Efficient and Highly Accurate Diagnosis of Malignant Hematological Diseases Based on Whole-Slide Images Using Deep Learning
Hematological disorders
Feature (linguistics)
DOI:
10.3389/fonc.2022.879308
Publication Date:
2022-06-10T06:17:07Z
AUTHORS (8)
ABSTRACT
Hematopoietic disorders are serious diseases that threaten human health, and the diagnosis of these is essential for treatment. However, traditional methods rely on manual operation, which time consuming laborious, examining entire slide challenging. In this study, we developed a weakly supervised deep learning method diagnosing malignant hematological requiring only slide-level labels. The improves efficiency by converting whole-slide image (WSI) patches into low-dimensional feature representations. Then patch-level features each WSI aggregated representations an attention-based network. model provides final diagnostic predictions based By applying proposed to our collection bone marrow WSIs at different magnifications, found area under receiver operating characteristic curve 0.966 independent test set can be obtained 10× magnification. Moreover, performance microscopy images achieve average accuracy 94.2% two publicly available datasets. conclusion, have novel fast accurate in scenarios disorders.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (45)
CITATIONS (6)
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....