Heart Scan Warning: Low Flow Reserve on PET Quadruples Risk of Death and Heart Failure

A single number from a heart scan may predict who will die, suffer a heart attack, be hospitalized for heart failure, or experience a dangerous heart rhythm — even when traditional measures of ischemia look reassuring. That is the central message of a new study from Nippon Medical School in Tokyo, published in the European Journal of Nuclear Medicine and Molecular Imaging. Using nitrogen-13 ammonia positron emission tomography, or ¹³N-ammonia PET, researchers found that patients with chronic coronary syndrome whose global myocardial flow reserve fell below 2.0 faced more than four times the risk of a hard cardiac event compared with patients above that threshold, and the association held even after accounting for the amount of relative ischemia visible on the same scan.

Myocardial flow reserve, or MFR, is a quantitative measure of how well the heart’s microvascular plumbing works under stress. It is calculated by dividing myocardial blood flow during pharmacological stress — typically induced with a vasodilator such as adenosine or dipyridamole — by blood flow at rest. A healthy coronary circulation can multiply its flow several-fold when demand rises; an MFR of 2.0 means the heart can only double its perfusion, a level widely regarded as abnormal. Unlike visual perfusion images, which reveal regional defects only when blood flow disparities are severe enough to create contrast between territories, quantitative MFR captures diffuse and microvascular disease that conventional relative imaging can miss entirely.

The ¹³N-ammonia tracer occupies a special place in this field. Short-lived and demanding to produce — it requires an on-site cyclotron because of its roughly ten-minute half-life — nitrogen-13 ammonia nonetheless offers excellent extraction characteristics and a long track record in quantitative myocardial perfusion imaging. The Japanese team, led by Yoshimitsu Fukushima of the Department of Radiology, retrospectively analyzed consecutive adults with suspected or established chronic coronary syndrome who underwent ¹³N-ammonia PET/CT at their center. The cohort spanned two generations of PET scanners, a detail the investigators built directly into their statistical design by stratifying their primary model by scanner generation.

Of the 168 patients studied, 102 — just over 60 percent — had a global MFR below 2.0. These patients showed a distinctive hemodynamic fingerprint: higher resting blood flow and lower stress blood flow than their counterparts with preserved reserve. That pattern is physiologically telling. Elevated resting flow often reflects compensatory mechanisms or altered hemodynamics, while a blunted stress response signals that the coronary microcirculation cannot augment delivery when the myocardium demands it. The gap between those two numbers is precisely what the reserve metric compresses into a single value.

Over a median follow-up of 3.5 years, the study’s primary endpoint — a deliberately hard composite that excluded revascularization procedures — occurred in 28 of the 102 patients with low MFR, a rate of 27.5 percent, versus only 5 of the 66 patients with preserved reserve, or 7.6 percent. The composite counted all-cause death, non-fatal myocardial infarction, heart-failure hospitalization, and malignant ventricular arrhythmia. By excluding revascularization, the endpoint avoids the circularity that can plague imaging-outcome studies, in which a scan finding triggers an intervention that then counts as an event. The Kaplan-Meier separation between the two groups was stark, with a log-rank P value below 0.001.

The pre-specified Cox proportional hazards model adjusted for age, sex, and the summed difference score — a standard semi-quantitative index of relative ischemia derived from the perfusion images themselves. Even with that adjustment, low MFR remained independently associated with the composite endpoint, carrying a hazard ratio of approximately 4.52 with a 95 percent confidence interval of 1.71 to 11.92. The independence from the summed difference score is the study’s most consequential finding: it means the prognostic signal in MFR is not simply a restatement of visible ischemia but reflects something additional, most plausibly the health of the microvasculature and the myocardium itself.

Because the composite endpoint was weighted toward heart-failure hospitalization, the investigators performed a sensitivity analysis excluding patients who already had heart failure at baseline. The association survived: the hazard ratio was 4.17 with a P value of 0.011. This matters because it suggests reduced flow reserve is not merely a marker of established pump dysfunction. It aligns with a growing body of evidence linking coronary microvascular dysfunction to the later development of heart failure, including heart failure with preserved ejection fraction, a condition in which the muscle stiffens and fills poorly despite a seemingly normal pumping fraction. Prior work from groups including Taqueti and colleagues has shown that global coronary flow reserve predicts adverse events independently of angiographic stenosis severity, and meta-analyses have consolidated the prognostic value of PET-derived flow measures across tracers and populations.

The study also waded into a technical controversy that has divided the cardiac PET community: whether raw MFR should be corrected for the rate-pressure product, or RPP, the product of heart rate and systolic blood pressure that serves as a proxy for myocardial oxygen demand. Proponents of RPP correction argue that it adjusts for hemodynamic variability between rest and stress acquisitions; critics counter that it can distort the physiological meaning of the reserve and obscure genuine microvascular impairment. In this cohort, RPP correction reclassified 52 patients — 31 percent of the sample — across the 2.0 threshold, and the movement was predominantly downward, meaning correction pushed patients into the abnormal category. Event rates told a coherent story: patients reclassified downward had a 7.0 percent event rate, while those impaired on both the uncorrected and corrected values had a 29.0 percent rate, with a four-group log-rank P value of 0.003.

Those reclassification figures carry practical weight. If nearly a third of patients change risk category depending on whether a correction is applied, laboratories must be transparent about which value they report. The authors conclude that the uncorrected global MFR should be the primary reported metric, with any RPP-corrected value explicitly labeled as such. Their data suggest that uncorrected MFR below 2.0 identifies higher-risk patients independently of relative ischemia, and that correction, while informative, should supplement rather than replace the raw physiological measure. This position echoes recent comparative studies, including work from Houston and elsewhere examining corrected versus uncorrected reserve in patients without obstructive coronary disease, and a 2026 multicenter analysis confirming MFR’s prognostic value for mortality in ¹³N-ammonia imaging.

Limitations are worth keeping in view. The study was retrospective and single-center, and its cohort of 168 patients, while adequate for the pre-specified model, came from a Japanese population imaged across two scanner generations — a heterogeneity the authors addressed statistically but which nonetheless reflects real-world practice rather than a controlled trial. The endpoint composite, though hard, was dominated by heart-failure hospitalizations, so the findings speak most directly to heart-failure-related risk rather than to sudden death or arrhythmia in isolation. Still, the message for clinicians is concrete: when a quantitative ¹³N-ammonia PET study returns a global flow reserve below 2.0, that number deserves attention in its own right, not merely as a footnote to the perfusion images. For patients with chronic coronary syndrome, the microcirculation’s reserve capacity may be one of the most powerful predictors of what comes next.

Subject of Research: Prognostic value of ¹³N-ammonia PET myocardial flow reserve in chronic coronary syndrome

Article Title: Reduced ¹³N-Ammonia PET Myocardial Flow Reserve and Heart-Failure–Related Risk Independent of Relative Ischemia in Chronic Coronary Syndrome

Article References: Fukushima, Y., Kobayashi, R., Yodogawa, K., Shimizu, S., Imai, S., Hayashi, H., & Kumita, S. (2026). Reduced ¹³N-Ammonia PET Myocardial Flow Reserve and Heart-Failure–Related Risk Independent of Relative Ischemia in Chronic Coronary Syndrome. European Journal of Nuclear Medicine and Molecular Imaging. https://doi.org/10.1007/s00259-026-08183-9

Image Credits: AI Generated

DOI: 10.1007/s00259-026-08183-9

Keywords: myocardial flow reserve, ¹³N-ammonia PET, chronic coronary syndrome, heart failure, coronary microvascular dysfunction, rate-pressure product, myocardial perfusion imaging, risk stratification, cardiac PET, prognosis, ischemia, nuclear cardiology