Clinical vs automated BI-RADS density measures

By Naveed Saleh, MD, MS, for MDLinx
Published May 15, 2018

Key Takeaways

Automated and clinical Breast Imaging Reporting and Data System (BI-RADS) density measures similarly predict risk of interval and screen-detected cancer, according to a new study published in the Annals of Internal Medicine.

Currently, 30 states mandate that women be apprised of their breast density. BI-RADS breast density categories are the standard when reporting breast density; it can be determined either clinically by the radiologist or by automated measures using computer algorithms. By nature, automated BI-RADS is more objective than clinical BI-RADS.

“Concern has been raised about using clinical BI-RADS breast density for prevention strategies, calling into question the subjectivity and reproducibility of the measure for individual women,” wrote the authors, led by Karla Kerlikowske, MD, University of California, San Francisco, CA.

Radiologists classify clinical BI-RADS breast density at the time of mammography using the following categories:

  1. Almost entirely fatty
  2. Scattered fibroglandular densities
  3. Heterogeneously dense
  4. Extremely dense

In this study, the researchers drew participants from two case-control studies embedded within large prospective breast imaging cohorts. The researchers compared 4,400 matched controls with 1,609 patients with screen-detected invasive cancer and 351 with interval invasive cancer.

The researchers obtained three measurements:

  • Automated and clinical BI-RADS density at two time points between September 2006 and October 2014
  • Interval and screen-detected breast cancer risk
  • Mammography sensitivity

Volpara, version 1.5.3, was used for automated density measures.

“We used conditional logistic regression to assess the association of clinical and automated BI-RADS density with screen-detected and interval cancer,” wrote the authors.

They added: “Associations of interval and screen-detected cancer with clinical BI-RADS density were similar to those with automated BI-RADS density, regardless of whether density was measured more than 6 months to less than 2 years or 2 to 5 years before diagnosis.”

In addition to finding that automated and clinical BI-RADS density measures exhibited comparable discriminatory accuracies, the researchers found that these accuracies were higher for interval than for screen-detected cancers.

The team also found that mammography sensitivities were similar between automated and clinical BI-RADS categories.

One strength of this study is that it defined invasive cancer as that occurring within 12 months of a negative screening, which is shorter and more relevant than definitions used in other studies.

“Because automated BI-RADS breast density is more reproducible than clinical density and is being used increasingly in the clinical setting,” conclude the researchers, “our results suggest that automated density measures may be used to predict risk and help identify women most in need of supplemental screening.”

In an accompanying editorial, Joann G. Elmore, MD, MPH, David Geffen School of Medicine at University of California, Los Angeles, CA, and Jill Wruble, DO, Yale School of Medicine, New Haven, CT, reflect on the potential of automated breast density measurement.

“Automated measurement technology is new, but is it ‘improved?’ That remains to be determined. In the meantime, the breast imaging community's experience with the parallel technology of computer-aided detection (CAD) may be a cautionary tale.”

They added: “…Automated density measurement has the potential to improve reproducibility and workflow efficiency. However, we are in an era of ‘choosing wisely’ and seeking value in health care. Therefore, we must be cautious before implementing and paying for medical technology.”

To read more about this study, click here

To read more about the accompanying editorial, click here.

Share with emailShare to FacebookShare to LinkedInShare to Twitter