Direct Segmentation of Mammography and Tomosynthesis Sinograms for Lesion Localization
Tomography · 03.03.2026 · 12 (3) · 34
DOI: 10.3390/tomography12030034 ↗Kurz gefasst
Fachartikel in der Zeitschrift Tomography, online erschienen am 03.03.2026. Verfasst von Estefanía Ruíz Muñoz und 5 weiteren Autor:innen. Beteiligte Einrichtung: Instituto Nacional de Astrofísica, Óptica y Electrónica, Sta. María Tonantzintla, Puebla 72840, Mexico. Hauptthemen laut MEDLINE: Mammography, Breast Neoplasms, Radiographic Image Interpretation, Computer-Assisted. Die Arbeit nennt 23 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Secretaría de Ciencia, Humanidades, Tecnología e Innovación.
Abstract (Originalsprache)
Background
The Detection and localization of breast lesions remain challenging in mammography and digital breast tomosynthesis (DBT) due to tissue overlap and information loss during volumetric reconstruction. Sinograms preserve the full angular projection data acquired during scanning, enabling analysis of tissue structure without reconstruction.
Methods
This study proposes a direct segmentation approach for mammography and DBT sinograms using a U-Net architecture. Experiments were conducted on 1082 annotated mammography mass images from the CBIS-DDSM dataset (521 benign, 561 malignant) and 272 annotated DBT images from the Breast Cancer Screening DBT dataset (136 benign, 136 malignant). Dataset splitting was performed at the patient level to prevent data leakage, and all reported quantitative results correspond to the independent test set, with the validation set used solely for model selection and early stopping. Three input configurations were evaluated: mammography sinograms, DBT sinograms, and a combined model.
Results
The mammography model achieved the highest performance (Dice: 0.94 training, 0.90 test), outperforming DBT alone (0.77 training, 0.70 test). Multimodal fusion improved DBT results (Dice: 0.84 test). Centroid analysis showed 99.11% correspondence with reference annotations (average distance: 2.83 pixels), and partial back-projection reconstructions confirmed anatomical consistency. Compared with YOLOv5x, the proposed approach provided superior lesion localization, particularly for small or multiple lesions.
Conclusions
Direct sinogram segmentation is an efficient, clinically viable strategy for breast lesion detection and localization.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Instituto Nacional de Astrofísica, Óptica y Electrónica, Sta. María Tonantzintla, Puebla 72840, Mexico.
- Secretaria de Ciencia, Humanidades, Tecnología e Innovación (SECIHTI), Mexico City 03940, Mexico.
ORCID-Profile: Estefanía Ruíz Muñoz, Leopoldo Altamirano Robles, Raquel Díaz Hernández, Kelsey Alejandra Ramírez Gutiérrez, Saúl Zapotecas-Martínez, José de Jesús Velázquez Arreola
Förderung
- Secretaría de Ciencia, Humanidades, Tecnología e Innovación
Zitieren
APA
Muñoz, E. R., Robles, L. A., Hernández, R. D., Gutiérrez, K. A. R., Zapotecas-Martínez, S., & Velázquez Arreola, J. J. (2026). Direct Segmentation of Mammography and Tomosynthesis Sinograms for Lesion Localization. Tomography, 12(3), 34. https://doi.org/10.3390/tomography12030034
Vancouver
Muñoz ER, Robles LA, Hernández RD, Gutiérrez KAR, Zapotecas-Martínez S, Velázquez Arreola JJ. Direct Segmentation of Mammography and Tomosynthesis Sinograms for Lesion Localization. Tomography. 2026;12(3):34. doi:10.3390/tomography12030034
BibTeX
@article{Muoz2026,
author = {Muñoz, Estefanía Ruíz and Robles, Leopoldo Altamirano and Hernández, Raquel Díaz and Gutiérrez, Kelsey Alejandra Ramírez and Zapotecas-Martínez, Saúl and Velázquez Arreola, José de Jesús},
title = {Direct Segmentation of Mammography and Tomosynthesis Sinograms for Lesion Localization},
journal = {Tomography},
year = {2026},
volume = {12},
number = {3},
pages = {34},
doi = {10.3390/tomography12030034}
}
EUVIMED