arxiv_cv 99% Match Challenge Submission / Technical Report Medical Imaging Researchers,Radiologists,Oncologists,AI Developers in Healthcare 6 days ago

Auto3DSeg for Brain Tumor Segmentation from 3D MRI in BraTS 2023 Challenge

computer-vision › medical-imaging

📄 Abstract

Abstract: In this work, we describe our solution to the BraTS 2023 cluster of challenges using Auto3DSeg from MONAI. We participated in all 5 segmentation challenges, and achieved the 1st place results in three of them: Brain Metastasis, Brain Meningioma, BraTS-Africa challenges, and the 2nd place results in the remaining two: Adult and Pediatic Glioma challenges.

Authors (4)

Andriy Myronenko

Dong Yang

Yufan He

Daguang Xu

Submitted

October 29, 2025

arXiv Category

cs.CV

arXiv PDF

Key Contributions

This work describes the successful application of Auto3DSeg from the MONAI framework to achieve top performance in the BraTS 2023 challenges. The solution secured 1st place in three segmentation challenges and 2nd place in two others, demonstrating the effectiveness of automated segmentation tools for complex medical imaging tasks.

Business Value

Accelerates the diagnosis and treatment planning for brain tumors by providing highly accurate and automated segmentation, potentially improving patient outcomes and reducing radiologist workload.

Paper Metadata

Innovation Type

Application of Existing Framework

Deployment Feasibility

High, as it leverages a well-established framework (MONAI) designed for medical imaging applications.

Limitations Addressed

Addresses the challenge of achieving state-of-the-art performance in complex medical image segmentation tasks, particularly for brain tumors.

Technical Tags

3D MRI segmentationbrain tumor segmentationBraTS 2023Auto3DSegMONAImedical imaging challengesdeep learning segmentation

Research Topics

Medical Image AnalysisBrain Tumor SegmentationDeep LearningAutomated Segmentation

Methods & Architectures

Auto3DSegMONAI framework 3D Segmentation Networks (implied by MONAI)

Applications & Tasks

Healthcare Medical Imaging Neuroscience Oncology Accurate brain tumor segmentationPerformance in medical imaging challenges Brain Tumor Segmentation

Datasets & Benchmarks

Datasets

BraTS 2023 dataset

Benchmarks

1st place in Brain Metastasis, Brain Meningioma, BraTS-Africa challenges • 2nd place in Adult and Pediatric Glioma challenges

Related Fields

Medical ImagingRadiologyDeep LearningComputer VisionNeuro-oncology

Keywords

brain tumor segmentation3D MRIBraTS 2023MONAIAuto3DSegmedical imagingdeep learningsegmentationradiologyoncology

Academic Context

#Medical Image Analysis#Brain Tumor Segmentation#Deep Learning#Automated Segmentation

Technology Stack

Frameworks & Libraries

MONAI

Commercial Potential

Potential Products

Automated Brain Tumor Segmentation Software

Target Industries

HealthcarePharmaceuticalsMedical Device Manufacturing

Use Case Examples

Assisting radiologists in segmenting brain tumors for treatment planningQuantifying tumor volume and changes over timeSupporting clinical trial analysis

Competitive Edge

Achieved state-of-the-art results in a highly competitive challenge, demonstrating the effectiveness of the Auto3DSeg approach within the MONAI ecosystem.

Market Opportunity

Large and growing market for AI in medical diagnostics.

Revenue Models

Software licensing to hospitals and research institutionsservice contracts.

Resource Requirements

Compute Needs

High, requires GPUs for training and inference of 3D deep learning models.

Data Requirements

3D MRI scans (e.g., T1, T1ce, T2, FLAIR) with expert annotations.

Deployment Constraints

Integration into clinical workflows, regulatory approval (e.g., FDA).

Scalability

Scalable with appropriate computational infrastructure.

Regulatory Considerations

Medical device regulationsdata privacy (HIPAA).

Production Readiness

Maturity Level

Research / Applied

Time to Market

2-4 years

View Full Paper Back to Papers