ECG-Free Cardiac Gating From Routine FDG PET Works in Danish Study

Researchers in the Capital Region of Denmark reported that data-driven gating, a method that needs no ECG leads or other external device, extracted a cardiac motion signal from 138 of 169 routine [18F]FDG PET/CT examinations, or 81.7%. The study was published online in Physics in Medicine and Biology on October 7, 2026, and appeared in the October 14, 2026, issue (vol. 71, issue 19, article 195006). For departments already scanning cancer patients on cardiotoxic chemotherapy, the idea is to get a heartbeat, and possibly an ejection fraction, out of data they are already collecting.

The paper is a feasibility study, and its authors do not pitch it as anything more. According to the study abstract on the Capital Region of Denmark research portal, left ventricular ejection fraction (LVEF) was estimated in 101 patients, reported as 60%.

Why chemotherapy cardiotoxicity monitoring is the target

Chemotherapy-induced cardiotoxicity can lead to irreversible heart failure. LVEF is routinely monitored for that reason, but conventional assessment requires dedicated cardiac imaging and clinical resources.

If part of that surveillance could someday come from the FDG scan itself, the work would shift rather than disappear. In principle, there would be no extra acquisition. Instead, a processing task would fall to physics or software, and the physician would get one more number to interpret. That is why the agreement data matter more than the gating success rate.

Fludeoxyglucose F 18 Injection is an FDA-approved PET agent. One manufacturer’s U.S. prescribing information on DailyMed lists oncology (assessing abnormal glucose metabolism to evaluate malignancy), cardiology (identifying left ventricular myocardium with residual glucose metabolism and reversible loss of systolic function in coronary artery disease with LV dysfunction, used with perfusion imaging) and neurology (identifying epileptic foci). LVEF estimation and cardiotoxicity monitoring are not among the labeled uses. The approach in this study is software applied to existing emission data, tested as a research method for feasibility.

How data-driven gating works on routine FDG PET

The primary aim was extracting cardiac motion. LVEF was secondary. The team, with Nanna Overbeck of the Department of Clinical Physiology and Nuclear Medicine as corresponding author, applied anatomical masking and frequency-domain filtering to histo images, then ran the method across the 169 examinations.

“Deviceless” carries the practical weight here. Because the gating signal is derived from the emission data rather than from external hardware, the method can in principle be applied after the fact, which is the retrospective use the authors describe.

On the primary aim, the results are encouraging. The pulse frequencies were physiologically plausible and agreed well with the reference exams. Recovering a heartbeat from a routine cancer protocol, without hardware, in better than four of five exams is a real technical result.

The paper’s title specifies long-axial field-of-view (LAFOV) PET, and LAFOV appears in its keywords. The abstract itself refers only to routine [18F]FDG PET and names no scanner model or vendor. Whether the method carries over to standard axial field-of-view systems is an open question.

Where the LVEF estimates fall short

The LVEF step depends on something the gating step does not: enough FDG uptake in the myocardium for software to estimate ejection fraction. Only 101 patients (60%) met that bar, and LVEF was estimated in all 101.

The study’s counts mix units. The 169 figure refers to examinations and the 101 figure to patients, and the number of unique patients behind the 169 exams is not given. The two numbers cannot be read as a simple fraction of one another.

Each patient’s data-driven gating LVEF was compared with that patient’s own clinical reference test, either echocardiography or multigated acquisition (MUGA) scanning. Mean bias was +3.4% against echocardiography and −2.6% against MUGA. The authors describe “good agreement with established clinical reference methods” and say agreement was strongest with echocardiography. They also say the limits of agreement were too wide for the methods to be used interchangeably in the clinic.

A small average bias can hide large disagreement in individual patients, and in cardiotoxicity surveillance the individual patient’s measurement is what counts. The authors call the method a promising foundation for retrospective cardiac function assessment that could complement standard LVEF evaluation.

What the study leaves open

The comparison is cross-sectional: one PET-derived LVEF against one reference measurement. Cardiotoxicity monitoring is a longitudinal task, and whether the method can detect a decline from baseline in the same patient has not been tested in this study.

The numerical limits of agreement, the interchangeability threshold, the split between echocardiography and MUGA comparisons and the time between PET and reference exams all matter for interpreting the results. A gap of weeks during cardiotoxic therapy would be significant.

The reasons 31 exams failed gating, and why about 40% of patients had too little myocardial uptake for LVEF estimation, are unknown. Co-authors Paul Schleyer, Jorge Cabello and Anders Bertil Rodell are listed without affiliations, and funding and any industry involvement are not stated.

The study that would move this from add-on to something clinicians could act on is a serial one: the same patients, scanned before and during chemotherapy, with the PET-derived ejection fraction tracking a decline the reference test also sees.

About the sources: Study figures and author statements come from the abstract posted on the Capital Region of Denmark research portal (PMID 42697242), which also gives the online and issue publication dates. The full text, including methods and agreement statistics, was not reviewed. The DailyMed label is background on FDG’s approved uses and does not refer to this study.