The two partners have launched ‘Bianca’, the first large-scale digitalisation initiative in Italy focused on histopathological samples, aimed at training AI algorithms to support pathologists’ diagnostic activities.
The Pathology Division of the European Institute of Oncology (IEO) and Laife Reply, the Reply Group company specialised in AI and Big Data solutions for the healthcare sector, have entered into a collaboration to develop Bianca, the first project in Italy aimed at creating an AI-based digital biobank designed as an integral part of clinical diagnostic practice.
The initiative is part of a broader technological innovation journey that structurally integrates research and development into routine diagnostic processes in pathology, transforming the traditional histopathological sample workflow into an end-to-end digital ecosystem. The complete digitalisation of histopathological and molecular diagnostic workflows aims to make analysis more efficient, scalable and reproducible, laying the foundations for the evolution of AI-supported oncological diagnostics.
Selected under the ‘Agreements for Innovation’ programme promoted by the Italian Ministry of Enterprises and Made in Italy (MIMIT) and building on a well-established collaboration between IEO and Reply, the BIANCA project initially involves the digitalisation of histopathological slides using state-of-the-art scanners capable of generating ultra-high-resolution digital images of tissue samples.
On this platform, Artificial Intelligence algorithms are developed and trained to analyse images and support pathologists in histological and molecular analysis, as well as in the formulation of clinical and diagnostic hypotheses.
Launched at the end of 2024 with a planned overall duration of 30 months, the project has now reached its midway point. The extensive archive of histopathological samples collected over the years by the IEO Biobank is currently at an advanced stage of digitalisation.
At the same time, Laife Reply is working closely with the joint IEO Pathology and Information Systems team to train AI algorithms on different types of cancer. This includes the introduction of advanced algorithmic solutions based on self-annotation mechanisms, capable of automatically labelling pathological findings on images, reducing the manual workload for clinicians, accelerating model training, improving accuracy and enabling large-scale replicability.
The project is also exploring the use of multimodal algorithms capable of combining histopathological images with structured clinical data to identify and analyse new biomarkers.
In particular, research activities are already underway on specific diseases, with the aim of predicting information that is currently obtainable only through complex tests, thereby reducing time, costs and the overall impact on patients.
“Bianca represents a turning point for oncological pathology,” said Professor Nicola Fusco, Director of the Pathology Division, IEO. “The integration of digitalisation and AI enables a significant improvement in the quality, standardisation and reproducibility of diagnosis – both histopathological and molecular – by optimising the entire workflow, reducing reporting times, rationalising costs and improving the overall efficiency of diagnostic services for our patients. At the same time, the project contributes to the training of a new generation of pathologists with highly specialised skills, capable of combining morphological and molecular expertise with advanced digital tools and AI algorithms, paving the way for a sustainable evolution of oncological diagnostics.”
“With Bianca, we are collaborating with IEO to support the evolution of pathology in the oncological field,” said Carlo Malgieri, Partner, Laife Reply. “This is not just about applying Artificial Intelligence to individual cases, but about building a scalable and industrialisable framework designed to be offered to hospitals and smaller healthcare organisations. The framework integrates sample digitalisation, advanced algorithms and high-performance analytics infrastructures. This approach makes it possible to support clinicians, enable new services for healthcare systems and oncological research and ensure transparency and explainability – key elements to guarantee that every algorithm-supported decision is trustworthy.”
