Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability
New Biotechnology · 2023-12 · 78 · 52-67
DOI: 10.1016/j.nbt.2023.09.008 ↗Kurz gefasst
Fachartikel in der Zeitschrift New Biotechnology, 2023-12. Verfasst von Matej Gallo und 4 weiteren Autor:innen. Beteiligte Einrichtung: Faculty of Informatics, Masaryk University, Botanická 68a, 602 00 Brno, Czech Republic. Electronic address: 422328@mail.muni.cz. Hauptthemen laut MEDLINE: Neural Networks, Computer, Prostatic Neoplasms. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 91 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministerstvo Školství, Mládeže a Tělovýchovy, Ministerstvo Zdravotnictví Ceské Republiky.
Abstract (Originalsprache)
Diagnostic histopathology faces increasing demands due to aging populations and expanding healthcare programs. Semi-automated diagnostic systems employing deep learning methods are one approach to alleviate this pressure. The learning models for histopathology are inherently complex and opaque from the user's perspective. Hence different methods have been developed to interpret their behavior. However, relatively limited attention has been devoted to the connection between interpretation methods and the knowledge of experienced pathologists. The main contribution of this paper is a method for comparing morphological patterns used by expert pathologists to detect cancer with the patterns identified as important for inference of learning models. Given the patch-based nature of processing large-scale histopathological imaging, we have been able to show statistically that the VGG16 model could utilize all the structures that are observable by the pathologist, given the patch size and scan resolution. The results show that the neural network approach to recognizing prostatic cancer is similar to that of a pathologist at medium optical resolution. The saliency maps identified several prevailing histomorphological features characterizing carcinoma, e.g., single-layered epithelium, small lumina, and hyperchromatic nuclei with halo. A convincing finding was the recognition of their mimickers in non-neoplastic tissue. The method can also identify differences, i.e., standard patterns not used by the learning models and new patterns not yet used by pathologists. Saliency maps provide added value for automated digital pathology to analyze and fine-tune deep learning systems and improve trust in computer-based decisions.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Faculty of Informatics, Masaryk University, Botanická 68a, 602 00 Brno, Czech Republic. Electronic address: 422328@mail.muni.cz.
- Faculty of Informatics, Masaryk University, Botanická 68a, 602 00 Brno, Czech Republic.
- Department of Pathology, Masaryk Memorial Cancer Institute, Žlutý kopec 7, 656 53 Brno, Czech Republic.
- Institute of Computer Science, Masaryk University, Šumavská 416/15, 602 00 Brno, Czech Republic.
Förderung
- Ministerstvo Školství, Mládeže a Tělovýchovy
- Ministerstvo Zdravotnictví Ceské Republiky
Zitieren
APA
Gallo, M., Krajňanský, V., Nenutil, R., Holub, P., & Brázdil, T. (2023). Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability. New Biotechnology, 78, 52-67. https://doi.org/10.1016/j.nbt.2023.09.008
Vancouver
Gallo M, Krajňanský V, Nenutil R, Holub P, Brázdil T. Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability. New Biotechnology. 2023;78:52-67. doi:10.1016/j.nbt.2023.09.008
BibTeX
@article{Gallo2023,
author = {Gallo, Matej and Krajňanský, Vojtěch and Nenutil, Rudolf and Holub, Petr and Brázdil, Tomáš},
title = {Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability},
journal = {New Biotechnology},
year = {2023},
volume = {78},
pages = {52--67},
doi = {10.1016/j.nbt.2023.09.008}
}
EUVIMED