# Development of a 1 × 512 Ring Transducer Array-Based 3D Ultrasound Imaging System for Accurate Breast Lesion Detection: Phantom and Preliminary Clinical Feasibility Study

## TL;DR

This paper introduces a high-resolution 3D ultrasound system using a ring transducer array and adaptive beamforming to detect breast microlesions with high accuracy.

## Contribution

A novel adaptive beamforming algorithm with optimal sound speed estimation is introduced to enhance 3D ultrasound imaging resolution and accuracy.

## Key findings

- The system achieves a spatial resolution of 0.5 mm and linear measurement accuracy of 2.1%.
- Preliminary clinical results show a Dice similarity coefficient of 0.87 when compared to MRI imaging.

## Abstract

The work presents an algorithm for early detection of breast microlesions using a high resolution three-dimensional ultrasound imaging system. The system employs a 1 × 512 ring transducer array and a triaxial displacement platform with an accuracy of 0.1 mm, achieving high-density acquisition of three-dimensional volumetric data through fixed-step scanning. To improve imaging quality, an adaptive beamforming algorithm incorporating optimal sound speed estimation is proposed, effectively compensating for phase distortion caused by sound speed heterogeneity within tissues and improving spatial coherence and imaging resolution. The three-dimensional volumetric data is visualized using volume rendering to achieve high-fidelity three-dimensional ultrasound image reconstruction. The in vitro experimental results demonstrate that the proposed algorithm improves the system’s spatial resolution to 0.5 mm, with a linear measurement accuracy of 2.1%. A preliminary clinical feasibility case study comparing breast image reconstruction with MRI imaging results shows a Dice similarity coefficient of 0.87 for the lesion region, high anatomical structure reconstruction accuracy, and good spatial consistency. These results demonstrate preliminary clinical feasibility for early detection of breast microlesions.