DigiPatICS: AI to improve cancer diagnosis and share results

The Image and Video Processing Group (GPI) at the Universitat Politècnica de Catalunya – BarcelonaTech (UPC), from to the Intelligent Data Science and Artificial Intelligence Research Centre (IDEAI-UPC), is taking part in the DigiPatICS project, led by the Institut Català de la Salut (ICS), to improve cancer diagnosis in public hospitals in Catalonia. This is achieved through the digitalisation of histopathological samples (that is, microscopic samples of biological tissue) and the incorporation of AI tools that automate the analysis to refine the diagnosis and make it more reliable. With this system, a larger number of samples can be analysed more quickly, allowing patients to receive a more precise diagnosis sooner.

Previously, tissue samples were analysed using microscopes, a process in which pathologists manually counted a small number of cells reactive to immunohistochemical stains (a set of techniques used to detect the presence of cancer). Since 2022, when the first phase of the project — dedicated to digitalisation — began, ICS hospitals have had 24 high-precision scanners connected to 13 stations for digitising samples. This has made it possible to analyse a much greater number of cells per sample and to share both the images and the diagnoses in real time and securely with the rest of the hospitals in the network, creating a shared database of more than four million digitised images.

Based on the digitised samples, AI algorithms using deep learning automatically detect, quantify and classify the cells reactive to immunohistochemical stains. This not only speeds up the process, but also improves statistical accuracy and enables diagnoses to be reproducible — one of the objectives of DigiPatICS. Just one year after the algorithms were implemented, in 2024, 13,885 images had already been analysed, supporting the treatment of more than 4,400 patients.

The results generated by the algorithms can be viewed by more than 170 pathologists across the seven ICS hospitals that provide this service: Arnau de Vilanova University Hospital (Lleida), Bellvitge University Hospital (L’Hospitalet de Llobregat), Germans Trias i Pujol University Hospital (Badalona), Girona Doctor Josep Trueta University Hospital, Joan XXIII University Hospital (Tarragona), Vall d’Hebron University Hospital (Barcelona), and Tortosa Verge de la Cinta Hospital. The diagnoses and images are also shared with Viladecans Hospital, completing the Institute’s network. This allows the hospitals to remain connected, jointly review results, and offer remote diagnoses, thus creating one of the largest digital pathology networks in the world. At the same time, the introduction of these tools has enabled the ICS to advance in the digital transformation of this type of medical service.

Although the project initially focused on breast cancer due to its high prevalence — with estimates suggesting that one in eight women has had or will have this type of cancer — the study is now being extended to other oncological pathologies, such as lung, colon and endometrial cancer.

The DigiPatICS project received, on 19 November in Barcelona, the award of the 15th UPC Research Valorisation Prize, in the category of best technology transfer project, granted by the UPC Social Council.

Image credits: UPC, Vall d'Hebron University Hospital.