• AI
  • Molecular Imaging
  • CT
  • X-Ray
  • Ultrasound
  • MRI
  • Facility Management
  • Mammography

Ultrasound BI-RADS brings consistency to lesion classification

Article

The Breast Imaging Reporting and Data System for ultrasound can reliably characterize breast tumors, according to a study published in the April issue of the American Journal of Roentgenology.

The Breast Imaging Reporting and Data System for ultrasound can reliably characterize breast tumors, according to a study published in the April issue of the American Journal of Roentgenology.

Clinical data validating ultrasound's ability to differentiate benign from malignant breast lesions keep accumulating. But the modality's value for breast cancer screening remains controversial because previous attempts to devise classification systems for the modality did not provide consistent results.

Dr. Andrea S. Hong and colleagues in the radiology department at Duke University weighed the BI-RADS nomenclature that describes tumors according to their shape, orientation, margins, and several ultrasound-specific features. They found these ultrasound-based descriptors, developed by the American College of Radiology in 2003, were significantly different in malignant and benign masses.

The investigators' trial included sonograms of 403 breast lesions - 202 palpable and 201 unpalpable - from 369 patients. Each case was evaluated blindly by one of three dedicated breast radiologists, who correlated descriptions against mammography and biopsy. Histology confirmed 35% of the masses as malignant.

The ACR's ultrasound BI-RADS lexicon identified masses with spiculated margins, irregular shapes, and nonparallel orientation as having high positive predictive values (86%, 62%, and 69%, respectively). Lesions described as having high negative predictive values using the same scheme included circumscribed margins (90%), parallel orientation (78%), and oval shapes (84%). The differences in ultrasound BI-RADS features that helped identify benign from malignant masses were statistically significant (p

BI-RADS ultrasound-specific markers showing high predictive value for benign lesions included posterior acoustic enhancement (67%) and no posterior acoustic features (79%). Posterior acoustic shadowing, on the other hand, showed the highest predictive value for malignancy (52%).

Although mammography remains the imaging standard for population screening, a growing number of clinical studies agree that more powerful, widely available ultrasound technology has improved diagnosis and management of breast tumors.

As in mammography, sonographic BI-RADS descriptors showing spiculated margins suggest the occurrence of infiltrating lesions, while irregular shapes can indicate inconsistent growth and advancement of a lesion's edge. On the other hand, most lesions showing circumscribed margins or oval shapes indicating smooth uniform growth can be associated with benign tumors, the Duke researchers said.

The study had some limitations. The observers were generally not blinded to mammograms when they evaluated ultrasound results. And though blinded to biopsy, they were aware their assessments did not directly affect patient care, which might have influenced their analysis. However, long-term follow-up studies with larger population and histology samples using Duke's guidelines may provide more conclusive results, they said.

The ACR BI-RADS lexicon provides standardized terminology that facilitates accurate and consistent breast sonography and mammography reporting, investigators said.

For more information from the Diagnostic Imaging archives:

Research supports ultrasound for breast cancer screening

Breast ultrasound wins in younger women

MRI, ultrasound find larger role in staging breast cancer

Ultrasound finds favor among breast clinicians

Related Videos
Making the Case for Intravascular Ultrasound Use in Peripheral Vascular Interventions
Nina Kottler, MD, MS
Radiology Challenges with Breast Cancer Screening in Women with Breast Implants
Related Content
© 2024 MJH Life Sciences

All rights reserved.