BraTS 2024 Cluster of Challenges (BraTS + Beyond-BraTS)

This document describes the experimental design of the 'International Brain Tumor Segmentation (BraTS) Cluster of Challenges', in partnership with the 'A.I. for Response Assessment in Neuro-Oncology' (AI-RANO) cooperative group.
 
Since 2012, the annual BraTS challenge has focused on the generation of a fair benchmarking environment and an associated dataset for the delineation of adult brain tumors. After 10 years, we conducted the RSNA-ASNR-MICCAI BraTS 2021 challenge, which spearheaded a partnership of MICCAI with two major clinical societies in the US (RSNA & ASNR) and contributed to extending the BraTS dataset to >2,000 cases. Building upon this effort, in 2023, we conducted the first BraTS Cluster of Challenges, with a substantially expanded dataset of >4,500 cases and a scope to address additional i) underserved patient populations, ii) tumor types, and iii) clinical concerns (e.g., missing data).
 
This year the focus of the BraTS 2024 Cluster of Challenges remains the generation of a common benchmarking environment, but with a further expanded 1) clinical relevance, 2) scope, and 3) dataset. Specifically, the BraTS 2024 Cluster of Challenges partners with the AI-RANO group to present newly proposed clinically relevant challenges, in a synergistic attempt to maximize the potential clinical impact of the innovative algorithmic contributions made by the participating teams. The scope extends further to address additional i) underserved populations (i.e., sub-Saharan Africa patients), ii) timepoints (i.e., pre- & post-treatment), iii) tumor types (e.g., meningioma), iv) modalities (i.e., histology samples), v) clinical concerns (e.g., missing data), and iv) technical considerations (e.g., generalizability). Finally, the BraTS 2024 datasets describes a further contribution to the community of additional well-curated manually-annotated cases, comprising a) MRI scans from 4,000 previously unseen patients & 280,000 histology samples.
 
The focus of each challenge (referred to as "task" here onwards - based on predefined terminology by the submission platform) of the BraTS 2024 Cluster of Challenges is to identify the current state-of-the-art algorithms for addressing (Task 1) Post-Treatment Adult Glioma, (Task 2) Post-treatment Intracranial Meningioma, (Task 3) Pre- and Post-Treatment Brain Metastases, (Task 4) Brain Glioma in the underserved sub-Saharan African patient population, (Task 5) Pre-Treatment Pediatric Tumor Patients in partnership with multiple related societies, (Task 6) Generalizability of Segmentation Methods Across (Pre-treated) Tumors, (Task 7) Evaluation of Augmentation Techniques, in partnership with FDA, (Task 8) MRI Synthesis, (Task 9) MRI Inpainting, as well as (Task 10) Assessing the Heterogeneous Histologic Landscape of Glioma. Worth highlighting that 6/10 tasks (1-3,6,7,10) are newly introduced tasks.
 
Details for each 'Task' are listed in the rest of this document. Notably, all data for tasks 1-9 are routine clinically acquired, multi-site multiparametric magnetic resonance imaging (mpMRI) scans of brain tumor patients. The BraTS 2024 challenge participants are able to obtain the training and validation data of the challenge at any point from the Synapse platform. These data will be used to develop, containerize, and evaluate their algorithms in unseen validation data until August 2024, when the organizers will stop accepting new submissions and evaluate the submitted algorithms in the hidden testing data. Ground truth reference annotations for all datasets are created and approved by expert neuroradiologists/neuropathologists for every subject included in the training, validation, and testing datasets to quantitatively evaluate the performance of the participating algorithms.

This document includes information on both BraTS and Beyond-BraTS events.