AI-IBD

Personalized medicine in IBD: integration of clinical, laboratory and AI models for outcome prediction - AI-IBD (IP-UNIST-32)


 
Project title: Personalized medicine in IBD: integration of clinical, laboratory and AI models for outcome prediction - AI-IBD (IP-UNIST-32)
Project applicant: University of Split, School of Medicine
Project value: EUR 238,000.00
Project duration: 01/10/2025 - 30/09/2029
Project leader: Prof. Joško Božić, MD, PhD
Contact for further information: Prof. Joško Božić, MD, PhD (josko.bozic@mefst.hr)
 
Project description
Although IBD is primarily classified as a disease of the gastrointestinal system, in a substantial proportion of cases, estimated between 25% and 40%, manifestations outside the digestive tract are also present, affecting various organ systems. In addition, in approximately 10% of patients, extraintestinal manifestations (EIM) represent the initial clinical presentation of the disease. Despite their relatively high prevalence, these manifestations often remain unrecognized, resulting in delayed diagnosis and inadequate therapeutic management. Furthermore, there are currently no reliable prognostic factors that would enable prediction of EIM development.
Accordingly, the aim of this research project is to conduct a comprehensive clinical analysis of patients with IBD, with a particular emphasis on detecting systemic manifestations of the disease, including those that have not yet become clinically apparent. Using machine learning algorithms, the project will seek to identify potential predictive patterns and interrelationships that could contribute to earlier recognition and a better understanding of the systemic nature of IBD.
In recent years, alongside significant advances in computing technology and machine learning algorithms, several research groups have attempted to develop models for predicting extraintestinal manifestations (EIM), with varying degrees of success. For example, Verma et al. developed a predictive model for one of the most common EIMs, arthropathy, based on clinical and demographic data from a relatively limited number of patients. Other authors developed models combining clinical parameters with genetic variations. Finally, in a recently published study, Baumgart et al. applied an artificial intelligence-based network analysis approach to a cohort of nearly 30,000 patients with IBD, identifying EIM clusters and developing an interactive model that enables clinicians to visualize and recognize interrelationships among different EIMs.
 
Project collaborators

• Assist. Prof. Anteo Bradarić-Šlujo, PhD
• Assist. Prof. Josipa Bukić, PhD
• Prof. Darko Duplančić, PhD
• Assoc. Prof. Tea Galić, PhD
• Assoc. Prof. Duška Glavaš, PhD
• Assoc. Prof. Iris Jerončić Tomić, PhD
• Prof. Ivana Kolčić, PhD
• Assoc. Prof. Slavica Kozina, PhD
• Assoc. Prof. Mladen Krnić, PhD
• Marko Kumrić, PhD
• Assist. Prof. Slaven Lupi Fernandin, PhD
• Prof. Valdi Pešutić Pisac, PhD
• Prof. Željko Puljiz, PhD
• Assist. Prof. Doris Rušić, PhD
• Prof. Tina Tičinović Kurir, PhD
• Assoc. Prof. Marion Tomičić, PhD
• Prof. Marija Tonkić, PhD
• Assoc. Prof. Ivana Unić, PhD
• Assist. Prof. Marino Vilović, PhD
• Josip Vrdoljak, PhD
• Marija Franka Žuljević, PhD
• Assoc. Prof. Daniela Šupe-Domić, PhD (FZZ Split)
• Marina Rudan Dimlić, PhD (MedILS)

Measurable indicators of project result achievement
The project is designed to achieve clearly defined, measurable results aligned with the program objectives of the call. The expected outcomes derive from four core work packages (WP1-WP4) and can be summarized as follows:
 
Expected project results

• Creation of a high-quality, multidisciplinary database of clinical and subclinical data on patients with IBD (WP1).
• Development and validation of a predictive model to identify patients at increased risk of developing EIM using machine learning algorithms (WP2).
• Publication of scientific results in high-ranking journals and dissemination through congresses and professional meetings (WP3).
• Active involvement and training of young researchers and PhD candidates, with encouragement of interdisciplinary collaboration (WP3, WP4).
• Implementation of popular-science activities and public health education in the community (WP4).
• Establishment of international scientific collaboration and dissemination of results beyond the Republic of Croatia (WP3, WP4).

 
These indicators directly contribute to the following program objectives:

• Objective 1.1 - Increasing scientific productivity and competitiveness of public higher education institutions
• Objective 1.4 - Strengthening human resources for scientific work (through PhD candidates and training)
• Objective 1.7 - Contributing to open science (FAIR database and availability of ML models)
• Objective 3.3 - Internationalization and international mobility
• Objective 4.6 - Popularization of science and collaboration with citizens

 
Through a clearly structured work plan (WP1-WP4), all indicators are methodologically grounded and realistically achievable over the 48-month project duration. Transparency, replicability, and open science ensure long-term sustainability and broad applicability of the results in clinical practice and further research.
 
“Funded by the European Union - NextGenerationEU.”
The views and opinions expressed are those of the author only and do not necessarily reflect the official views of the European Union or the European Commission. Neither the European Union nor the European Commission can be held responsible for them.
  Print page