Deep Multi-Magnification Networks for multi-class breast cancer image segmentation

•Multi-Magnification Network segments multiple tissue subtypes on pathology images.•Features from both high magnifications and low magnifications are fully utilized.•Partial annotation approach is proposed to reduce labeling burdens for pathologists.•Sharp boundaries delineate tissue subtypes and ou...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Computerized medical imaging and graphics Ročník 88; s. 101866
Hlavní autori: Ho, David Joon, Yarlagadda, Dig V.K., D’Alfonso, Timothy M., Hanna, Matthew G., Grabenstetter, Anne, Ntiamoah, Peter, Brogi, Edi, Tan, Lee K., Fuchs, Thomas J.
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: United States Elsevier Ltd 01.03.2021
Elsevier Science Ltd
Predmet:
Annotations
Breast
Breast cancer
Breast carcinoma
Breast Neoplasms - diagnostic imaging
Coders
Computational pathology
Deep Multi-Magnification Network
Female
Humans
Image processing
Image segmentation
Internal Medicine
Multi-class image segmentation
Neural Networks, Computer
Other
Partial annotation
Deep Multi-Magnification Network
Breast cancer
Multi-class image segmentation
Computational pathology
Partial annotation
ISSN:0895-6111, 1879-0771, 1879-0771
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:•Multi-Magnification Network segments multiple tissue subtypes on pathology images.•Features from both high magnifications and low magnifications are fully utilized.•Partial annotation approach is proposed to reduce labeling burdens for pathologists.•Sharp boundaries delineate tissue subtypes and outperform the state-of-the-art. Pathologic analysis of surgical excision specimens for breast carcinoma is important to evaluate the completeness of surgical excision and has implications for future treatment. This analysis is performed manually by pathologists reviewing histologic slides prepared from formalin-fixed tissue. In this paper, we present Deep Multi-Magnification Network trained by partial annotation for automated multi-class tissue segmentation by a set of patches from multiple magnifications in digitized whole slide images. Our proposed architecture with multi-encoder, multi-decoder, and multi-concatenation outperforms other single and multi-magnification-based architectures by achieving the highest mean intersection-over-union, and can be used to facilitate pathologists’ assessments of breast cancer.
Bibliografia:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
David Joon Ho: Methodology, Software, Validation, Formal analysis, Investigation, Visualization, Writing - Original Draft, Dig V. K. Yarlagadda: Software, Timothy M. D’Alfonso: Validation, Data Curation, Investigation, Writing - Review & Editing, Matthew G. Hanna: Data Curation, Anne Grabenstetter: Data Curation, Peter Ntiamoah: Conceptualization, Edi Brogi: Conceptualization, Lee K. Tan: Conceptualization, Investigation, Supervision, Thomas J. Fuchs: Conceptualization, Investigation, Supervision, Writing - Review & Editing
ISSN:0895-6111
1879-0771
1879-0771
DOI:10.1016/j.compmedimag.2021.101866