Breast Tumor Classification Based on the Scatterer Spacing Estimated with Matching Pursuit
PhysioPlus Open · 2025
DOI: 10.64473/ppo.02.02.032 ↗Kurz gefasst
Fachartikel in der Zeitschrift PhysioPlus Open, 2025. Verfasst von Xiuhua Zeng und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Background
Mean Scatterer Spacing (MSS) estimated from ultrasound radiofrequency (RF) echo signals provides valuable microstructural information for breast tumor characterization. However, conventional MSS estimation methods suffer from limited noise robustness, low objectivity, and strong dependence on signal assumptions.
Objective
This study aims to develop an objective and noise-robust approach for estimating MSS to improve the differentiation between benign and malignant breast tumors.
Methods
A Matching Pursuit (MP)–based method is proposed to decompose ultrasound RF echo signals using a redundant Gabor atom dictionary. The Second-Order Difference (SOD) of the Reconstructed Signal-to-Residual Ratio (RSRR) is employed to automatically identify Gabor atoms corresponding to coherent scattering components while suppressing diffuse scattering and noise. Inter-Scatterer Spacing (ISS) and MSS are then estimated from the spatial locations of the selected atoms. The method was evaluated on the OASBUD database, including 48 benign and 52 malignant breast lesion regions of interest.
Results
The estimated mean ISS for benign lesions was 0.81 ± 0.10 mm, which was significantly higher than that of malignant lesions (0.75 ± 0.12 mm). One-way ANOVA analysis confirmed a statistically significant difference between the two groups (p = 0.024).
Conclusion
The proposed MP-based MSS estimation approach enhances the accuracy, stability, and objectivity of scatterer spacing estimation from ultrasound RF signals. By effectively isolating coherent tissue microstructures, the method provides a reliable quantitative tool for distinguishing benign and malignant breast tumors and shows strong potential for clinical ultrasound tissue characterization.
Zitieren
APA
Zeng, X., Xia, W., Wang, Z., & Yang, T. (2025). Breast Tumor Classification Based on the Scatterer Spacing Estimated with Matching Pursuit. PhysioPlus Open. https://doi.org/10.64473/ppo.02.02.032
Vancouver
Zeng X, Xia W, Wang Z, Yang T. Breast Tumor Classification Based on the Scatterer Spacing Estimated with Matching Pursuit. PhysioPlus Open. 2025. doi:10.64473/ppo.02.02.032
BibTeX
@article{Zeng2025,
author = {Zeng, Xiuhua and Xia, Wenzhi and Wang, Zhengwei and Yang, Taoxiang},
title = {Breast Tumor Classification Based on the Scatterer Spacing Estimated with Matching Pursuit},
journal = {PhysioPlus Open},
year = {2025},
doi = {10.64473/ppo.02.02.032}
}
EUVIMED