Diagnostic architectural and dynamic features at breast MR imaging: multicenter study

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Breast magnetic resonance (MR) imaging is a vital tool in breast cancer detection and diagnosis.
• Understanding the predictive value of various imaging features is essential for improving diagnostic accuracy.

Purpose of the Study:

• To prospectively evaluate the prevalence and predictive value of three-dimensional (3D) and dynamic breast MR imaging features.
• To assess the diagnostic performance of individual and combined imaging features in predictive models for breast cancer.

Main Methods:

• Prospective analysis of 995 lesions in 854 women using 3D and dynamic 2D breast MR imaging.
• Readers assessed enhancement shape, distribution, border architecture, intensity, and kinetic curve patterns.
• Multivariate models were constructed using architectural and dynamic features, with diagnostic accuracy measured by areas under the receiver operating characteristic curve (Az).

Main Results:

• The absence of enhancement showed an 88% negative predictive value for cancer.
• Focal mass margins (Az = 0.76) and signal intensity (Az = 0.70) were highly predictive.
• Multivariate models incorporating multiple features achieved a high diagnostic accuracy (Az = 0.880).

Conclusions:

• Architectural and dynamic features are critical for accurate breast MR imaging interpretation.
• Multivariate models integrating these features demonstrate superior diagnostic accuracy over qualitative dynamic enhancement assessment.