Deep Learning-Based Quantification of Pulmonary Hemosiderophages in Cytology Slides

@article{Marzahl2020DeepLQ,
  title={Deep Learning-Based Quantification of Pulmonary Hemosiderophages in Cytology Slides},
  author={Christian Marzahl and Marc Aubreville and Christof A. Bertram and Jason Stayt and Anne-Katherine Jasensky and Florian Bartenschlager and Marco Fragoso-Garcia and Ann Kristin Barton and Svenja Elsemann and Samir Jabari and Jens Krauth and Prathmesh Madhu and J{\"o}rn Voigt and Jenny Hill and Robert Klopfleisch and Andreas K. Maier},
  journal={Scientific Reports},
  year={2020},
  volume={10}
}
Exercise-induced pulmonary hemorrhage (EIPH) is a common condition in sport horses with negative impact on performance. Cytology of bronchoalveolar lavage fluid by use of a scoring system is considered the most sensitive diagnostic method. Macrophages are classified depending on the degree of cytoplasmic hemosiderin content. The current gold standard is manual grading, which is however monotonous and time-consuming. We evaluated state-of-the-art deep learning-based methods for single cell… Expand
Inter-Species Cell Detection: Datasets on pulmonary hemosiderophages in equine, human and feline specimens
TLDR
This work introduces a novel, fully annotated multi-species P-Hem dataset which consists of 74 cytology whole slide images (WSIs) with equine, feline and human samples and aims to tackle several gaps present in currently available datasets. Expand
Fast Pre-Diagnosis of Neoplastic Changes in Cytology Images Using Machine Learning
TLDR
The experiments showed that the prototype model achieved 0.7416 mAP, which outperforms the state-of-the-art machine learning and human estimated results and provided a series of analyses that may facilitate ML-based solutions by casting more light on some aspects of its performance. Expand
Learning to be EXACT, Cell Detection for Asthma on Partially Annotated Whole Slide Images
TLDR
A WSI-based training approach which can train a deep learningbased object detection model with partially annotated WSIs and compensate class imbalances on the fly and yielded human-like performance when compared to the annotations of ten trained pathologists is developed. Expand
Artificial intelligence-based pathology for gastrointestinal and hepatobiliary cancers
TLDR
Artificial intelligence can automatically detect tumour tissue, easing the exponentially increasing workload on pathologists and capture prognostically relevant tissue features and thus predict clinical outcome across GI and liver cancer types. Expand
Assessing domain adaptation techniques for mitosis detection in multi-scanner breast cancer histopathology images
TLDR
Two mitosis detection models and two style transfer methods are trained and evaluated and it is found that the best of these models, U-Net without style transfer, achieved an F1-score of 0.693 on the MIDOG 2021 preliminary test set. Expand
Classification for avian malaria parasite Plasmodium gallinaceum blood stages by using deep convolutional neural networks
TLDR
The Darknet has the best performance in multiple class-wise classification, with average values of greater than 99% in accuracy, specificity, and sensitivity, and has a low misclassification rate (< 1%) than the other three models. Expand
Dogs as Model for Human Breast Cancer: A Completely Annotated Whole Slide Image Dataset
TLDR
A novel dataset of 21 WSIs of CMC completely annotated for mitotic figures, using machine learning to identify previously undetected MF and representation learning and two-dimensional projection to increase the consistency of the annotations. Expand
MitoDet: Simple and robust mitosis detection
TLDR
This method description presents the submitted algorithm for the Mitosis Domain Generalization Challenge, which employs a RetinaNet trained with strong data augmentation and achieves an F1 score of 0.7138 on the preliminary test set. Expand
Are fast labeling methods reliable? A case study of computer-aided expert annotations on microscopy slides
TLDR
A novel loss function is developed which incorporates the experts' annotation consensus in the training of a deep learning classifier and shows that a classifier trained with annotations from a single expert with computer-aided support can outperform the combined annotations from up to nine experts. Expand
Is Crowd-Algorithm Collaboration an Advanced Alternative to Crowd-Sourcing on Cytology Slides?
TLDR
This study shows that crowd-algorithm collaboration is a promising new approach to generate large datasets when it is ensured that a carefully designed setup eliminates potential biases. Expand
...
1
2
...

References

SHOWING 1-10 OF 60 REFERENCES
Alveolar macrophage graded hemosiderin score from bronchoalveolar lavage in horses with exercise-induced pulmonary hemorrhage and controls.
TLDR
The conclusion was made that careful assessment and scoring of alveolar macrophages for hemosiderin by means of the Golde scoring system shows promise as a more sensitive approach than repeated postexertional endoscopy alone to detect EIPH. Expand
Alveolar Macrophage Graded Hemosiderin Score from Bronchoalveolar Lavage in Horses with Exercise-Induced Pulmonary Hemorrhage and Controls
TLDR
The conclusion was made that careful assessment and scoring of alveolar macrophages for hemosiderin by means of the Golde scoring system shows promise as a more sensitive approach than repeated postexertional endoscopy alone to detect EIPH. Expand
Comparison of two cytological methods for detecting pulmonary haemorrhage in horses
TLDR
BAL allows detection of microscopic bleeding and identification of haemosiderophages, that is, alveolar macrophages containing haemOSiderin (an iron-storage complex produced by digestion of erythrocytes). Expand
Modern convolutional object detectors for nuclei detection on pleural effusion cytology images
TLDR
This work proposes the first implementation of the faster region-based convolutional neural network (Faster R-CNN), the region- based fully Convolutional network (R-FCN), and the single shot multibox detector (SSD) methods for nuclei detection on pathology images. Expand
Field Of Interest Proposal for Augmented Mitotic Cell Count: Comparison of two Convolutional Networks
TLDR
This work compares two methods for region of interest proposal, both based on convolutional neural networks (CNN), and indicates that, while the overall correlation to the ground truth mitotic activity is considerably higher for the approach based upon the fine resolution network, the field of interest choices are only marginally better. Expand
Diagnosis of pulmonary hemorrhage in the immunocompromised host.
TLDR
Thrombocytopenia and invasive fungal infections were statistically associated with severe hemorrhage, as was an increased percentage of alveolar macrophages in the BAL sample. Expand
Haemosiderin-laden macrophages in the bronchoalveolar lavage fluid of patients with diffuse alveolar damage
TLDR
In patients with an acute onset of diffuse lung infiltrates and respiratory distress, ≥20% haemosiderin-laden macrophages in BALF can occur with DAD, and is not necessarily diagnostic of DAH. Expand
Exercise-induced pulmonary hemorrhage.
TLDR
The most commonly performed method to diagnose EIPH at the present time is endoscopy of the upper airway alone or in combination with tracheal wash analysis for the presence of erythrocytes and hemosiderophages. Expand
Computerized Calculation of Mitotic Count Distribution in Canine Cutaneous Mast Cell Tumor Sections: Mitotic Count Is Area Dependent
TLDR
The findings of this study highlight the need to further standardize how to select the region of the tumor in which to determine the mitotic count. Expand
Bronchoalveolar lavage: sampling technique and guidelines for cytologic preparation and interpretation.
  • A. Hoffman
  • Medicine
  • The Veterinary clinics of North America. Equine practice
  • 2008
TLDR
An important advantage of BAL over tracheal wash cytology is that BAL cytology relates well to the clinical signs and pathophysiologic consequences of inflammatory airway disease. Expand
...
1
2
3
4
5
...