EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Avoiding Tissue Overlap in 2D Images: Single-Slice DBT Classification Using Convolutional Neural Networks

João Mendes iD, Nuno Matela iD, Nuno Garcia

Tomography · 2023

DOI: 10.3390/tomography9010032 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Tomography, 2023. Verfasst von João Mendes und 2 weiteren Autor:innen. Beteiligte Einrichtung: Faculdade de Ciências, Instituto de Biofísica e Engenharia Biomédica, Universidade de Lisboa, 1749-016 Lisboa, Portugal. Hauptthemen laut MEDLINE: Mammography, Breast Neoplasms. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fundação para a Ciência e Tecnologia–Portugal, Bolsa de Investigação para Doutoramento.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Breast cancer was the most diagnosed cancer around the world in 2020. Screening programs, based on mammography, aim to achieve early diagnosis which is of extreme importance when it comes to cancer. There are several flaws associated with mammography, with one of the most important being tissue overlapping that can result in both lesion masking and fake-lesion appearance. To overcome this, digital breast tomosynthesis takes images (slices) at different angles that are later reconstructed into a 3D image. Having in mind that the slices are planar images where tissue overlapping does not occur, the goal of the work done here was to develop a deep learning model that could, based on the said slices, classify lesions as benign or malignant. The developed model was based on the work done by Muduli et. al, with a slight change in the fully connected layers and in the regularization done. In total, 77 DBT volumes—39 benign and 38 malignant—were available. From each volume, nine slices were taken, one where the lesion was most visible and four above/below. To increase the quantity and the variability of the data, common data augmentation techniques (rotation, translation, mirroring) were applied to the original images three times. Therefore, 2772 images were used for training. Data augmentation techniques were then applied two more times—one set used for validation and one set used for testing. Our model achieved, on the testing set, an accuracy of 93.2% while the values of sensitivity, specificity, precision, F1-score, and Cohen’s kappa were 92%, 94%, 94%, 94%, and 0.86, respectively. Given these results, the work done here suggests that the use of single-slice DBT can compare to state-of-the-art studies and gives a hint that with more data, better augmentation techniques and the use of transfer learning might overcome the use of mammograms in this type of studies.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

MammographyBreast NeoplasmsHumansFemaleNeural Networks, ComputerImaging, Three-DimensionalRadiographic Image Interpretation, Computer-Assisted

Stichwörter der Autor:innen

CNNDBTcarcinoma of the breastdeep learning

Autor:innen und Institutionen

  • Faculdade de Ciências, Instituto de Biofísica e Engenharia Biomédica, Universidade de Lisboa, 1749-016 Lisboa, Portugal.
  • Faculdade de Ciências, LASIGE, Universidade de Lisboa, 1749-016 Lisboa, Portugal.

ORCID-Profile: João Mendes, Nuno Matela

Förderung

  • Fundação para a Ciência e Tecnologia–Portugal
  • Bolsa de Investigação para Doutoramento

Zitieren

APA

Mendes, J., Matela, N., & Garcia, N. (2023). Avoiding Tissue Overlap in 2D Images: Single-Slice DBT Classification Using Convolutional Neural Networks. Tomography. https://doi.org/10.3390/tomography9010032

Vancouver

Mendes J, Matela N, Garcia N. Avoiding Tissue Overlap in 2D Images: Single-Slice DBT Classification Using Convolutional Neural Networks. Tomography. 2023. doi:10.3390/tomography9010032

BibTeX
@article{Mendes2023,
  author = {Mendes, João and Matela, Nuno and Garcia, Nuno},
  title = {Avoiding Tissue Overlap in 2D Images: Single-Slice DBT Classification Using Convolutional Neural Networks},
  journal = {Tomography},
  year = {2023},
  doi = {10.3390/tomography9010032}
}
Kostenlos starten →