Machine learning-driven spleen imaging and genomics uncover a splenic connection to coronary artery disease.
Authors
Affiliations (18)
Affiliations (18)
- Harvard Medical School, Boston, MA 02115, USA.
- Cardiovascular Imaging Research Center, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114, USA.
- Artificial Intelligence in Medicine Program, Mass General Brigham, Boston, MA 02114, USA.
- Harvard T. H. Chan School of Public Health, Boston, MA 02115, USA.
- Program in Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA.
- Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA 02114, USA.
- Heart and Vascular Institute, Mass General Brigham, Boston, MA 02114, USA.
- Computational Precision Health Program, University of California, Berkeley, Berkeley, CA 94720, USA.
- Computational Precision Health Program, University of California, San Francisco, San Francisco, CA 94143, USA.
- Faculty of Medicine, KU Leuven, 3000 Leuven, Belgium.
- Data Sciences Platform, Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA.
- Layer Health, Boston, MA 02111, USA.
- Division of Cardiovascular Medicine, Brigham and Women's Hospital, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, MA 02115, USA.
- Cardiovascular Disease Initiative, Broad Institute of Harvard and MIT, Cambridge, MA 02142, USA.
- Google Ventures, Cambridge, MA 02142, USA.
- Clinical and Translational Epidemiology Unit, Massachusetts General Hospital, Boston, MA 02114, USA.
- Personalized Medicine, Mass General Brigham, Boston, MA 02114, USA.
- Amgen Inc., Thousand Oaks, CA 91320, USA.
Abstract
Despite advances in managing traditional risk factors, coronary artery disease (CAD) remains the leading cause of mortality. Circulating hematopoietic cells influence risk for CAD separately from traditional risk factors, but the role of a key regulating organ, the spleen, is unknown. The understudied spleen is a representation of the hematopoietic system optimally suited for unbiased radiologic investigations toward mechanistic insights. Here, we leveraged deep learning to extract 107 splenic radiomic features from abdominal magnetic resonance imaging (MRI) scans of 42,059 UK Biobank participants and of 2745 Mass General Brigham Biobank (MGBB) participants. Of these, 10 features from UK Biobank were associated with CAD. Genome-wide association analysis of CAD-associated features identified 219 loci, including 9p21. Variants at 9p21, the strongest yet mechanistically elusive CAD locus, were associated with splenic features such as run-length nonuniformity, reflecting heterogeneity of continuous texture regions. Research MRI findings were consistent internally, but external clinical validation highlighted challenges in translating analyses of abdominal MRI scans to routine clinical practice because of variability in imaging protocols and greater clinical heterogeneity among patients. Our study, combining deep learning with genomics, presents a framework to uncover potential splenic involvement in CAD and emphasizes translational gaps between research and clinical radiomics.