In silico Pathology

Not only will it allow for equal health care regardless of geography, but it has the potential to address many of the challenges faced in the pathology field. This can be achieved mainly by implementing methods for in silico pathology, i.e. automated analyses of pathological samples such as tissue sections stained with haematoxylin and eosin (H&E) scanned with high-resolution scanners. Deep learning combined with ground truth based on the pathology assessment and patient outcome forms the basis for methods developed in the DoMore! project.

This is a project related to the DoMore! project. 

The current procedures needed to render a prognosis are time-consuming and subjective. During the project, we have utilized Deep Learning and Big Data to develop robust systems and ICT solutions to supplement or replace methods in pathology to increase productivity, quality, and hence treatment. All experiments described in this report have been designed to examine multiple samples from each tumour and patient. With this approach, we have addressed the heterogeneity issue by analysing more of the tissue samples with fewer resources.

With more digital tools on the market, we can increase efficiency in pathology and provide a more accurate prognosis to cancer patients. Many of the current tasks are performed manually under a microscope. One obvious advantage of digitalisation is that it would allow the pathologist to work from anywhere using only a computer. It would also be easier for clinicians to receive both second and expert opinions when one has the opportunity to make an assessment remotely.