Winter School XAI 2025: Registration Now Open!

At the beginning of the new year, the Winter School XAI 2025 will occur from January 6 to 10, 2025, at the Masaryk Oncology Institute in Brno and the Medical University of Graz. This unique program will provide practical insights into the latest artificial intelligence (AI) advancements in digital pathology.

Participants will engage in training sessions led by experts, covering essential topics such as AI-assisted diagnostics, the significance of explainability in medical AI, and certification processes according to the In Vitro Diagnostic Regulation (IVDR). The program also includes a practical component, allowing participants to work directly with digital pathology data, including whole-slide images and clinical data.

The education will be conducted by prominent professionals in the field who will share valuable insights into the future of AI in pathology. Participants in this school will deepen their knowledge and skills in digital pathology, data preparation for AI, and critical aspects of privacy protection.

Registration is open until December 15, 2024. To secure your place, please fill out the registration form. Take advantage of this opportunity to expand your knowledge!

BioMedAI Team Showcases Research at the first ASDP Congress in Seoul

Representatives of the BIoMedAI project participated in the 1st Annual Congress of the Asian Society of Digital Pathology (ASDP), which took place from October 2 to 4, 2024, in Seoul, South Korea.

The delegation included Vít Musil, the team leader from the RationAI lab and Assistant professor at the Department of Computer Science, along with PhD student Adam Bajger and master's student Adam Kukučka, who presented the results of his research. This research, initially part of his bachelor’s thesis, is now under review.

The ASDP 2024 congress focused on the current status and guidelines for Asia's digital pathology (DP) and artificial intelligence (AI). It brought together experts from academia and industry to discuss how these technologies can be standardized and ethically applied to advance healthcare across the region. The program included sessions on the practical application of AI in pathology, emerging regulatory issues, and the latest innovations in digital pathology imaging. The conference provided a unique opportunity for attendees to network and exchange knowledge with global leaders in the field, further bridging the gap between research and real-world applications.

This conference was a valuable platform for connecting experts from pathology, informatics, academia, and business sectors, enabling a comprehensive view of the latest research and practical applications in digital pathology and AI.

XAI Summer School 2024 Brings Together Experts and Enthusiasts in Brno

From September 2 to 6, 2024, the Faculty of Informatics and the Institute of Computer Science at Masaryk University in Brno hosted a successful summer school focused on Explainable Artificial Intelligence (XAI) in digital pathology. Organized by members of the BioMedAI project, the event attracted nearly 40 participants eager to explore the applications of AI in medicine.

Petr Holub from Masaryk University, one of the main organizers of the summer school and Project Lead of the BioMedAI project, shared his assessment of the event: "The atmosphere was incredibly positive, with students showing great enthusiasm and actively participating in all sessions. We managed to create a collaborative environment where participants could openly share their experiences and knowledge. The summer school's greatest achievement was providing practical skills in using XAI tools in digital pathology, which is crucial for many of the attendees’ research. The hands-on workshops, where participants could apply new methods and tools to real-world data, were particularly well-received and sparked the most interest."

Throughout the week, attendees engaged in lectures and hands-on exercises led by experts in the field, including Norman Zerbe from Charité, Heimo Müller from the Medical University of Graz, Aray Karjauv from Technical University Berlin, Tomáš Brázdil, Petr Holub and Vít Musil from Masaryk University, and Rudolf Nenutil from the Masaryk Memorial Cancer Institute. An online lecture on the state of the art of XAI in digital pathology was presented by David Steiner from Google. The program covered a wide range of topics, from the basics of pathology to advanced AI research, data management, and practical training using tools like RatioPath and RatioViz.

Aleksandra Zec, a PhD student from Politecnico di Milano who is working with digital pathology data in her research, shared her positive experience: "I had a great time in Brno, especially since I am working with digital pathology data in my PhD research. Even though I’m quite familiar with the topic, I found the sessions to be both useful and comprehensive. The hands-on exercises and lectures by leading experts provided valuable insights into the applications of AI in digital pathology, which will definitely benefit my ongoing research."

The summer school provided participants with valuable practical experience and a deep understanding of the latest technologies in XAI. The event fostered a collaborative learning environment, inspiring attendees to actively engage with the future of medical innovation and contribute to the advancement of AI in medicine.

Successful Start of the XAI Summer School 2024

September 4, 2024

The XAI Summer School has successfully started, with nearly 40 participants attending lectures by experts in the field. Renowned speakers are sharing valuable insights and inspiring the next generation to explore this high-potential area that promises significant advancements in medicine.

The event fosters a collaborative learning environment, motivating participants to engage deeply with the future of medical innovation.

Program

Monday
10:00-10:15 Intro (Petr Holub)
10:15-12:15 Introduction to pathology (Rudolf Nenutil)
12:15-13:15 Lunch break
13:15-15:15 Intro to XAI on (not only digital pathology) imaging (Aray Karjauv)
15:15-15:30 Break
15:30-17:00 State of the art in (X)AI in digital pathology (Dave Steiner)

Tuesday
9:00-11:00 Data in digital pathology - images, annotations, metadata, clinical data. Clinical AI applications. (Norman Zerbe)
11:00-12:00 Handling sensitive data (Petr Holub)
12:00-13:00 Lunch break
13:00-13:20 Metacentrum Registration (Vít Musil)
13:20-14:20 Handling data quality (Karel Štěpka)
14:20-14:30 Break
14:30-16:30 Mining clinical data (Aleš Horák)

Wednesday
9:00-12:00 Hands-on: Training a simple AI model in digital pathology (part 1) (Vít Musil)
12:00-13:00 Lunch break
13:00-15:30 Hands-on: Training a simple AI model in digital pathology (part 2) (Vít Musil)
15:30-16:00 Break
16:00-17:30 XAI in digital pathology (Aray Karjauv)

Thursday
9:00-10:30 Hands-on: Explainability of the trained model from Wednesday (Vít Musil)
10:30-11:00 Break
11:00-12:00 Clinical data relevant for pathology (Rudolf Nenutil)
12:00-13:00 Lunch break
13:00-16:00 Hands-on: Training AI model on Whole Slide images (part 1) (Vít Musil)

Friday
9:00-10:30 Hands-on: Model inference on WSI (Vít Musil)
10:30-12:00 Hands-on: Explainability (Vít Musil)
12:00-13:00 Lunch break
13:00-13:30 Wrap up of the school (Petr Holub)

Peter Holub is a representative of the EOSC Association Health Data Task Force

August 15, 2024

The EOSC Association Task Forces play a key role in the development of the European Open Science Cloud (EOSC). These teams, composed of experts and volunteers from across Europe, focus on specific areas essential for the implementation and advancement of EOSC.

The recent reorganisation for the 2024–2025 period has brought significant changes. Peter Holub has been appointed as the Co-Chair of the Health Data Task Force, which has been operating in its new form since July this year. He will be directly involved in addressing issues related to the interoperability and use of health data in research, in accordance with FAIR principles.

"Medical data is the subject of numerous activities aimed at standardising it (e.g., ISO, HL7, and other standardisation bodies) and making it discoverable and accessible. My goal is to find ways in which EOSC can effectively support the use of medical data in scientific research. Access to medical data is one of the essential conditions for training AI models that advance medicine and improve treatment outcomes. This is also our focus at BioMedAI, where we use AI in oncopathology."

XAI Winter School 2025

Dates: January 6-10, 2025
Place: the Masaryk Memorial Cancer Institute in Brno, Czech republic & the Medical University of Graz, Austria

From January 6 to 10, 2025, the Masaryk Memorial Cancer Institute in Brno and the Medical University of Graz will host a Winter School focused on Artificial Intelligence (AI) in digital pathology. This specialized program offers participants a hands-on introduction to the latest advancements in AI, with a particular focus on its application in digital pathology.

Expert-led training on AI and Digital Pathology

Throughout the week, participants will attend lectures and practical sessions led by prominent experts in the field. These sessions will cover various topics, including the role of AI in assisting diagnoses, the importance of explainability in medical AI systems, and the certification processes required under the In Vitro Diagnostic Regulation (IVDR). Participants will also gain insight into preparing data for AI training, such as whole-slide images (WSIs), annotations, clinical and omics data, and data privacy considerations.

Hands-on experience in Brno and Graz

The program includes practical, hands-on sessions allowing participants to work directly with digital pathology data. Participants will practice scanning and saving WSIs, creating annotations, and extracting structured data from unstructured clinical data using advanced AI tools such as Natural Language Processing (NLP) and large language models (LLMs). The week will conclude with a tour of the state-of-the-art digital pathology infrastructure at the Medical University of Graz.

What Will Participants Gain from the Winter School?

• Getting pathologists familiar with state-of-the-art in digital pathology
  ( Hands-on in Graz on entire digital pathology dept. )
  
  
• Getting pathologists familiar with the possibilities of different AI approaches - from assistance to diagnosis
  
  
• Getting familiar with explainability
  
  
• Getting familiar with IVDR & certification processes
  
  
• Training pathologists on how to prepare data for AI
  WSIs - formats
  WSI annotations - formats, semantics, SOPs
  clinical data - extraction manual vs. NLP vs. LLMs
  omics data - basic formats
  
  
• Data privacy aspects - pseudonymized vs. anonymized data, types of contracts

Organizational aspects

1. Part - BRNO
The first part of the Winter School will take place from January 6 to 8, 2025, at the Masaryk Memorial Cancer Institute in Brno, where participants will focus on introductory sessions, AI use in digital pathology, and practical hands-on training.

>>> Transfer to Medical University Graz by train on Wednesday evening

2. part - GRAZ
On the evening of January 8, participants will travel by train to Medical University Graz, where the second part of the program will be held from January 9 to 10, 2025. In Graz, participants will experience a guided tour of the digital pathology infrastructure and gain additional practical insights.

Registration information

Registration takes place via the form. You can register until 15 December. After that time, registration will be closed.
The registered participants will receive an informative email regarding the organization and other details of the school at the latest by 20 December 2024.

Fees

Participation is free of charge.
Participants are expected to cover their travel and accommodation expenses.

Bridging Medicine and Technology: The BioMedAI Project Sheds Light on Cancer Diagnosis with AI

16. 4. 2024

The non-profit organization Science Party hosted an event spotlighting the collaborative efforts of pathologists and IT specialists from the BioMedAI project as they presented the challenge of cancer using AI!

Through the BioMedAI project, they demonstrated to the public how they're reshaping medicine and computer science. Their efforts in training AI to identify prostate cancer cells are speeding the diagnosis and treatment of the disease and setting a new standard in interdisciplinary collaboration.

Pathologist Rudolf Nenutil from the Masaryk Cancer Memorial Institute captivated everyone by comparing studying cells to detective work. At the same time, IT scientist Tomáš Brázdil explained the complexities of training AI to the audience.

This lecture wasn't just for specialists - it was a gathering for the wider community, showcasing the incredible strides in the fight against cancer through the power of AI.

Event: https://www.scienceparty.cz/udalost/24_02

Photo: Lucie Skřičková, Masaryk University

Rudolf Nenutil on the program BILANCE featured by Czech Television

6. 3. 2024

BILANCE is a journalistic format focusing on the analysis and evaluation of current events and topics in Czech society. The March edition of the program tackled the theme "Will We Beat Cancer?" and brought many updates from the field of cancer diagnostics and treatment.

Rudolf Nenutil, a medical oncopathologist, introduced software utilizing artificial intelligence for the recognition of cancerous findings. He collaborates with a team of computer scientists from Masaryk University on the BioMedAI project, which holds revolutionary potential in the field of cancer diagnostics.

To watch the program in the Czech language: BILANCE - Porazíme rakovinu? 

Photo: MOÚ
Source: www.mou.cz

Pathologist Rudolf Nenutil featured on Czech Radio podcast: Discussion on Software Development utilizing artificial intelligence for cancer detection

7. 2. 2024

Pathologist Rudolf Nenutil recently appeared as a guest on a popular Czech Radio podcast, where he shared insights into the development of software utilizing artificial intelligence for the detection of cancerous findings. Collaborating with a team of computer scientists from Masaryk University, he contributes to the BioMedAI project, which holds revolutionary potential in the field of cancer diagnostics.

During the interview, Nenutil emphasized that traditional methods of pathological examination of tissue under a microscope have seen little advancement over the past one hundred and twenty years. However, the new software has the potential to significantly increase the efficiency and accuracy of diagnostics, enabling faster and more precise identification of tumors. The software can capture even the smallest changes that might escape the attention of pathologists.

Additionally, the podcast delves into the importance of pathologists' work, which plays a crucial role in the cancer treatment process, and the significance of the new software, which could greatly facilitate and improve the entire treatment process.

You can listen to the whole interview in the Czech language: Metoda rozpoznávání rakovinových nálezů se 120 let nezměnila. Práci pomůže zrychlit nový software | Brno (rozhlas.cz)

Author: Ludmila Opltová from The Czech Radio BRNO

Petr Holub, BioMedAI representative, contributes to discussion on FAIR data and the impact of AI on research  

24.–24. 1. 2024

In late January, a discussion regarding the significance of integrating FAIR data into scientific research occurred in Milan. The event ESFRI-EOSC Policy Workshop on "FAIR Data Productivity and Advanced Digitalization" brought together representatives from European research infrastructures to delve into the future of research.

Petr Holub, representing BioMedAi and BBMRI, actively participated in this discussion and highlighted the indispensable role of artificial intelligence in medical research. Specifically, he emphasized its significance in digital pathology for cancer diagnosis and treatment and pointed out its contribution to anonymizing and synthesizing data for publication, addressing the ongoing challenge of managing sensitive data. Holub's speech underscored that the role of EOSC extends beyond mere funding; it plays a crucial part in facilitating these essential tasks.

Experts also emphasized the importance of collaboration within research communities to ensure effective implementation of FAIR data practices and discussed challenges including staff training, data accessibility, and management of sensitive data.

Photo: J. Heyrovsky Institute of Physical Chemistry
Source: https://www.jh-inst.cas.cz/cs/news/fair-data-vliv-ai-na-vyzkum-v-milane-se-diskutovalo-o-tom-proc-je-dulezite-oba-aspekty

Czech Open Source Policy Forum Highlights Contributions from BioMedAI

On April 24, 2024, the Czech Open Source Policy Forum took place in Brno, featuring prominent discussions on the sustainability of open-source projects within the public sector and academia. Among the attendees was Petr Holub, a representative from the BioMedAI project. He actively participated in discussions focused on the importance of community engagement for the sustainability of these initiatives. Holub's insights contributed to a deeper understanding of how collaborative efforts can enhance the longevity and impact of open-source projects.