Researchers using automated breast density measurements have found that women with mammographically dense breast tissue have higher recall and biopsy rates and increased odds of screen-detected and interval breast cancer, according to a large new study from Norway published online in the journal Radiology. The study supports automated measurements as a future standard to ensure objective breast density classification for breast cancer screening, the researchers say.
Previous studies have shown that women with mammographically dense breasts face a higher risk of breast cancer and missed cancers than those with nondense breasts, partly because the superimposition of dense breast tissue on mammograms leads to a masking effect, causing some cancers to go undetected. However, the majority of those studies relied on subjective density assessments, most commonly the radiologist's subjective interpretation using the ACR's BI-RADS. This approach introduces potential mammogram reader variability into density categorization, says the study's principal investigator, Solveig Hofvind, PhD, from the Cancer Registry of Norway in Oslo.
For the new study, Hofvind and colleagues used automated software to help classify mammographic density in 107,949 women ages 50 to 69 from BreastScreen Norway, a national program that offers women screening every two years. The researchers looked at a total of 307,015 digital screening examinations that took place from 2007 to 2015.
The automated software classified breasts as dense in 28% of the screening examinations. Rates of screen-detected cancer were 6.7 per 1,000 examinations for women with dense breast tissue and 5.5 for women with nondense breasts. Interval breast cancer, or breast cancer detected between screenings—usually by palpation—was 2.8 per 1,000 in the dense breast tissue group and 1.2 for women with nondense tissue.
The recall rate was 3.6% for women with dense breasts, compared with 2.7% for women with nondense breasts. The biopsy rate of 1.4% in the dense breast group was higher than the 1.1% rate for women in the nondense category.
"The odds of screen-detected and interval breast cancer were substantially higher for women with mammographically dense versus fatty breasts in BreastScreen Norway," Hofvind says. "We also found substantially higher rates of recalls and biopsies among women with mammographically dense breast tissue."
Mammographic sensitivity for detecting breast cancer was only 71% for women with dense breasts, compared to 82% for women with nondense breasts. Cancers detected at screening were more advanced among women in the dense breast tissue group. Average tumor diameter for screen-detected cancers was 16.6 mm in the dense breasts group, compared with 15.1 mm for the nondense group. Lymph node-positive disease was found in 24% of women with dense breasts, compared with 18% for women with nondense breasts.
The results add another piece to the puzzle of screening protocols for women with dense breasts. Hofvind notes that more than one-half of states in the United States have enacted breast density legislation, with some states requiring that women be informed about their breast density or that additional imaging could be beneficial. However, supplemental screening for women with dense breasts currently isn't recommended by any of the major health care societies or organizations, and more research is needed before widespread changes are made.
"We need well-planned and high-quality studies that can give evidence about the cost-effectiveness of more frequent screening, other screening tools such as tomosynthesis, and/or the use of additional screening tools like MRI and ultrasound for women with dense breasts," she says. "Further, we need studies on automated measurement tools for mammographic density to ensure their validity."
In the meantime, Hofvind says, the findings will help inform how automated volumetric density categorization will change population-based screening performance and outcomes through a more objective breast density measurement method.
In an accompanying editorial, "Machine Detection of High Breast Density: Worse Outcomes for Our Patients," Liane E. Philpotts, MD, FACR, from Yale School of Medicine writes, "This study is important for two main reasons: (1) the validation that automated means of density classification can correctly identify a percentage of women with dense tissue, and (2) women with dense tissue have poorer performance of screening mammography. It lends support to the density notification movement along with redoubling efforts towards optimizing supplemental screening methods."
She adds that automated volumetric determination could better standardize density measurements and provide greater confidence that women with dense breast tissue can consistently be identified.
"Breast density is here to stay, and it is in everyone's best interest to embrace understanding and optimization of breast imaging practice to best address the needs of women with dense tissue," Philpotts concludes.— Source: RSNA