MICCAI 2014 will provide an excellent opportunity for a day long cluster of events in Brain Tumor computation (September 14, 2014), composed of a Workshop, and radiologic and pathology image processing Challenges to discuss and showcase the value of open science in addressing the challenges of Big Data in the context of brain cancer. 

1. Workshop: Computational Precisian Medicine
2. Imaging Challenge: BRATS 2014
3. Digital Pathology Challenge

Workshop: Computational Precision Medicine

The goal of the workshop is to present and discuss basic requirements and current resources for open science development of systems in support of computational precision medicine in brain tumor diagnosis and treatment planning.  This half-day workshop (8:30 am – 12:30 pm) will include invited talks as well as proffered oral and poster presentations. 

Imaging Challenge: BRATS 2014

The goal of the imaging Challenge in multi-modal Brain Tumor Image Segmentation (BRATS) is to gauge the current state-of-the-art in automated brain tumor segmentation and compare different methods.  In addition, BRATS 2014 will include sub-challenges on analysis of longitudinal data sets and classification of tumor grades.  BRATS-2014 Challenge will contain three sub-challenges:

Sub-Challenge 1: Segmentation – Automatic Evaluation of a collection of over 50 multi-parametric MRI cases.

Sub-Challenge 2: Longitudinal Evaluation – Segmentation of time series images.

Sub-Challenge 3: Classification – Automatic classification into one of the three classes of Low Grade II, Low Grade III, and High Grade IV (glioblastoma multiforme or GBM).

For more information please visit: http://www.braintumorsegmentation.org

Brain Tumor Digital Pathology Challenge

As the technology for digital imaging has advanced, there is now increasing use of digital images ("virtual slides") for pathologic analysis of surgical specimens.  Automated tumor segmentation, by defining tumor regions with critical histologic features has the potential to increase both the speed and accuracy of diagnosis by pathologists and/or computer software.  There will be two sub-challenges in the proposed Brain Tumor Digital Pathology Challenge:

Sub-Challenge 1: Classification - Automated classification of LGG and GBM from a collection of 30+ high-resolution digital pathology slides.

Sub-Challenge 2: Segmentation – Automated segmentation of necrotic and normal brain regions on regions of digital pathology slides from a collection of 20+ GBM cases.

Source of Test Data: All cases selected for the Test phase of each challenge will be drawn from the Cancer Genome Atlas (TCGA) and made available for download through this site (TCIA).

For more information please visit:  http://pais.bmi.stonybrookmedicine.edu

Manuscript and Poster Submissions

Participants in Brain Tumor Challenges will be required to submit short manuscripts outlining their approach and preliminary results on the training data in July.  Respective Challenge organizers will review all submissions and submitters will be notified of the results. Challenge Test (Phase 3) results will be announced at the workshop and the top three scoring teams will be invited to give 12 min presentations of their methodology and results.  All participants will be invited to submit posters for exhibition at MICCAI 2014.  Reports may be prepared by the respective organizers of each Challenge after MICCAI 2014 for submission to appropriate peer reviewed journals.

More information about these events, including links to specific Brain Tumor Challenge websites, for more details about each challenge, will be provided here by early May 2014.