Direct award, national procurement; AI-based image analysis

University of Turku together with Universities Helsinki, Eastern Finland and Oulu (Units) have been granted funds to promote joint research in the following expertise area. Using the same software in all aforesaid Units, together with a dedicated human resource in exercising high level and comparable research to develop digital pathology, would facilitate unparallel networking and distribution of various quantitation and segmentation algorithms and machine learning-based tools for identification of structural and pathological features in tissue samples. The planned infrastructure development will enable FinGMice to provide modern/up-to-date high-throughput quantitative tissue phenotyping research and services, and to make the quantitative digital histopathological analysis easily accessible for the biomedical research community. Other consortium parties already have a common AI-based image analysis software (Visiopharm quantitation software) installed and available. To facilitate national level high-throughput quantitation of molecular pathology analyses as part of the FinGMice services together with other consortium parties, it is necessary to install the same full version of the software at Turku Center for Disease Modeling (TCDM, at the University of Turku) premises.

Reasoning In order to ensure the complete, uncompromised comparability of validated scientific research and it´s results it is utmost necessary to use same tools (software in this case) as other parties. The selected product also has unique set of necessarily needed features that enables efficient and comprehensive measurements annotated structures and objectives from tissue sections placed on widely used microscope glass slides: Deep Learning based AI as Standard • Flexible image analysis solution • No programming knowledge required • Academy training and service and support team • Train by example • Pre-trained Nuclei segmentation APPs • Adapt pre-built APPs or built your own • Scanner agnostic for flexibility. • Deployable on premise or cloud based • Access images from shared database • Batch-processing of images • Tools to analyze tissue-microarrays • Ability to analyse both Brightfield, Fluorescence and Multiplex images