Methodology for a fully automated pipeline of AI-based body composition tools for abdominal CT

• DOI: 10.1007/s00261-025-04951-7
• Publication Date: 2025-04-28T04:25:28Z
• Authors: John W. Garrett; Perry J. Pickhardt; Ronald M. Summers
• Keywords: Normalization; DICOM

Abstract Accurate, reproducible body composition analysis from abdominal computed tomography (CT) images is critical for both clinical research and patient care. We present a fully automated, artificial intelligence (AI)-based pipeline that streamlines the entire process—from data normalization anatomical landmarking to automated tissue segmentation quantitative biomarker extraction. Our methodology ensures standardized inputs robust models compute volumetric, density, cross-sectional area metrics range of organs tissues. Additionally, we capture selected DICOM header fields enable downstream scan parameters facilitate correction acquisition-related variability. By emphasizing portability compatibility across different scanner types, image protocols, computational environments, ensure broad applicability our framework. This toolkit basis Opportunistic Screening Consortium in Abdominal Radiology (OSCAR) has been shown be versatile, large multi-center studies. Graphical abstract