Zheng Feng Lu

Clinical Diagnostic Physicist
Biological Sciences Division
The University of Chicago

zlu@radiology.bsd.uchicago.edu

Dept. and Web

Radiology

Project Description

Automated Diagnostic Ultrasound Performance Analysis

Diagnostic ultrasound is susceptible to equipment-related deficiencies, including array transducer element dropouts, beam-forming circuit board malfunctions, and unintentional changes in imaging protocol presets. The effects of these issues on clinical images can be subtle and difficult for users to recognize, making it essential to have a robust quality control (QC) program so the deficiencies can be detected early, before patient care is compromised. Annual ultrasound QC testing is required for ACR ultrasound accreditation; however, more frequent QC testing is desirable.

A major barrier to routine ultrasound QC is the lack of accessible and automated QC image analysis tools. This project aims to develop an automated phantom image analysis tool using an ultrasound phantom composed of randomly distributed spherical voids within a tissue-mimicking background. Preliminary studies demonstrate that the lesion signal-to-noise ratio (LSNR) and the number of detected voids are promising indicators of ultrasound system performance.

The proposed project will focus on characterizing sources of measurement variability and evaluating the sensitivity of the automated method for detecting common equipment deficiencies such as array transducer element dropouts. The ultimate goal is to enable routine, user-friendly QC using an affordable phantom that ultrasound operators can routinely apply to assess and maintain system performance over time. This project is part of the AAPM TG400* effort.

AAPM TG400: https://www.aapm.org/org/structure/default.asp?committee_code=TG400

Requirement

The ideal candidate will be a motivated student with a background in physics, biomedical engineering, or computer science, and an interest in learning automated diagnostic image analysis.