# Quantitative outperforms visual assessment of nonparallel orientation in ultrasound breast imaging reporting and data system

## TL;DR

This study shows that using quantitative measurements for nonparallel orientation in ultrasound improves breast cancer diagnosis accuracy compared to visual assessment.

## Contribution

The study introduces a quantitative method for nonparallel orientation in ultrasound BI-RADS, improving diagnostic performance.

## Key findings

- Quantitative angle measurement outperformed visual assessment with an AUC of 0.838 versus 0.651.
- Modified BI-RADS with quantitative angle had higher diagnostic performance (AUC = 0.922) than standard BI-RADS (AUC = 0.905).
- Quantitative orientation angle is more reproducible and objective than visual assessment.

## Abstract

The visual assessment of the “nonparallel orientation” descriptor in the ultrasound Breast Imaging Reporting and Data System (BI-RADS) is subjective and may affect both the accurate interpretation of breast masses and the overall diagnostic performance of the ultrasound BI-RADS. The objective of this study was to determine whether quantitative measurement of the nonparallel orientation descriptor improves diagnostic performance in the evaluation of breast malignancy assessment.

This prospective study, conducted at a tertiary hospital, analyzed 253 out of 6,893 patients with ultrasound BI-RADS 3-5 solid breast masses. For each mass, parallel or nonparallel orientation was assessed visually, and the orientation angle of breast mass was measured quantitatively on ultrasound image using built-in ultrasound software. Histopathological diagnosis served as the reference standard. Receiver Operating Characteristic (ROC) curve was plotted to determine the optimal cutoff value for the orientation angle in assessing breast malignancy. The diagnostic performances of the standard BI-RADS (using visual nonparallel orientation) and a modified BI-RADS (using the quantitative orientation angle) were compared for malignancy stratification.

McNemar testing demonstrated significant differences in diagnostic outcomes between visual nonparallel orientation assessment and quantitative angle measurement, as well as between the standard and modified ultrasound BI-RADS classifications (all p < 0.001). The area under the curve (AUC) for visual assessment was 0.651, compared to 0.838 for the orientation angle (p < 0.001). Incorporating orientation angle into ultrasound BI-RADS showed higher diagnostic performance (AUC = 0.922) compared to the standard ultrasound BI-RADS in this cohort (AUC = 0.905, p = 0.024).

The quantitative orientation angle is a more reproducible and objective measure. It can serve as a valuable complementary descriptor within ultrasound BI-RADS. Integrating this quantitative angle into a multi-descriptor assessment of breast masses improves the overall diagnostic performance.