A new study led by investigators from Mass General Brigham reveals that certain features in the spleen hidden within imaging scans may signal a person's risk of coronary artery disease (CAD), suggesting targets for prevention and treatment. The findings are published in Science Translational Medicine .
"Great advances have been made against the classical risk factors for CAD. LDL cholesterol can be lowered very effectively today, for example, yet substantial residual risk remains, and CAD is still the world's leading cause of death," said co-senior author Zhi Yu, MB, PhD, an investigator in the Clinical and Translational Epidemiology Unit and Center for Genomic Medicine at Mass General Brigham. Yu is also an associate member at the Broad Institute of MIT and Harvard. "Evidence increasingly ties the blood-forming system to heart disease, and the spleen is a central hub of that system, storing and filtering blood and producing immune and inflammatory cells. But the spleen is hard to assess through routine tests, so imaging is a powerful way to see nuanced changes in the organ and ask whether they reflect, or even forecast, disease."
To investigate, Yu and colleagues leveraged artificial intelligence (AI) tools, abdominal imaging, and clinical outcomes data to identify CAD-relevant splenic features. They also performed genomic analyses to see if known CAD-associated genes are linked to splenic features seen in their imaging data.
The team's research drew on information from 42,059 participants in the UK Biobank. Ten of 107 splenic features observed on abdominal imaging scans were associated with CAD. Also, genome-wide association analyses identified genes linked to both splenic features and CAD. These genes were often involved in processes such as inflammation, smooth muscle cell function, hypertension, and fat cell formation. Many of the associated variants fell in non-coding, regulatory regions of the genome. Two variants at the genetic locus 9p21 on chromosome 9 were associated with a nonuniform, irregular texture in the spleen, increased odds of CAD, independent of blood pressure, cholesterol, and other conventional risk factors.
To see how these general-population findings hold up in patients, the team tested whether the UK Biobank patterns would also appear in people imaged as part of routine clinical care. In 2,745 patients from the Mass General Brigham Biobank, most associations did not carry over. Research cohorts like the UK Biobank tend to be healthier and scanned under uniform protocols, whereas clinical patients tend to have more complex medical histories and are imaged under widely varying protocols — so the two settings may capture genuinely different biological states.
"These findings shed light on novel mechanisms linking the spleen to CAD, providing potential targets for therapeutic intervention to address this unexplored axis," said lead author Meghana Kamineni, MD, an internal medicine resident at Mass General Brigham.
Authorship: In addition to Yu and Kamineni, Mass General Brigham authors include Vineet Raghu, Haodong Tian, Buu Truong, Romit Bhattacharya, Peter Libby, Patrick T. Ellinor and Whitney Hornsby. Additional authors include Pradeep Natarajan, Zhanqing Hua, Ahmed Alaa, Art Schuermans, Sam Friedman, Christopher Reeder, Mahnaz Maddah and Anthony Philippakis.
Disclosures: A list of author disclosures can be found in the paper.
Funding: This research was funded in part by the Harvard-MIT Health Sciences and Technology program; the Harvard Medical School Office of Scholarly Engagement; the American Heart Association (Career Development Award 935176; 18SFRN34110082); Harvard Catalyst K12 Award; Bayer AG; Fondation Leducq (14CVD01; TNE-18CVD04); the National Institutes of Health (1K01HL168231, 1R01HL092577, K24HL105780, 1R01HL134892, 1R01HL163099-01, R01AG063839, R01HL151627, R01HL157073, R01HL166538, 1K99HG012956-01, R01HL142711, R01HL127564, R01HL148050, R01HL151283, R01HL148565, R01HL135242, R01HL151152, and R01DK125782); the RRM Charitable Fund; the Simard Fund; and Massachusetts General Hospital through the Paul and Phyllis Fireman Endowed Chair in Vascular Medicine.
Paper cited: Kamineni M et al. "Machine learning–driven spleen imaging and genomics uncover a splenic connection to coronary artery disease," Science Translational Medicine. DOI: 10.1126/scitranslmed.aeh2517
