For the estimated three million Americans with a form of heart failure where the heart pumps normally but the body never quite gets enough oxygen, diagnosis has always been more art than science. The condition, called heart failure with preserved ejection fraction (HFpEF), produces shortness of breath and fatigue. It does not always show up on the standard tests. And until recently, there were no drugs that clearly helped. That changed when GLP-1 and SGLT2 inhibitors started posting results that cardiologists could not ignore. Now the question is whether a new three-minute MRI add-on, developed at Cedars-Sinai Medical Center and published March 25 in Science Translational Medicine, can give doctors the metabolic readout they need to use those drugs well.
The test measures myocardial oxygen consumption — how efficiently the heart muscle uses oxygen — using a signal borrowed from brain imaging. Blood oxygen level-dependent (BOLD) MRI has been used for years in neuroimaging to track oxygenation changes in the brain. Hsin-Jung Yang and colleagues at Cedars-Sinai's Biomedical Imaging Research Institute adapted it for the heart, solving a harder problem: the heart never stops moving. They built motion correction into the imaging sequence, calibrated the signal against a physical model of how hemoglobin interacts with the MRI field, and produced numbers, not just brightness differences on an image.
"A cardiac MRI scan is a routine clinical procedure that people can order," Yang said in an interview with STAT News. "This is a three-minute add-on, without contrast agent, without breath-holding."
The team validated the approach in two stages. First in pigs, then in 22 patients who had suffered heart attacks. The MRI-derived oxygen consumption figures matched results from older invasive techniques — specifically, measurements taken via catheter in the coronary sinus, the vein that drains blood from the heart muscle. That correlation is the method's proof of concept: the test works well enough to replace a more burdensome procedure in some settings.
The clinical motivation is HFpEF, and it is not subtle. GLP-1 drugs like semaglutide (Ozempic, Wegovy) and SGLT2 inhibitors like empagliflozin (Jardiance) have both shown meaningful benefits in HFpEF populations in late-stage trials. Both drug classes are now in clinical guidelines as treatments for the metabolic abnormalities that underlie much of HFpEF, according to research published in JACC. The drugs work, or appear to work, by changing how the heart muscle metabolizes fuel. But knowing whether they are working in a given patient requires knowing something about the heart's metabolism to begin with — and that has historically meant a catheter procedure no routine cardiology practice can offer.
"GLP-1 drugs and SGLT2 drugs to treat that HFpEF population show very promising results," Yang said. "That is now in the clinical guidelines. And that's also the key trigger of why we're doing this. We wanted to directly see the metabolism in the heart."
The broader market the team is watching is metabolic syndrome: patients with obesity, diabetes, and hypertension who are at high risk of developing cardiac oxygenation abnormalities before they develop symptoms severe enough to trigger a diagnosis. Yang noted that earlier identification in those populations is where the test's greatest clinical impact could land.
The paper has real limits. Twenty-two patients is a standard early human validation size — it establishes that the method works, not that it changes outcomes. The test was studied in post-heart-attack patients, not in the HFpEF population where the metabolic angle is most commercially interesting. Large prospective trials in diverse patient groups will be needed before the method could become standard of care. The researchers are also adjacent to another Cedars-Sinai group doing metabolite imaging in mitochondria, suggesting this is one front in a broader effort rather than a standalone advance.
The competitive landscape for non-invasive cardiac metabolic imaging is active. A paper published in the Journal of Cardiovascular Magnetic Resonance in early 202501309-7/fulltext) described a separate approach using coronary sinus MR blood oximetry in pigs, and PET-based oxygen consumption measurements remain the clinical gold standard despite their radiation and cost. The Yang team's advantage is speed and hardware: BOLD MRI is available at most major hospitals, requires no tracers, and adds three minutes to an existing protocol.
What to watch next is straightforward: whether Cedars-Sinai pursues a larger clinical trial and whether HFpEF drug manufacturers show interest in pairing diagnostic readout with their therapies. The logic is intuitive — if a drug class works by improving cardiac metabolism, a test that measures cardiac metabolism could guide who gets the drug, whether they are responding, and how aggressively to dose. That is a compelling commercial story if the validation data hold at scale.
The source material is a STAT News Q&A with Yang and the Science Translational Medicine paper. The paper is paywalled; STAT News first reported the research on March 25, 2026.
