Published: 14-11-2024 14:44 | Updated: 14-11-2024 14:44

AI expert wins big in medical imaging competition

A computer generated 3D-visualization of a brain from a patient with glioma where the tumor is red, the necrotic core is blue and the edema is green.
This 3D visualization demonstrates the substantial impact of a glioma on brain structures. Enhancing tumor (red), necrotic core (blue), and edema (green) comprise the tumor mass, which extends throughout both hemispheres.

Mehdi Astaraki, a specialist in artificial intelligence (AI) for medical imaging, has achieved remarkable success in the MICCAI BraTS Challenges. He secured first place in the Pediatric Brain Tumor Segmentation and second place in the Meningioma RT Segmentation categories.

Medical Image Computing and Computer-Assisted Intervention (MICCAI) BraTS Challenge is a prestigious event where experts from around the world compete to develop the best AI models for medical imaging.

Mehdi Astaraki is standing by a desk with a computer screen showing a 3D image of a brain.
PhD Mehdi Astaraki. Photo: N/A

“It’s truly an honor to receive these awards, says Mehdi Astaraki, postdoctoral researcher at the Department of Oncology-Pathology in an interview performed by Farhad Abtahi. We all know how AI techniques are advancing rapidly, and many aspects of computational methods are already changing. However, unlike entertainment and general applications of AI, in health care fields thorough evaluation and validation of these techniques are quite vital. In this context, challenges play an important role. The MICCAI challenges are held to the highest quality standard to benchmark the developed solutions for medical image processing/analysis applications. Participating in these challenges provides invaluable feedback and helps the whole community, including us, to identify the pros and cons of the current solutions. More importantly, it is an opportunity to keep ourselves updated with state-of-the-art methods, ultimately improving our models to contribute to better clinical tools and patient care. By participating in these challenges, I tried to validate our research efforts and demonstrate our progress in this competitive field.”

Mehdi’s success in the competition highlights the importance of creating AI models that are not only accurate but also adaptable to different medical scenarios. His work helps improve tools used by doctors to diagnose and treat patients.

Mehdi is also a strong supporter of open science, which means sharing research findings, codes, and data openly. This approach helps other researchers build on existing work, leading to faster and more reliable advancements in medical technology.

At the Stockholm Medical Artificial Intelligence and Learning Environments (SMAILE) core facility at Karolinska Institutet and Karolinska University Hospital, Mehdi has access to advanced tools and a collaborative environment. This support has been crucial in developing and refining his AI models.

Mehdi Astaraki’s achievements in the MICCAI BraTS Challenges demonstrate the potential of AI in improving healthcare. His dedication to open science and collaboration is paving the way for better medical tools and patient care.