Skip to content


Deep-Learning based Techniques for Precision Tissue-Cell Microarray Processing


The Project includes the following challenges:

  1. Using Artificial Intelligence (AI) and Deep-Learning (DL) algorithms to perform Quantitative Image Analysis of histological tissue slides to select the area with Cancer Cells which will be processed with the Tissue Microarrayer (TMA)
  2. Develop a Fully Automatic Tissue Microarrayer punching mechanism (Galileo-Hub) using the coordinates from Artificial Intelligence algorithms.


The innovative solutions developed refer to:

  1. Whole slide image pre-processing: The ability to train the AI and DL to analyse whole slide images and from each image identify the areas of interest (tumour and non-tumour regions).
  2. Automatic image alignment: The ability to automatically match the glass slide image (marked on the area of interest) with donor tissue image, using either: the feature-based methods and/or the “pixel intensity-based” algorithms, and sending the coordinates of the position of the Tissue Block to be cored (e.g.; where to extract the tissue for the TMA block) to the Galileo Hub.
  3. Fully Automation of the Tissue Microarray punching mechanism which will include sensors to verify, at each cycle, if the tissue is correctly extracted from the donor block and if the punching needles are broken or damaged.
  4. Develop the Galileo HUB fully automated Tissue Microarrayer integrating the AI-DL algorithms with the automatic punching mechanism and the Galileo CKxxx software.


We developed a prototype of a Fully Automated Tissue Microarrayer (Galileo-Hub) integrating Artificial Intelligence and Deep-Learning based Techniques for Precision Tissue-Cell Processing, which will be industrialized and planned for market release in the first quarter of 2022


The main impact of the Galileo-Hub will be to reduce the time of Tissue Microarray production workflow by: (1) reducing the time spend by the  Pathologist in selecting the area of interest from analysing hundreds of tissue samples; (2) increasing the precision of core picking thanks to the use of Artificial Intelligence and Deep learning algorithms; (3) make better and faster Tissue Microarray thanks to the automatic punching mechanism and the open architecture of the Galileo-Hub which allows the use of different sizes donor blocks for prospective and retrospective (from pathology archive) projects.