Smart algorithms against breast cancer: 14 teams will develop AI solutions to improve breast cancer diagnostics

Smart algorithms against breast cancer: Onco:Innovate partner AI4Health.Cro organises startup competition

During the competition 14 selected teams will develop AI solutions to improve breast cancer diagnostics. At the final demo day event on April 08.2025 in Zagreb, Onco:Innovate will select its winning teams aimed at improving breast cancer diagnostics. The winning teams will receive support, resources, and awareness to help bring their innovative ideas to life. This initiative highlights the potential of AI to enhance early detection and patient care, paving the way for a future where faster, more accurate diagnoses save lives.

The call for applications opened in December 2024, and a total of 34 teams with 106 competitors – IT experts, innovators and healthcare professionals, ready to face the challenges of breast cancer diagnostics – applied to the competition. After a thorough evaluation, the 14 best teams were selected and presented today at the kick-off event held at the Zagreb Innovation Center – ZICER. They continue the competition and enter the race for a prize fund worth 9,000 euros.

The competitors come from different parts of Croatia, from Čakovec, Karlovac, Osijek, Split and Zagreb. These are the names of the teams that continue the competition: Breast Friends, BreastQuestAI, Code to Heal, D308, Deregularizatori, GANga, Guard(AI)ns of Precision, JMM Kronos, MDT, MediSense, OncoVision, ProtostarLabsTeam and SCIOM.

“The fact that we had more than 100 applicants for the new AI4Health.Cro innovation competition in a specific niche area confirms the great interest, desire to contribute and the need for innovation in the healthtech sector. I congratulate all the teams starting the competition today, who were selected from among strong competition. ZICER’s mission is to encourage innovations that bring real change. By supporting projects like these, we develop technological solutions and build the foundations for a better and healthier society. We are proud to be part of a story that can improve healthcare in Croatia,” said Frane Šesnić , Director of ZICER.

How can AI help improve breast cancer diagnosis?

Breast cancer is the most common malignant tumor in women worldwide and is responsible for more than 670,000 deaths annually. Early diagnosis literally saves lives and national breast cancer screening programs have already proven their key role in saving lives by enabling timely mammography examinations, but the system is under great pressure. The number of images that need to be processed is increasing , and there is a shortage of skilled personnel and consequently difficult access to experts in rural regions. This is where artificial intelligence comes into play.

“AI-based solutions can significantly speed up and facilitate the analysis of mammographic images , enabling faster and more accurate diagnostic decisions. In this way, we help radiologists and relieve the burden on the healthcare system, optimize resources, and directly contribute to better patient care,” explains Dr. Anja Barešić , coordinator of AI4Health.Cro and head of the Laboratory for Computational Biology and Translational Medicine at the Ruđer Bošković Institute (IRB).

The goal is AI solutions applicable in clinical practice.

“The contestants will try their hand at four interesting tasks that examine how artificial intelligence can help in the analysis of mammograms. First, they need to develop an AI model that recognizes and classifies breast lesions according to the BI-RADS system, which is important for assessing the risk of breast cancer. Then, in the second task, they will have to mark suspicious areas on the images , just like an experienced radiologist. The third task asks them to explain how their model reaches conclusions, which helps to detect important patterns and possible biases. Finally, the fourth task is to design a simple and useful interface that makes it easier for doctors to interpret the results, with clear explanations of the diagnoses and their reliability,” explains Dr. Tomislav Šmuc,  Head of the Department of Electronics at the Ruđer Bošković Institute (IRB) and the Science and Technology Coordinator of the AI4Health.Cro project.

In order to solve these tasks, the competitors will use real, but anonymized mammography data, with strict privacy protection. This type of data allows testing solutions in a secure environment thanks to the cooperation of the AI4Health.Cro Center with hospitals and clinics, ensuring that the developed technologies are truly applicable in everyday clinical practice. The large amount of computing resources required to perform the tasks were provided by the Ruđer Bošković Institute and the University Computing Center – SRCE.

Grand finale in April – Onco:Innovate will provide support for the winning teams

The top three teams will present their solutions in April at the annual conference of the AI4Health.Cro Center, which brings together leading experts in healthcare and AI technology. Will there be a solution among them that will change the future of breast cancer diagnostics? Time will tell, but one thing is for sure – the race is on!