RSNA: Breast ultrasound screening hampered by low positive predictive value

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon

CHICAGO—Technologist performed handheld screening breast ultrasound offered to women with heterogeneous or dense breasts can detect small mammographically occult breast cancers with a cancer detection rate of 0.32 and a cost of $55,000 per breast cancer, according to a retrospective review presented Nov. 27 at the 97th Scientific Assembly and Annual Meeting of the Radiological Society of North America (RSNA). However, the overall positive predictive value of screening ultrasound is low.

On Oct. 1, 2009, Connecticut became the first state to pass a law that required radiologists to include breast density in their reports and to inform patients with heterogeneous or extremely dense breasts that they may benefit from the addition of screening breast ultrasound or MRI. The law also requires insurers to pay for screening ultrasound in those patients with dense breasts. 

Prior to October, 2009, Yale University Medical Center had not performed breast screening ultrasound, explained Kathryn Greenburg, MD, chief resident, division of clinical radiology at Yale School of Medicine in New Haven, Conn. Thus, she and her colleagues sought to determine the performance of handheld screening breast ultrasound in these women during the first year of the law’s implementation. Greenburg noted that patients were not automatically instructed to schedule screening ultrasound, but instead, they were encouraged to discuss the risks and benefits with their primary care providers.

The first year produced a database of more than 13,000 women, receiving more than 14,000 mammograms. From this group, 1,038 breast ultrasound exams were performed. Due to the exclusionary criteria, Greenburg et al performed a retrospective review on 937 women with dense breasts on mammography and who subsequently underwent handheld screening breast ultrasound at Yale facilities from Oct. 1, 2009 through Sept. 30, 2010. 

All studies were initially performed by a breast ultrasound technologist, and a radiologist immediately reviewed and interpreted each exam.

The average age of the study population was 52 years, and the average length of time between the mammogram and the ultrasound was 60.4 days. 

Of 937 women, 65.6 percent of the patients were low-risk, 13.1 percent were intermediate-risk and 12.1 percent were high risk for breast cancer. The researchers classified unknown risk factors at 9.2 percent.

Of the screening ultrasounds, 75 percent were classified BI-RADS 1 or 2, 20 percent as BI-RADS 3 and 5 percent as BI-RADS 4. Of 60 biopsies performed in 52 patients, nine were BI-RADS 3 lesions and 51 were BI-RADS 4. “While the BI-RADS 3 rate was very high, with 20 percent of women requiring short-interval follow-up, but to date, none of these women have been found to have a malignancy,” Greenburg said.

Overall, 57 of the 60 biopsies were benign and three of the 60 biopsies were malignant (all BI-RADS 4). Greenburg reported that all three cancers were found in post-menopausal patients and were solid masses, including a 9 mm and 4 mm infiltrating ductal carcinoma and a 5 mm ductal carcinoma in situ. One cancer was found in each of the three risk groups, and 4.7 percent of patients had a false positive exam. 

The overall positive predictive value for biopsy of BI-RADS 4 masses was 5.9 percent. The PPV in the low-risk group was 3.4 percent and 14.2 percent for both the intermediate and high-risk groups. “The overall positive predictive value is low,” Greenburg said.

The overall cancer detection rate was 0.32 percent. The cancer detection rate in the low, intermediate and high-risk groups was 0.16 percent, 0.81 percent and 0.88 percent, respectively. 

In total, 72 women had 85 procedures. Based on the Connecticut Medicare reimbursement rates, the costs of ultrasound screening in this study population was approximately $165,000, or $55,000 per breast cancer because three cancers were detected, Greenberg explained.

“Strategies and methods are needed to decrease the number of benign lesions recommended for short-term follow-up and biopsy,” she concluded.