Deep learning for rapid and reproducible histology scoring of lung injury in a porcine model
(2023)- Abstract
- Acute respiratory distress syndrome (ARDS) is a life-threatening condition with mortality rates between 30-50%. Although in vitro models replicate some aspects of ARDS, small and large animal models remain the primary research tools due to the multifactorial nature of the disease. When using these animal models, histology serves as the gold standard method to confirm lung injury and exclude other diagnoses as high-resolution chest images are often not feasible. Semi-quantitative scoring performed by independent observers is the most common form of histologic analysis in pre-clinical animal models of ARDS. Despite progress in standardizing analysis procedures, objectively comparing histological injuries remains challenging, even for... (More)
- Acute respiratory distress syndrome (ARDS) is a life-threatening condition with mortality rates between 30-50%. Although in vitro models replicate some aspects of ARDS, small and large animal models remain the primary research tools due to the multifactorial nature of the disease. When using these animal models, histology serves as the gold standard method to confirm lung injury and exclude other diagnoses as high-resolution chest images are often not feasible. Semi-quantitative scoring performed by independent observers is the most common form of histologic analysis in pre-clinical animal models of ARDS. Despite progress in standardizing analysis procedures, objectively comparing histological injuries remains challenging, even for highly-trained pathologists. Standardized scoring simplifies the task and allows better comparisons between research groups and across different injury models, but it is time-consuming, and interobserver variability remains a significant concern. Convolutional neural networks (CNNs), which have emerged as a key tool in image analysis, could automate this process, potentially enabling faster and more reproducible analysis. Here we explored the reproducibility of human standardized scoring for an animal model of ARDS and its suitability for training CNNs for automated scoring at the whole slide level. We found large variations between human scorers, even for pre-clinical experts and board-certified pathologies in evaluating ARDS animal models. We demonstrate that CNNs (VGG16, EfficientNetB4) are suitable for automated scoring and achieve up to 83% F1-score and 78% accuracy. Thus, CNNs for histopathological classification of acute lung injury could help reduce human variability and eliminate a time-consuming manual research task with acceptable performance. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/47881d2f-1ac2-452c-b445-f7677c023ae0
- author
- Augusto Silva, Iran LU ; Kazemi Rashed, Salma LU ; Hedlund, Ludwig ; Lidfeldt, August ; Gvazava, Nika LU ; Stegmayr, John LU ; Skoryk, Valeriia LU ; Aits, Sonja LU and Wagner, Darcy LU
- organization
-
- LU Profile Area: Light and Materials
- LTH Profile Area: Nanoscience and Semiconductor Technology
- StemTherapy: National Initiative on Stem Cells for Regenerative Therapy
- NanoLund: Centre for Nanoscience
- Lung Bioengineering and Regeneration (research group)
- WCMM-Wallenberg Centre for Molecular Medicine
- Stem Cell Center
- LU Profile Area: Natural and Artificial Cognition
- Cell Death, Lysosomes and Artificial Intelligence (research group)
- MultiPark: Multidisciplinary research focused on ParkinsonĀ“s disease
- Medical Microspectroscopy (research group)
- LTH Profile Area: Engineering Health
- EpiHealth: Epidemiology for Health
- LTH Profile Area: AI and Digitalization
- eSSENCE: The e-Science Collaboration
- publishing date
- 2023-05-14
- type
- Working paper/Preprint
- publication status
- published
- subject
- keywords
- acute respiratory distress syndrome, lung, histology, deep learning, convolutional neural network, computer vision
- publisher
- bioRxiv
- DOI
- 10.1101/2023.05.12.540340
- project
- Lund University AI Research
- Deep Learning for Lung Image Analysis
- Using artificial intelligence and advanced computational tools to identify biological pathways, disease mechanisms and therapeutic opportunities
- Artificial intelligence for lung histology analysis
- language
- English
- LU publication?
- yes
- id
- 47881d2f-1ac2-452c-b445-f7677c023ae0
- date added to LUP
- 2023-05-15 14:06:36
- date last changed
- 2024-05-16 09:25:14
@misc{47881d2f-1ac2-452c-b445-f7677c023ae0, abstract = {{Acute respiratory distress syndrome (ARDS) is a life-threatening condition with mortality rates between 30-50%. Although in vitro models replicate some aspects of ARDS, small and large animal models remain the primary research tools due to the multifactorial nature of the disease. When using these animal models, histology serves as the gold standard method to confirm lung injury and exclude other diagnoses as high-resolution chest images are often not feasible. Semi-quantitative scoring performed by independent observers is the most common form of histologic analysis in pre-clinical animal models of ARDS. Despite progress in standardizing analysis procedures, objectively comparing histological injuries remains challenging, even for highly-trained pathologists. Standardized scoring simplifies the task and allows better comparisons between research groups and across different injury models, but it is time-consuming, and interobserver variability remains a significant concern. Convolutional neural networks (CNNs), which have emerged as a key tool in image analysis, could automate this process, potentially enabling faster and more reproducible analysis. Here we explored the reproducibility of human standardized scoring for an animal model of ARDS and its suitability for training CNNs for automated scoring at the whole slide level. We found large variations between human scorers, even for pre-clinical experts and board-certified pathologies in evaluating ARDS animal models. We demonstrate that CNNs (VGG16, EfficientNetB4) are suitable for automated scoring and achieve up to 83% F1-score and 78% accuracy. Thus, CNNs for histopathological classification of acute lung injury could help reduce human variability and eliminate a time-consuming manual research task with acceptable performance.}}, author = {{Augusto Silva, Iran and Kazemi Rashed, Salma and Hedlund, Ludwig and Lidfeldt, August and Gvazava, Nika and Stegmayr, John and Skoryk, Valeriia and Aits, Sonja and Wagner, Darcy}}, keywords = {{acute respiratory distress syndrome; lung; histology; deep learning; convolutional neural network; computer vision}}, language = {{eng}}, month = {{05}}, note = {{Preprint}}, publisher = {{bioRxiv}}, title = {{Deep learning for rapid and reproducible histology scoring of lung injury in a porcine model}}, url = {{http://dx.doi.org/10.1101/2023.05.12.540340}}, doi = {{10.1101/2023.05.12.540340}}, year = {{2023}}, }