2022
DOI: 10.3389/fphys.2022.946099
|View full text |Cite
|
Sign up to set email alerts
|

LAD-GCN: Automatic diagnostic framework for quantitative estimation of growth patterns during clinical evaluation of lung adenocarcinoma

Abstract: Quantitative estimation of growth patterns is important for diagnosis of lung adenocarcinoma and prediction of prognosis. However, the growth patterns of lung adenocarcinoma tissue are very dependent on the spatial organization of cells. Deep learning for lung tumor histopathological image analysis often uses convolutional neural networks to automatically extract features, ignoring this spatial relationship. In this paper, a novel fully automated framework is proposed for growth pattern evaluation in lung aden… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
3
0

Year Published

2023
2023
2025
2025

Publication Types

Select...
5

Relationship

0
5

Authors

Journals

citations
Cited by 5 publications
(3 citation statements)
references
References 27 publications
0
3
0
Order By: Relevance
“…The best classification accuracy was 89.24%, while, in the study of Di-Palma et al, the histological classification of the known five patterns of lung ADC resulted in a classification accuracy of 94.51% [65]. Xiao et al created a novel framework combining CNNs and graph convolutional networks for quantitative estimation of histopathological growth patterns in lung ADC slides [66]. Another lung ADC subtyping problem was performed by Sheikh et al achieving a high accuracy rate of 0.946 and outperforming the state-of-the-art models [67].…”
Section: Lung Adc Predominant Architectural Patterns Classificationmentioning
confidence: 99%
See 1 more Smart Citation
“…The best classification accuracy was 89.24%, while, in the study of Di-Palma et al, the histological classification of the known five patterns of lung ADC resulted in a classification accuracy of 94.51% [65]. Xiao et al created a novel framework combining CNNs and graph convolutional networks for quantitative estimation of histopathological growth patterns in lung ADC slides [66]. Another lung ADC subtyping problem was performed by Sheikh et al achieving a high accuracy rate of 0.946 and outperforming the state-of-the-art models [67].…”
Section: Lung Adc Predominant Architectural Patterns Classificationmentioning
confidence: 99%
“…Graph CNNs have been used to identify regions or cell structural features that are highly associated with the class label. In this category, three approaches have been proposed, where Graph-based modules are combined with AlexNet [54], VGG16 [66], and ResNet. [37].…”
Section: Deep Learning Approachesmentioning
confidence: 99%
“…The correlation between the predominant pattern and disease outcome was not analysed. Xiaoet al [24] proposed to combine GNN and CNN modules to predict the LUAD growth patterns. The Graph convolutional networks (GCN) model was trained using the nuclear features, while the CNN model was trained using the whole image.…”
Section: Related Workmentioning
confidence: 99%