Newsletter.

Welcome to the November 2024 issue of the
Medical Imaging and Data Resource Center (MIDRC) newsletter!

MIDRC Tools to Empower Your Research

MIDRC offers a suite of easy-to-use tools to support medical imaging and AI research. Designed to simplify tasks like cohort building, analyzing radiology reports, and selecting AI evaluation metrics, these tools are accessible to researchers of all skill levels—no programming required. Here’s what you can explore:

• MIDRC Data Portal: Access and navigate extensive medical imaging datasets with powerful search and filtering capabilities.

• MIDRC BioMedical Imaging Hub: A centralized repository for imaging data, tools, and documentation, supporting seamless data search, visualization, and analysis.

• RadGPT: MIDRC LLM for Radiology Reports: A tool that explains radiology reports in plain language and suggests follow-up questions for patients to discuss with their doctors.

• MIDRC-MetricTree: An interactive guide to help you choose the most suitable performance metrics for medical imaging models.

• MIDRC Bias Awareness Tool:Identify and address biases in AI models to ensure fair and reliable imaging outcomes.

• New Tools in Development: Keep an eye out for the Helper AI tool for seamless MIDRC dataset navigation and the MIDRC's Interoperability Tool for integrating imaging with other data types.

Access Tools on Our Website

• Web-Based Tools: Use interactive tools directly online. Explore here.

• Downloadable Tools: Access downloadable resources like algorithms for offline use. Explore here.

Visit our website to explore these tools and stay updated on upcoming features. For guidance, check out our seminar series, video tutorials, and FAQs, or reach out with your questions.

MIDRC at RSNA 2024!
Chicago, IL from December 1–5, 2024

The Radiological Society of North America's Annual Meeting (RSNA 2024) is just around the corner! This year, the Medical Imaging and Data Resource Center (MIDRC) is set to play a significant role, offering informative sessions and interactive experiences aimed at advancing the integration of AI in medical imaging.

Highlighted MIDRC Sessions:

At RSNA 2024, MIDRC will host several key sessions, including specialized workshops in the Deep Learning Lab and interactive events covering data diversity, cohort building, and cancer imaging research. These are designed to offer hands-on insights into MIDRC’s extensive data resources and expertise in AI.

1. Cohort Building for AI Research

• Date/Time: Monday, December 2, 11 a.m. - 12 p.m.

• Location: Deep Learning Lab, Course #DLL04

• Overview: Discover how MIDRC’s data cohorts can support your AI research. This workshop will guide participants through MIDRC’s data access tools and cohort-building methodologies, providing an essential foundation for imaging research projects.

2. Contributing Diverse Data for Inclusive AI

• Date/Time: Tuesday, December 3, 1 p.m. - 2 p.m.

• Location: Deep Learning Lab, Course #DLL09

• Overview: Emphasizing the importance of diversity in AI training datasets, this session covers MIDRC’s work in ensuring representation within imaging data, from annotation to data sharing strategies that support equitable outcomes in medical imaging.

3. Engaging the Cancer Imaging Research Community

• Date/Time: Tuesday, December 3, 4:30 p.m. - 5:30 p.m.

• Location: Education Session, Course #T8-RCP19

• Overview: Experts delve into MIDRC'S cancer imaging projects, sharing how its resources can empower research on cancer detection, diagnosis, and treatment, and enhance collaboration within the research community.

4. Innovation Theater: MIDRC Overview and Updates

• Date/Time:Wednesday, December 4, 4 p.m. - 5 p.m.

• Location: Innovation Theater

• Overview: This session provides a comprehensive update on MIDRC’s initiatives to advance medical imaging through its collaborative efforts. Learn about MIDRC’s ethically curated imaging commons, which supports AI development and biomedical research, and explore how MIDRC’s infrastructure assists researchers, both new and experienced, in tackling major public health challenges beyond COVID-19. Key speakers will discuss MIDRC’s creation, goals, and the impact of its resources on public health and patient care. For further details, visit MIDRC.org.

Explore the MIDRC Booth

Throughout RSNA 2024, the MIDRC booth at the Lakeside Learning Center will be a hub of activity. Visitors can engage with MIDRC staff, learn about the latest advancements in AI for imaging, and explore how MIDRC data resources and tools can fuel their own research endeavors.

Why Attend?

MIDRC’s sessions are designed to empower researchers, radiologists, and AI professionals with practical knowledge, collaborative opportunities, and access to one of the world’s most comprehensive imaging databases. The RSNA Annual Meeting brings together the best in radiology, AI, and medical imaging, making it the perfect venue for MIDRC to showcase its initiatives and foster deeper engagement with the imaging community.

For more details, visit RSNA Annual Meeting 2024 and MIDRC at RSNA 2024.

Celebrating the Winners of
the MIDRC XAI Grand Challenge!

The MIDRC XAI Grand Challenge has successfully concluded, bringing together top researchers and teams to tackle the complex task of making artificial intelligence (AI) more interpretable for medical imaging. This year’s competition, titled “Decoding AI Decisions for Pneumonia on Chest Radiographs,” invited participants to advance methods for understanding AI decision-making in pneumonia detection on chest X-rays.

Here are the outstanding winners:

1st Place: Dr. Ian Pan, Brigham and Women’s Hospital
Prize: $15,000

2nd Place: Dr. Mathieu Goulet, Centre régional intégré de cancérologie (CRIC) and Centre de recherche du CISSS de Chaudière-Appalaches
Prize: $8,000

3rd Place:Dr. Yijie Yuan, Johns Hopkins University
Prize: $7,000

4th Place: Team MIA-UniBern: Members of this team include Haozhe Luo, Zixin Shu, Aurélie Pahud de Mortanges, and Mauricio Reyes from the University of Bern’s ARTORG Center, with Ziyu Zhou from Shanghai Jiao Tong University.
Prize: $5,000

5th Place: Team AIRI, Moscow: This team is a collaborative effort from the Artificial Intelligence Research Institute and the Kharkevich Institute for Information Transmission Problems, including members Valentin Samokhin, Aleksei Bokov, Evgeniia Przhedzetskaia, Mikhail Goncharov, Iaroslav Bespalov, Boris Shirokikh, and Vladimir Shaposhnikov.
Prize: $5,000

6th Place: Dr. Joseph Paul Cohen, Stanford University
Prize: $5,000

7th Place: Kerol Djoumessi, Hertie Institute for AI in Brain Health, University of Tübingen
Prize: $5,000

Why This Challenge Matters:
Explainable AI is essential for integrating AI into clinical workflows in a safe and transparent way. By fostering research in XAI, MIDRC is supporting innovations that make AI tools for medical imaging more understandable, accessible, and trustworthy for healthcare professionals.

Congratulations to all the winners, and thank you to everyone who participated in and supported this significant effort, including the CodaLab Challenge platform!

MIDRC XAI Grand Challenge Review:
Decoding AI Decisions for Pneumonia on Chest Radiographs

Join us on December 17 for an exciting deep-dive webinar featuring the top three finalists from the MIDRC Explainable AI (XAI) Grand Challenge! This event, titled "MIDRC XAI Grand Challenge Review: Decoding AI Decisions for Pneumonia on Chest Radiographs," will showcase leading approaches to understanding how AI models identify pneumonia in chest radiographs—a critical step toward building trust and interpretability in medical AI applications.

This webinar is an opportunity to see live presentations and demonstrations by the top ranked teams, explore their innovative methodologies, and understand how explainable AI could reshape clinical diagnostics. Don’t miss out!

Now Available: NIH Seminar on
MIDRC's Role in Advancing Medical Imaging AI

Dr. Maryellen L. Giger’s seminar is now available on the NIH Office of Data Science Strategy YouTube channel!

In this engaging session, Dr. Giger, a pioneering expert in AI for medical imaging, explores how the Medical Imaging and Data Resource Center (MIDRC) is shaping the future of healthcare through cutting-edge AI technologies. Part of the NIH Data Sharing and Reuse Seminar Series, this seminar sheds light on MIDRC’s innovative contributions to AI research and clinical practice.

Key Takeaways from the Seminar:

The Power of Curated Datasets: Learn how real-world, curated datasets are crucial for building accurate and trustworthy AI systems in healthcare.

Driving Global Innovation: See how MIDRC’s open-access resources are fueling AI research and development on a global scale.

Addressing AI Challenges: Gain insights into MIDRC’s approaches to tackling data bias, model validation, and sustainable AI practices.

Transforming Diagnostics: Discover the transformative role of AI in enhancing diagnostics and enabling personalized patient care.

Don’t miss this opportunity to learn from a leader in the field about MIDRC’s mission and its impact on medical imaging AI.

Access the Seminar: Watch here
Explore the NIH Data Sharing and Reuse Seminar Series: Learn more