[論文レビュー] MHub.ai: A Simple, Standardized, and Reproducible Platform for AI Models in Medical Imaging

Leonard Nürnberg, Dennis Bontempi+18 more|arXiv (Cornell University)|Jan 15, 2026

Radiomics and Machine Learning in Medical Imaging被引用数 0

ひとこと要約

tldr: MHub.ai introduces an open-source, container-based platform that standardizes access to AI models in medical imaging, adding reference data, unified interfaces, and dashboards to promote reproducibility and benchmarking.

ABSTRACT

Artificial intelligence (AI) has the potential to transform medical imaging by automating image analysis and accelerating clinical research. However, research and clinical use are limited by the wide variety of AI implementations and architectures, inconsistent documentation, and reproducibility issues. Here, we introduce MHub$.$ai, an open-source, container-based platform that standardizes access to AI models with minimal configuration, promoting accessibility and reproducibility in medical imaging. MHub$.$ai packages models from peer-reviewed publications into standardized containers that support direct processing of DICOM and other formats, provide a unified application interface, and embed structured metadata. Each model is accompanied by publicly available reference data that can be used to confirm model operation. MHub$.$ai includes an initial set of state-of-the-art segmentation, prediction, and feature extraction models for different modalities. The modular framework enables adaptation of any model and supports community contributions. We demonstrate the utility of the platform in a clinical use case through comparative evaluation of lung segmentation models. To further strengthen transparency and reproducibility, we publicly release the generated segmentations and evaluation metrics and provide interactive dashboards that allow readers to inspect individual cases and reproduce or extend our analysis. By simplifying model use, MHub$.$ai enables side-by-side benchmarking with identical execution commands and standardized outputs, and lowers the barrier to clinical translation.

研究の動機と目的

• Motivate the need for standardized, reproducible AI model access in medical imaging.
• Describe an open-source platform that packages models into standardized containers.
• Demonstrate how the platform supports direct processing of DICOM and other formats.
• Explain how embedded metadata and reference data enable transparent evaluation and reuse.
• Showcase a clinical use case to illustrate benchmarking and reproducibility benefits.

提案手法

• Package models from peer-reviewed publications into standardized containers.
• Provide a unified application interface and support direct processing of DICOM and other formats.
• Embed structured metadata and accompany each model with publicly available reference data.
• Enable community contributions and modular adaptation of any model.

実験結果

リサーチクエスチョン

• RQ1 How does MHub.ai standardize access to AI models in medical imaging with minimal configuration?
• RQ2 Can standardized containers and reference data enable transparent benchmarking and reproducibility across models?
• RQ3 What is the impact of a unified interface and dashboards on model evaluation and reuse in clinical settings?
• RQ4 To what extent can community contributions be integrated into the platform to extend model coverage?

主な発見

• Models from peer-reviewed publications are packaged into standardized containers.
• The platform supports direct processing of DICOM and other formats with a unified interface.
• Each model includes embedded metadata and publicly available reference data for verification.
• Generated segmentations and evaluation metrics are publicly released to promote transparency.
• Interactive dashboards allow inspection of individual cases and reproduction or extension of analyses.
• The framework enables side-by-side benchmarking with identical execution commands and outputs.