Evidence Hub

Performance Evaluation of a Dual Headed Molecular Breast Imaging System

K. Erlandsson, A. Cherlin, M. Exley, N. Atkins, I. Baistow, B. Hutton, K. Thielemans

2025 IEEE Nuclear Science Symposium (NSS), Medical Imaging Conference (MIC) and Room Temperature Semiconductor Detector Conference (RTSD), 2025

Abstract

A prototype dual-headed Molecular Breast Imaging (MBI) system has been developed at Kromek and UCL. Each head consists of an array of CZT detectors with sub-pixel resolution and depth-ofinteraction (DOI) capability equipped with multi-pinhole collimators, which allow multiplexing (MX). The system is stationary and provides limited-angle tomographic information. We have evaluated the performance of the prototype with phantom measurements in terms of lesion detection and dose reduction factor (DRF). We scanned a ${ }^{99 ~\mathrm{m}}$ Tc-filled phantom containing 5-9 mm diameter spherical inserts and a uniform background phantom separately. We generated data sets with different count statistics for target-to-background ratios (TBRs) of 10 and 20 using bootstrapping. Images were reconstructed with OSEM, including DOI-based de-MX. We estimated the lesion detection capability using contrast-to-noise ratio (CNR) and the scan statistic (SS) method. The analyses were done w and w/o noise reduction using the non-local means (NLM) filter. The filter had a significant effect on the CNR values but not on the SS results. The CNR results suggest that, with a DRF of $8,7-\text{mm}$ lesions were detectable with TBR=10 w or w/o NLM filter while, for TBR $=20,6-\text{mm}$ lesions were detectable w/o and $5-\text{mm}$ lesions w NLM filter. The SS evaluation resulted in AUC values of $0.86,0.75$ and 0.55 for DRFs of 1,2 and 4 with TBR $=10$, while with $\text{TBR}=20$, all AUC values were in…

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free