Search

Large epidemiological studies such as the UK Biobank (UKBB) or German National Cohort (NAKO) provide unprecedented health-related data of the general population aiming to better understand determinants of health and disease. As part of these studies, Magnetic Resonance Imaging (MRI) is performed in a subset of participants allowing for phenotypical and functional characterization of different organ systems. Due to the large amount of imaging data, automated image analysis is required, which can be performed using deep learning methods, e. g. for automated organ segmentation. In this paper we describe a computational pipeline for automated segmentation of abdominal organs on MRI data from 20,000 participants of UKBB and NAKO and provide results of the quality control process. We found that approx. 90% of data sets showed no relevant segmentation errors while relevant errors occurred in a varying proportion of data sets depending on the organ of interest. Image-derived features based on automated organ segmentations showed relevant deviations of varying degree in the presence of segmentation errors. These results show that large-scale, deep learning-based abdominal organ segmentation on MRI data is feasible with overall high accuracy, but visual quality control remains an important step ensuring the validity of down-stream analyses in large epidemiological imaging studies.

Purpose: To detail the rationale, design, and future perspective of implementing whole-body magnetic resonance (MR) imaging in the German National Cohort, a large multicentric population-based study. Materials and Methods: All institutional review boards approved the study, and informed consent is obtained before study enrollment. Participants are enrolled from a random sample of the general population at five dedicated imaging sites among 18 recruitment centers. MR imaging facilities are equipped with identical 3.0-T imager technology and use uniform MR protocols. Imager-specific hardware and software settings remained constant over the study period. On-site and centralized measures of image quality enable monitoring of completeness of the acquisitions and quality of each of the MR sequences. Certified radiologists read all MR imaging studies for presence of incidental findings according to predefined algorithms. Results: Over a 4-year period, six participants per day are examined at each center, totaling a final imaging cohort of approximately 30 000 participants. The MR imaging protocol is identical for each site and comprises a set of 12 native series to cover neurologic, cardiovascular, thoracoabdominal, and musculoskeletal imaging phenotypes totaling approximately 1 hour of imaging time. A dedicated analysis platform as part of a central imaging core incorporates a thin client-based integrative and modular data handling platform to enable multicentric off-site image reading for incidental findings. Scientific analysis will be pursued on a per-project hypothesis-driven basis. Conclusion: Population-based whole-body MR imaging as part of the German National Cohort will serve to compile a comprehensive image repository, will provide insight into physiologic variants and subclinical disease burden, and has the potential to enable identification of novel imaging biomarkers of risk.