Tartaglia is a pioneering project formed by a public-private partnership consortium of 16 organisations, including the Vall d’Hebron Research Institute (VHIR). It is financed with close to 8 million euros from the Ministry of Economic Affairs and Digital Transformation with funds from the Next Generation Recovery, Transformation and Resilience Plan. The overall aim of the project is to create a federated network that helps improve clinical research in our country by using AI.

And, in commemoration World Prostate Cancer Day, work package 4 is worth a special mention: Early diagnosis of clinically significant prostate cancer using AI. This work package involves three clinical partners: the VHIR, as the coordinator; the Foundation for the Promotion of Health and Biomedical Research of the Valencian Community (FISABIO) and the Galician Health Knowledge Agency (ACIS/SERGAS). It also has four technical partners: Barcelona Supercomputing Center (BSC), Veratech, Opinno and GMV, the global coordinator of the project.

The main objectives of this work package are to improve early diagnosis of prostatic neoplasia by generating AI algorithms to help improve clinical practice. The project focuses on analysing MRIs and digitised biopsies and on integrating the analysis of various clinical variables to improve predictions of the possible appearance of clinically significant prostate cancer. All of this will enable us to improve predictions and personalise information on the potential risk of suffering prostate cancer, thereby providing quality personalised medicine.

In particular, AI algorithms are being developed that can be applied to a federated network to analyse MRIs and digitised biopsies. The aim is also to integrate several clinical variables into the analysis to improve predictions of the appearance of clinically significant prostate cancer. The implementation of neural networks in this project will help process large volumes of medical data efficiently, thereby facilitating early and accurate detection of prostate cancer. These neural networks can learn complex and subtle patterns in MRIs and biopsies, as well as clinical variables. As the neural network is trained with more data, its prediction capacity is expected to improve. The use of a federated network is particularly significant in this context, as it allows multiple medical centres to work together without having to directly share patient data. Each centre can train its own neural network using its own data, and then share the knowledge with the federated network. This guarantees data privacy and fosters collaboration in developing more effective AI algorithms.

The work package is based on the use of clinical variables, MRI segmentation and biopsies to generate an algorithm to improve the diagnosis of clinically significant prostate cancer.

In the case of clinical variables, the value of the different variables used in clinical practice to diagnose prostate cancer is integrated. These include blood prostate-specific antigen levels, family history and age at the time of biopsy, among others factors.

In the case of segmenting MRIs, the aim is to develop AI algorithms that can precisely identify and delimit the regions of interest in the prostate. Segmentation is an essential step for analysing and quantifying different characteristics and patterns in MRIs, which can help detect abnormalities and possible tumours in the prostate.

The aim is to develop algorithms to identify and classify the microscopic characteristics of prostate tissues, an essential step in determining the presence of cancer. By digitising pathological anatomy samples, we can analyse them using AI algorithms, potentially speeding up and improving the diagnostic process.

By combining MRI analysis, biopsy segmentation and the integration of clinical variables, the aim is to obtain an AI model that can accurately predict the presence of clinically significant prostate cancer. This will permit more precise diagnosis, thus facilitating the application of personalised treatments and improving medical care for patients.

It is important to stress the importance of regular testing to detect and prevent prostate cancer. Identifying it early can make all the difference to prognosis and treatment. Prevention and healthcare are essential for maintaining a full and active life.

The Tartaglia project is at the forefront of clinical research in the field of prostate cancer, demonstrating the potential of AI to improve medical practice and make progress in the diagnosis and treatment of diseases.

TARTAGLIA is part of the program R&D Missions in Artificial Intelligence of the Spain Digital Agenda 2025 and of the National Artificial Intelligence Strategy, funded by the European Union through the Next GenerationEU funds. The actions carried out will be reported to the Ministry of Economic Affairs and Digital Transformation (file number MIA.2021.M02.0005), corresponding to the funds of the Recovery, Transformation and Resilience Plan.