Registry, the San Francisco Mammography Registry, and the Vermont Breast Cancer Surveillance System. The analysis included screening ultrasound examinations performed from 2017 through 2020 at 27 radiology facilities among women aged 40 to 74 years with heterogeneously or extremely dense breasts, no personal history of breast cancer, and no prior mastectomy.

In total, 25,328 screening examinations (4,218 ABUS and 21,110 HHUS) were performed among 21,446 women and interpreted by 98 radiologists. Each examination was followed for 90 days for a diagnosis of ductal carcinoma in situ (DCIS) or invasive breast cancer. Examinations were excluded if patients reported symptoms other than breast pain or had not had a mammogram within the prior 6 months.

Outcomes included abnormal interpretation (recall) rate, short-interval follow-up recommendation rate (final BI-RADS 3), biopsy recommendation rate, cancer detection rate, false-positive rates, and positive predictive values. Adjusted estimates accounted for BCSC 5-year invasive breast cancer risk, facility type and profit status, time since last mammogram, and screening round.

The 2 groups were similar by age and breast density, although ABUS was more often performed in non-Hispanic White women (80% vs 65%), in later study years, and in office settings rather than hospitals.

Did ABUS reduce false-positive results compared with handheld ultrasound?

In adjusted analyses, ABUS was associated with:

• A lower abnormal interpretation rate: 14.9% vs 19.2% (P < .0001)
• A lower short-interval follow-up recommendation rate: 3.0% vs 14.6% (P < .0001)
• A lower false-positive initial assessment rate: 14.6% vs 19.1% (P < .0001)

These figures correspond to absolute differences of approximately 4 percentage points for abnormal interpretation, nearly 12 percentage points for short-interval follow-up, and 4.5 percentage points for false-positive initial assessment.

The pattern held in a subset of 5 facilities that used both modalities. At those sites, abnormal interpretation rates were 16.9% for ABUS and 17.6% for HHUS, and short-interval follow-up rates were 35.2 vs 131.6 per 1,000 examinations, respectively.

How did cancer detection and biopsy rates compare?

Cancer detection and biopsy measures did not differ to a statistically significant degree between modalities. The unadjusted screen detection rate was 2.6 per 1,000 examinations with ABUS vs 1.7 per 1,000 with HHUS (P = .25). The positive predictive value of biopsy was 10.0% for ABUS vs 6.3% for HHUS, numerically about 50% greater but not statistically significant (P = .23).

Adjusted biopsy recommendation rates were 3.0% for ABUS and 2.9% for HHUS, and adjusted false-positive biopsy recommendation rates were 2.7% for both modalities.

Overall, 35 cancers were detected by HHUS and 11 by ABUS. The proportion of detected cancers that were in situ was 26% with HHUS vs 9% with ABUS (P = .24), a difference that was not statistically significant.

Why did the 2 modalities produce different recall patterns?

The type of initial recall differed by modality. ABUS was associated with a higher rate of initial BI-RADS 0 assessments (15% vs 8%), whereas HHUS produced far more BI-RADS 3 assessments at initial interpretation (10% vs 1%).

The investigators noted that this distinction is clinically appropriate. Because ABUS images are interpreted after the patient leaves, findings typically require diagnostic HHUS to resolve and are therefore designated BI-RADS 0. With HHUS, a sonographer or radiologist can obtain additional focused images in real time. The authors speculated that many ABUS recalls may relate to imaging artifacts that are eliminated during diagnostic workup, allowing a negative or benign final assessment. They added that reader training and experience likely play an important role in the accuracy of ABUS, whereas operator dependence and image quality variation have a greater effect on HHUS.

What are the key takeaways for oncology nurses?

• Prepare patients for what a recall means. Under ABUS, patients may be more likely to receive an initial BI-RADS 0 result requesting additional imaging, which indicates an incomplete assessment rather than a cancer diagnosis. In a recent Oncology Nursing News blog, Emily Mason Beard, BSN, RN, OCN, CBCN, described how a BI-RADS 0 report interpreted without clinical context led a patient to fear a cancer diagnosis, underscoring the value of timely nurse explanation.
• Expect fewer short-interval follow-up recommendations with ABUS. In this cohort, about 3% of ABUS examinations ended in a BI-RADS 3 short-interval follow-up recommendation vs nearly 15% with HHUS, which may reduce repeat imaging visits and associated waiting for patients.
• Discuss modality eligibility. Patients with chest-area implanted devices or piercings, breast implants, prior breast surgery, or sensitivity to the examination lotion may not be candidates for ABUS.
• Set expectations about biopsy. Biopsy recommendation and false-positive biopsy rates were similar between modalities, so patients should understand that a biopsy recommendation remains possible with either approach.
• Support shared decision-making. Patients notified of dense breast tissue may ask about supplemental screening options; nurses can explain that ABUS and HHUS showed similar cancer detection in this study, while noting that it was observational and did not measure sensitivity.

References

1. Rauscher GH, Alsheik N, Eavey J, et al. Comparison of supplemental breast cancer screening outcomes for automated versus handheld ultrasound. J Am Coll Radiol. 2026;23(10):1936-1945. doi:10.1016/j.jacr.2026.06.018
2. Study finds automated breast ultrasound is an effective alternative to handheld ultrasound for breast cancer screening. News release. Elsevier. October 8, 2026. Accessed October 9, 2026. https://www.elsevier.com/about/press-releases/study-finds-automated-breast-ultrasound-is-an-effective-alternative-to-handheld-ultrasound-for-breast-cancer-screening
3. Beard EM. Why using AI to interpret mammogram results can cause unnecessary panic. Oncology Nursing News. September 16, 2026. Accessed October 9, 2026. https://www.oncnursingnews.com/view/why-using-ai-to-interpret-mammogram-results-can-cause-unnecessary-panic