Ultrasound viscosity imaging empowers BI-RADS: toward precise breast lesion diagnosis and analysis of HER2 status

TL;DR

This study shows that ultrasound viscosity imaging improves breast cancer diagnosis and can help assess HER2 status noninvasively.

Contribution

The study introduces ultrasound viscosity imaging as a novel tool to enhance breast lesion diagnosis and HER2 status analysis.

Findings

V2.max, a viscosity parameter, improved differentiation between benign and malignant breast lesions when combined with BI-RADS.

The combined model achieved an AUC of 0.96, significantly higher than BI-RADS alone.

Four viscosity parameters correlated with HER2 positivity, offering potential noninvasive biomarkers.

Abstract

Breast cancer remains a major challenge in women’s health globally. Early screening and personalized treatment can improve outcomes. This study aimed to evaluate ultrasound viscosity imaging (UVI) for distinguishing benign from malignant breast lesions and noninvasively assessing human epidermal growth factor receptor 2 (HER2) status.

We conducted a retrospective analysis of 274 breast lesions, randomly divided into a derivation cohort and a validation cohort (VC) at a 7:3 ratio. Breast Imaging Reporting and Data System (BI-RADS) scores and UVI parameters were collected, with histopathology as the reference standard. The Boruta algorithm was used to identify the optimal viscous parameter (VP). A logistic regression model assessed the diagnostic performance of BI-RADS alone and in combination with VP. Associations between viscous parameters (VPs) and HER2 status were also examined.

Among 40 VPs, V2.max (maximum viscosity from the Voigt model within a perilesional 2-mm rim) was identified as the optimal marker. When combined with BI-RADS, V2.max enhanced the differentiation between benign and malignant lesions (p<0.001), increasing the area under the curve (AUC) from 0.91 (95% CI: 0.87-0.95) to 0.96 (95% CI: 0.94-0.98). The combined model also demonstrated superior calibration, which was revalidated in the VC. Subgroup analyses confirmed its effectiveness in younger patients and those with larger lesions. Furthermore, we identified four Voigt-model-derived VPs, including V2.max, that correlated with HER2 positivity, and explored their potential histological basis.

UVI-derived VPs enhance BI-RADS diagnostic performance for breast lesions and are associated with HER2 status.

Figures