Start date: 15-08-2020
End date: 15-02-2021
Whole slide imaging technology allowed the digitization of conventional glass slides, which led to several new opportunities in the Pathology field such as the integration of computational systems, most notably artificial intelligence (AI). However, the digitization process also brought along new challenges for the automated analysis of digitized images. One of the most important challenges arises from the presence of image artifacts.
Depending on the severity of artifacts that are present, tissue areas that are important for diagnosis could be unclear, unusable, or even be completely missing. Therefore, a quality control mechanism is needed to ensure that the whole slide image is in good condition to be analyzed by the pathologists. If a whole slide image contains too severe artifacts, the quality system could provide recommendations, such as rescanning the whole slide image.
In the Pathology department of Radboudumc, approximately 1.68% of scanned slides are rescanned due to poor quality. The high volume of slides scanned every day and the sheer size of the scanned images (reaching up to 100,000 x 200,000 pixels) make manual control and supervision a highly time-consuming task for technicians. We propose a solution to this problem in the form of an AI system that detects and highlights regions containing artifacts in whole slide images using deep learning tools. The goals of this AI system are to speed up the process of digitization of slides and to make other AI algorithms more robust against artifacts.
We will develop an AI system that can detect multiple types of artifacts. Specifically, this task can be seen as a semantic segmentation problem and we will use deep learning methods to solve it. With semantic segmentation, the goal is to recognize and understand what's in an image on a pixel level. During the project, experiments will be done to find the optimal combination of model architecture (e.g. U-net, PSPNet, and DeepLapV3), hyperparameters, data pre-processing steps, and prediction post-processing steps.
Initially, the AI system will be able to detect artifacts such as ink, out-of-focus, tissue folds, dust, and air bubbles. To our knowledge, these are the most common types of artifacts observed in whole slide images. We will consider other less common artifacts in the later stages of the project. The unique characteristics of some artifacts in combination with the wide range of tissue types and staining techniques make artifact detection a challenging task that requires sophisticated solutions. Therefore, an important goal of this project is to develop a generalizable AI solution that can accurately detect different types of artifacts in a wide range of tissue and staining types.
To train deep learning models we will use multiple data sources. These include an in-house data set of over 300 whole slide images from several organ types and staining. There will be additional data sets provided from partners (out-house data sets).
The desired outcome of the project is to deploy the AI-driven algorithm as part of the quality control (QC) systems in a clinical setting. An additional but important impact will be on the existing and upcoming AI algorithms by assisting them to make robust predictions in the presence of artifacts in whole slide images. A secondary project outcome will be in the form of a publication describing the AI system, deep learning methodology, used datasets, and obtained results within the scope of the project.
- Collect and label artifacts in whole-slide histopathology images.
- Development of AI for segmentation of regions with artifacts present.
- Development of a demonstrator to show the developed system in action.
- Writing of manuscript following publication guidelines.
The developed tool will be tested and subsequently implemented in the Department of Pathology of the Radboudumc. A major milestone of the project is to successfully deploy in the daily workflow of pathologists to speed up the process of digitization of slides as well as making other AI algorithms more robust against artifacts.