MAIA: a collaborative medical AI platform for integrated healthcare innovation

Abstract

Abstract The integration of Artificial Intelligence (AI) into clinical workflows requires collaborative platforms that bridge the gap between innovation and healthcare applications. This paper introduces MAIA (Medical Artificial Intelligence Assistant), an open-source platform designed to enable collaboration among clinicians, researchers, and AI developers. Built on Kubernetes, MAIA provides a modular, scalable environment with tools for data management, model development, annotation, deployment, and clinical feedback. Key features include project isolation, CI/CD automation, and integration with high-computing infrastructures and clinical workflows. MAIA supports real-world use cases in medical imaging AI, with deployments in academic and clinical environments. By promoting collaboration and interoperability, MAIA accelerates the translation of AI research into clinical solutions while promoting reproducibility, transparency, and user-centered design. We demonstrate MAIA in two end-to-end development and deployment projects at KTH Royal Institute of Technology and Karolinska University Hospital, focusing on bone metastasis segmentation in CT and brain metastasis segmentation in MRI.

Ähnliche Arbeiten