A Simple ECG Reading May Hint at a Hidden Heart Tumor, Study Finds

A routine electrocardiogram, the most common heart test in medicine, may carry a subtle clue to one of its rarest and most dangerous findings: an atrial myxoma. In a retrospective cohort study published in Clinical Research in Cardiology, researchers in Münster, Germany, analyzed standard 12-lead ECGs from 39 patients with histologically confirmed atrial myxomas and 44 age- and sex-matched controls, and found that patients carrying these benign but potentially lethal tumors showed a distinctly lower P-wave voltage in lead I, together with a relatively short P-wave duration. The finding, if confirmed, could turn an already universal test into an early, low-cost flag for a tumor that is too often discovered only after it has caused a stroke or blocked blood flow to a limb.

Atrial myxoma is the most common primary cardiac neoplasm, accounting for roughly half of all benign primary heart tumors, and it typically arises in the left atrium of women between the fourth and sixth decades of life. Although the overall prevalence of intracardiac tumors is estimated at less than 0.3 percent, their clinical weight is considerable: they can obstruct blood flow, shed emboli into the systemic circulation, and trigger arrhythmias. Embolic events have been reported in up to 30 to 40 percent of cases, predominantly systemic for left atrial lesions and pulmonary for right atrial ones. Diagnosis today rests on transthoracic or transesophageal echocardiography, with cardiac magnetic resonance imaging reserved for tissue characterization and surgical planning. Yet diagnostic delay is still described in the current literature, partly because the ECG, often the first cardiovascular test a patient receives, does not show obvious abnormalities.

The new analysis set out to ask whether a closer look at the P-wave, the small deflection on the ECG that reflects atrial depolarization, might reveal what standard interpretation misses. P-wave indices are established markers of atrial structural and electrical remodeling, and have been linked to atrial cardiomyopathy, interatrial conduction delay, and elevated risks of atrial fibrillation and cardioembolic stroke. Data on their behavior in the specific context of atrial tumors, however, has been scarce and largely limited to case reports and small series. The authors therefore conducted a retrospective, single-center analysis comparing patients with histologically confirmed atrial myxomas against a reference group selected by cardiologists on strict criteria: no known significant cardiac disease, no medication known to influence ECG morphology or cardiac conduction, and no cardiothoracic disease, including lung disease, that could alter ECG appearance.

The cohort comprised 83 patients. The tumor group included 39 patients whose myxomas were confirmed by histopathological examination of the surgical specimen; 34 tumors, or 87.2 percent, sat in the left atrium and 5, or 12.8 percent, in the right atrium, typically attached to the interatrial septum near the fossa ovalis. Median maximal tumor diameter was 4.1 centimeters, with a range from 0.5 to 9.0 centimeters. Median age in the tumor group was 61.0 years, and 74.4 percent of these patients were women. Age, sex, body mass index, heart failure, arterial hypertension, valve disease, prior stroke, and diabetes were evenly distributed between the groups. Coronary artery disease, dyslipidemia, and a positive smoking history, however, were significantly more prevalent among tumor patients, a residual imbalance the investigators addressed in a sensitivity analysis.

Seven P-wave indices were extracted from standard 12-lead resting ECGs, with measurements performed manually by two cardiologists: P-wave duration, P-wave voltage in lead I, P-wave morphology in lead V1, P-wave axis, P-wave area in lead II, the MVP (morphology-voltage-P-wave duration) score, and P-wave dispersion, defined as the difference between the maximum and minimum P-wave duration across all leads. Two indices separated the groups. Patients with myxomas had a significantly shorter P-wave duration, a median of 100 milliseconds with an interquartile range of 80 to 100, versus 100 milliseconds with an interquartile range of 100 to 120 in controls, with p = 0.005. P-wave voltage in lead I was likewise significantly lower in the tumor group, at 0.07 millivolts (interquartile range 0.05 to 0.08) versus 0.09 millivolts (interquartile range 0.06 to 0.10), with p < 0.001. Morphology in V1, axis, area, MVP score, and dispersion showed no statistically significant differences.

In univariable logistic regression, both P-wave duration (odds ratio 0.96 per millisecond; 95 percent confidence interval 0.94 to 0.99; p = 0.010) and P-wave voltage in lead I (odds ratio 0.73 per 0.01 millivolts; 95 percent confidence interval 0.61 to 0.88; p < 0.001) were each significantly associated with the presence of an atrial tumor. When the two were entered together, only voltage remained an independent predictor (odds ratio 0.76 per 0.01 millivolts; 95 percent confidence interval 0.63 to 0.91; p = 0.003), while duration was attenuated to borderline significance (odds ratio 0.97 per millisecond; 95 percent confidence interval 0.95 to 1.00; p = 0.078), an attenuation the authors attribute to collinearity between the two measurements.

The study also uncovered a size-dependent signal. Within the tumor group, a larger maximal tumor diameter was associated with lower P-wave voltage in lead I (Spearman ρ = −0.32; p = 0.044), an inverse relationship that held in a linear Pearson sensitivity analysis (r = −0.32; p = 0.047) but lost significance when the two most extreme tumor diameters were excluded (ρ = −0.28; p = 0.092), indicating a modest effect. Tumor size showed no relationship with P-wave duration or any other P-wave index. When patients were stratified by a maximal diameter of 4 centimeters, P-wave voltage decreased stepwise from the reference group (0.09 millivolts) to patients with tumors below 4 centimeters (0.08 millivolts) and to those with tumors of 4 centimeters or more (0.05 millivolts; p < 0.001). P-wave duration, by contrast, was identical in both size strata and differed only from the reference group. The authors propose a mass-effect hypothesis: a tumor that occupies or mechanically replaces functional atrial myocardium may simultaneously reduce the effective depolarizing mass, lowering P-wave amplitude, and shorten the atrial activation pathway by reducing functional atrial volume, shortening P-wave duration. They caution that this mechanism is hypothesis-generating and requires confirmation in larger cohorts with standardized volumetric imaging, and that subclinical pericardial effusion, a recognized accompaniment of atrial myxomas, may also contribute to generalized voltage reduction.

Diagnostic performance was quantified with receiver operating characteristic analysis, which the authors explicitly label exploratory because cut-offs were derived and tested in the same dataset. P-wave voltage in lead I separated patients best, with an area under the curve of 0.72 (95 percent confidence interval 0.60 to 0.82), followed by P-wave duration at 0.67 (0.55 to 0.78). Combining both indices produced a slightly higher apparent AUC of 0.76 (0.65 to 0.86), but the difference from voltage alone was not statistically significant (p = 0.15). A P-wave voltage of 0.08 millivolts or less identified myxoma patients with a sensitivity of 89.7 percent and a specificity of 50.0 percent, while a threshold of 0.06 millivolts yielded a sensitivity of 48.7 percent and a specificity of 72.7 percent. A P-wave duration of 80 milliseconds or less was highly specific (95.5 percent) but insensitive (38.5 percent). Adding clinical parameters to the two ECG variables did not meaningfully improve discrimination (apparent AUC 0.77; p = 0.08 versus voltage alone), and a simple two-item score combining duration of 100 milliseconds or less and voltage of 0.07 millivolts or less reached an AUC of only 0.66 (0.54 to 0.76).

The clinical logic is straightforward. Because the ECG is already a first-line investigation in patients presenting with stroke, dyspnea, or unexplained pulmonary embolism, a systematic look at P-wave voltage in lead I, and to a lesser extent P-wave duration, could help identify the small but clinically important subset of patients in whom a myxoma underlies these presentations. Both measurements are obtainable from any standard digital ECG without additional software or equipment, and could be integrated into routine interpretation at minimal cost. The authors stress that their findings do not challenge the diagnostic hierarchy in which echocardiography remains the mainstay of atrial tumor detection, with cardiac magnetic resonance complementing tissue characterization and surgical planning. Rather, they suggest that careful ECG analysis may surface underutilized diagnostic information that could help shorten the interval from presentation to definitive imaging.

The limitations are substantial and the authors enumerate them candidly. The retrospective single-center design limits generalizability and precludes causal inference. The sample of 83 patients is modest, and the study may have been underpowered to detect smaller effects among the remaining P-wave indices. The cohort consisted exclusively of atrial myxomas, so the findings cannot be extrapolated to papillary fibroelastomas, sarcomas, or cardiac metastases. The higher prevalence of coronary artery disease, dyslipidemia, and smoking in the tumor group introduces residual confounding that a retrospective design cannot fully resolve, although the sensitivity analysis confirmed that voltage remained an independent predictor. Postoperative ECGs were not systematically available, left- and right-atrial dimensions were not recorded, and the authors deliberately declined to propose a diagnostic risk score because sex was fixed by matching and prior stroke was not more frequent in the myxoma group. The authors themselves note that the discrimination of P-wave voltage in lead I was only modest and the absolute differences between groups small, meaning a reduced voltage cannot serve as a stand-alone test and may at best act as an additional argument for echocardiography. Within those bounds, the study adds a strikingly simple observation to cardiology’s diagnostic toolkit: the heart’s electrical signature may whisper what the tumor has been hiding.

Subject of Research: Electrocardiographic P-wave indices as potential diagnostic markers of atrial myxoma

Article Title: P-Wave indices in the presence of atrial myxomas: a retrospective cohort analysis

Article References: Doldi, F., Frommeyer, G., Martens, S., Hoffmeier, A., & Eckardt, L. (2026). P-Wave indices in the presence of atrial myxomas: a retrospective cohort analysis. Clinical Research in Cardiology. https://doi.org/10.1007/s00392-026-03049-4

Image Credits: AI Generated

DOI: 10.1007/s00392-026-03049-4

Keywords: atrial myxoma, ECG, P-wave voltage, P-wave duration, cardiac tumor, logistic regression, echocardiography, retrospective cohort, lead I, diagnostic adjunct, atrial cardiomyopathy, Clinical Research in Cardiology