Digital histopathology resource for image-based diagnosis

Imaging of soft tissues provides a basis for clinical diagnoses and biomedical research. However, traditional methods of soft tissue imaging are inherently 2-D and microscope-based. This project proposes a digital histopathology workflow incorporating 2-D whole slide and 3-D micro-computed tomography (CT) imaging of standard formalin-fixed paraffin-embedded (FFPE) soft tissue. Previous work at Southampton has shown that correlative whole slide and
CT imaging of FFPE lung tissue provides further insight into lung pathophysiology than traditional histopathological methods (Jones et al., 2016). This project aims to determine a workflow for correlative whole slide imaging (WSI) and CT of a range of standard FFPE soft tissue samples optimised for image quality and high throughput.The project will be split into four work packages:
1.Optimisation of WSI+CT workflow for maximal image quality
Identification of imaging parameters most relevant to biomedical research, eg. required resolution, sample preparation, contrast-to-noise ratio, etc, through a user survey.
Determination of a WSI+CT workflow resulting in optimal relevant image quality metrics, including sample preparation, image acquisition and reconstruction, with plans for data management.
2.Optimisation of WSI+CT workflow for high -throughput imaging
Determination of a WSI+CT scanning protocol optimised for high throughput while providing images of sufficient quality for post-processing (eg. segmentation and quantitative analysis), including sample mounting, image acquisition and reconstruction and data management. This could also involve design of specialised rigs to aid co-registration between imaging modalities.
Demonstration of the capabilities of the optimised workflow with a small initial study including a range of FFPE soft tissue samples from key collaborators.3.Simultaneous and interactive visualisation of correlative 2-D whole slide & 3
-D CT data
Survey of the research needs of the clinical and biomedical research community, ie. identification of features of the visualisation framework that would be most useful.
Development of a framework for interactive visualisation of co-registered 2-D WSI and 3-D CT data which addresses the research needs of the clinical and biomedical research community. This framework should also support quantitative maging features.
4.Application of correlative imaging and visualisation framework for a larger tissue archive study
Sourcing of a large sample size of a range of FFPE tissue samples from tissue banks and/or key collaborators
on the Wellcome Trust Biomedical Resource and Technology Development grant.
Large scale proof-of-principle study demonstrating the utility of the correlative imaging and visualisation framework.
Generation of a large dataset of images for training of machine learning algorithms to answer relevant research questions, as determined from the research needs of the clinical and biomedical research community.