Mean follow-up was 1.0 ± 0.1 years, and 35.4% of the cohort experienced recurrent atrial fibrillation or atrial tachycardia. Recurrence occurred in 46.2% of the 13 patients with one topological obstacle and 44.4% of the 27 with four or more, compared with 23.5% of 17 patients with two and 34.4% of 180 with three. The comparison between any recurrence and no recurrence across the four groups was not statistically significant (p = 0.433), while the overall distribution of no recurrence, atrial fibrillation and atrial tachycardia also did not differ significantly (p = 0.584).

Multivariable logistic regression adjusted for age, clinical type of atrial fibrillation, left atrial diameter, left ventricular ejection fraction, CHA2DS2-VASc score and low-voltage area burden; no obstacle-count group was significantly associated with recurrence relative to the three-obstacle group, and the overall likelihood-ratio p value was 0.194. The research was conducted at Seoul National University Hospital among patients who underwent radiofrequency catheter ablation between January 2023 and January 2024. Of 336 people initially considered, the investigators excluded 78 with a history of prior ablation, 13 without a complete left atrial voltage map, four without post-ablation follow-up and four whose recurrence was typical atrial flutter.

The final sample included 42.6% with paroxysmal atrial fibrillation, 43.5% with persistent disease and 13.9% with long-standing persistent disease. High-density electroanatomic maps acquired before ablation were used to identify low-voltage myocardium, with bipolar thresholds below 0.5 mV in sinus rhythm or pacing and below 0.2 mV during atrial fibrillation. Maps were acquired during sinus rhythm or pacing in 166 patients, or 70.0%, and during atrial fibrillation in 71, or 30.0%.

Low-voltage area burden was calculated as the percentage of the total mapped left atrial surface meeting the applicable voltage threshold. The investigators modeled the atrium as a closed two-dimensional surface and defined topological obstacles as non-conductive regions that interrupt electrical propagation. Each map initially contained three fundamental obstacles: the left pulmonary vein antrum, the right pulmonary vein antrum and the mitral valve.

Discrete low-voltage regions were then added when they met a size criterion, while extensive scar connecting anatomical structures could merge separate barriers into a single continuous obstacle. Incorporating the actual ablation lines completed the topological representation. To avoid counting negligible low-voltage spots as functional barriers, the researchers derived a minimum area from the wavelength required to support a macroreentrant circuit.

They used an estimated minimum atrial action potential duration of 150 milliseconds and a conduction velocity of 0.5 meters per second, producing a wavelength of 75 millimeters. Under an assumed circular geometry, that circumference corresponded to a radius of approximately 11.9 millimeters and an area of 4.45 square centimeters, rounded to 4.5 square centimeters for classification. Low-voltage regions smaller than 4.5 square centimeters were therefore not treated as independent obstacles.

The threshold was based on literature-derived rather than patient-specific physiological estimates, and the circular assumption could underestimate the boundary length of small, irregularly shaped regions. Obstacle count was related to low-voltage burden, but the relationship was not linear. Patients with one obstacle had the highest mean burden at 36.6% ± 24.8%, and all 13 had a burden of at least 5%, compared with 24.9% of the overall cohort.

The largest group, containing 180 patients with three obstacles, had a mean burden of only 2.1% ± 3.5%. Counts differed significantly across burden quartiles (p < 0.001). The highest-burden quartile, with a mean burden of 18.9% ± 17.6%, was the only quartile containing patients from all four obstacle groups, demonstrating that extensive low-voltage tissue can form either one large confluent barrier or several discrete ones.

The lowest quartile contained only patients with three obstacles. Patients with a single obstacle also differed clinically from the rest of the cohort. They were older, with a mean age of 67.8 years, and had a larger left atrium, lower left ventricular ejection fraction and higher CHA2DS2-VASc score.

Their mean left atrial diameter was 50.1 ± 5.2 mm, compared with 44.6 ± 7.0 mm in the three-obstacle group, while mean ejection fraction was 51.7% ± 12.6% and 57.6% ± 6.5%, respectively. Valvular heart disease was the only comorbidity that differed significantly across obstacle groups (p = 0.021). These imbalances limit attempts to separate the influence of topology from that of underlying remodeling: a lone obstacle was simultaneously associated with a greater burden of low-voltage tissue and less favorable cardiac measurements.

Most patients received pulmonary vein antrum isolation alone as the standard strategy, although 81, or 34.1%, also received additional lesions at the operator’s discretion; cavotricuspid isthmus ablation was the most common. The primary outcome was atrial fibrillation or atrial tachycardia lasting at least 30 seconds, documented on a 12-lead electrocardiogram or Holter monitor, after a three-month blanking period. Atrial tachycardia was less common than atrial fibrillation among recurrent cases, and the mean cycle length in patients with atrial tachycardia was 235 milliseconds.

Across low-voltage-area quartiles, recurrence was 20.0% in the lowest quartile, 41.7% in the second, 34.5% in the third and 45.8% in the highest; the overall quartile-level comparison was statistically significant (p = 0.018). Comparisons among obstacle groups within individual quartiles were limited by small subgroups and did not show a consistent pattern of increasing recurrence with obstacle count. The investigators also examined whether the structural differences were reflected in electrocardiographic measures of fibrillatory organization.

They analyzed lead V1 from 12-lead recordings obtained at recurrence, separated QRS and T waves with NeuroKit2, extracted and preprocessed fibrillatory-wave segments, and used fast Fourier analysis to calculate dominant frequency and organization index. Dominant frequency ranged from 5.04 ± 0.87 Hz in the single-obstacle group to 5.66 ± 1.51 Hz among patients with four or more obstacles. Organization index was highest in the single-obstacle group at 0.49 ± 0.09 and lowest in the three-obstacle group at 0.45 ± 0.07.

Neither measure differed significantly by obstacle count or low-voltage-area quartile, so the global signal-derived measures did not reproduce the distinctions identified by voltage mapping. The study’s retrospective, single-center design and small extreme-count groups limit generalizability. Operator-dependent additional ablation and incompletely documented comorbidities may also have influenced outcomes.

Voltage measurements can vary with mapping rhythm, wavefront direction and electrode configuration, while low-voltage area remains an imperfect surrogate for fibrosis. The fixed action-potential-duration and conduction-velocity assumptions and circular geometry make the 4.5-square-centimeter classification method-dependent. Threshold sensitivity against alternative cutoff values was not formally tested, although a supplementary shape-based analysis suggested that the circular assumption may have a modest effect for mildly elongated regions and a greater effect for markedly elongated ones.

Taken together, the results support further investigation of atrial substrate geometry, but they do not establish obstacle count as a recurrence predictor or clinical ablation target.

Subject of Research: Association of left atrial topological obstacles and low-voltage burden with atrial fibrillation recurrence after pulmonary vein antrum isolation

Article Title: Impact of low-voltage area burden and topological obstacles on atrial fibrillation recurrence after pulmonary vein antrum isolation

Article References: Choi, J., Go, Y.-H., Kwon, S., Lee, K.-Y., Ahn, H.-J., Lee, S.-R., Kim, T., Lee, J. C., Choi, E.-K., & Oh, S. (2026). Impact of low-voltage area burden and topological obstacles on atrial fibrillation recurrence after pulmonary vein antrum isolation. Frontiers in Network Physiology, 6, 1945351. https://doi.org/10.3389/fnetp.2026.1945351

Image Credits: AI Generated

DOI: 10.3389/fnetp.2026.1945351

Keywords: atrial fibrillation, pulmonary vein antrum isolation, topological obstacles, low-voltage area, cardiac ablation, arrhythmia recurrence, atrial substrate, left atrium, electrical conduction, network physiology, Impact, low-voltage

Researchers at a single center analyzed pre-ablation contrast-enhanced cardiac CT angiography images using a MedNeXt-based workflow that generated standardized meshes of the left atrial body. Six global descriptors were then calculated: Wadell sphericity, two coordinate-covariance eigenvalue-ratio indices known as E and F, convex hull ratio, and two diameter-normalized volume sphericity indices based on either the maximum diameter or the mean of three orthogonal diameters. The study was designed to test measurement methods and sensitivity, not to determine whether any descriptor predicts clinical outcomes.

Agreement with expert-guided correction was evaluated in a stratified subset of 40 cases. Relative to the corrected outputs, automated segmentations achieved a Dice coefficient of 0.956 ± 0.015 and a 95th-percentile Hausdorff distance of 1.62 ± 0.55 mm. Intraclass correlation coefficients for the six descriptors ranged from 0.871 to 0.947, indicating agreement between automated and expert-guided results, although the researchers stressed that this was not independent validation against fully manual segmentation. Running the workflow twice produced intraclass correlations of at least 0.996, demonstrating very high consistency in the implementation.

The comparisons also revealed substantial differences among metrics that may sound similar. Wadell sphericity was essentially unrelated to the two diameter-normalized volume sphericity indices, while those two variants were strongly correlated with one another. E, F, convex hull ratio, and both diameter-normalized indices were associated with left atrial volume. A component-based allometric analysis indicated that shape components do not necessarily scale together as overall size changes, so adjusting a measure for volume may not remove all geometric dependence. The AF-subtype-by-volume interaction was not evaluated, limiting interpretation of the exploratory subtype models.

Boundary perturbations produced further warnings about precision. Predefined changes to voxel resolution, mesh construction, and truncation planes by one voxel in either direction were used to assess sensitivity. Convex hull ratio showed the greatest tested boundary sensitivity, with a 95th-percentile change of 4.8%. Its association with AF subtype in an exploratory volume-adjusted model was modest and only nominally significant, with a standardized beta of −0.23 and a 95% confidence interval of −0.42 to −0.04; the authors characterized that result as hypothesis-generating rather than evidence of an independent clinical effect.

Overall, the study indicates that automated CT analysis can generate reproducible left atrial shape descriptors, but the six measures quantify different properties and carry different measurement uncertainties. Clinical use would require transparent formulas, standardized anatomical boundaries, appropriate modeling of size, independent manual validation, assessment of scan-to-scan variability, and validation in external populations. Because the work was retrospective, single-center, and methodological, it did not establish that any shape descriptor improves diagnosis, treatment selection, prediction of recurrence, or patient outcomes in atrial fibrillation.

Subject of Research: Automated three-dimensional left atrial shape analysis in cardiac CT for patients with atrial fibrillation

Article Title: Automated computation and comparison of three-dimensional left atrial shape descriptors from deep learning-assisted CT segmentation in atrial fibrillation: a retrospective methodological imaging study

Article References: Lu, L., & Shi, Z. (2026). Automated computation and comparison of three-dimensional left atrial shape descriptors from deep learning-assisted CT segmentation in atrial fibrillation: a retrospective methodological imaging study. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02945-5

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02945-5

Keywords: atrial fibrillation, cardiac CT, left atrium, deep learning, image segmentation, shape descriptors, sphericity, medical imaging, methodological study, Automated, computation, comparison

The research, led by Chiara Giordano and Giacomo Gamberini of the University of Milan together with senior author Stefano Carugo and a team spanning the Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, focused on a measurement called the Left Atrial Volumetric–Mechanical Coupling Index, abbreviated LACI-VM. The index is elegantly simple: it is calculated by dividing the left atrial volume index—the size of the heart’s upper left chamber adjusted for body surface area—by a′, the late diastolic mitral annular velocity measured with Doppler tissue imaging. That denominator reflects how well the atrium contracts during its mechanical squeeze. A large atrium that still contracts vigorously produces a modest ratio; a large, sluggish atrium produces a high one. In other words, LACI-VM fuses two complementary dimensions of atrial disease into a single number obtainable from any standard transthoracic echo.

The rationale is rooted in the biology of atrial cardiomyopathy. When the left atrium is chronically exposed to high filling pressures, hypertension, or aging-related fibrosis, it enlarges and its muscle fibers are progressively replaced by scar tissue. Electrical conduction through this remodeled chamber becomes slow and heterogeneous, creating the substrate for the reentrant circuits that drive atrial fibrillation. Structural dilation alone, captured by the left atrial volume index, has long been known to predict AF detection after cryptogenic stroke. Mechanical dysfunction, assessable through atrial strain or annular velocities, adds independent information. By combining both, LACI-VM aims to capture the full anatomical and functional footprint of an atrium that is primed to fibrillate.

The Milanese team applied this concept to 285 consecutive patients admitted with an embolic stroke or transient ischemic attack of undetermined source. All underwent transthoracic echocardiography during the index hospitalization, including measurement of LACI-VM and of the related Left Atrioventricular Coupling Index, or LACI, which is defined as the ratio of left atrial end-diastolic volume to left ventricular end-diastolic volume and reflects atrioventricular uncoupling. The patients, whose median age was 71 years and of whom 43 percent were women, were then followed for new-onset atrial fibrillation. Over a median follow-up of 10.9 months, AF appeared in 24 patients, or 8.4 percent of the cohort—and, critically, 62 percent of those events occurred within the first year, underscoring how early the arrhythmia tends to declare itself after a cerebral embolic event.

The statistical results were striking. LACI-VM was significantly higher in patients who went on to develop atrial fibrillation, and after multivariable adjustment for confounders it remained independently associated with incident AF, with a hazard ratio of 6.64 and a confidence interval of 1.34 to 59.88, reaching a p-value of 0.041. A hazard ratio approaching seven means that patients with an impaired coupling index faced roughly a six- to seven-fold higher rate of AF detection than their counterparts with preserved values, a magnitude of effect rarely seen for a single echocardiographic parameter. By contrast, the purely volumetric LACI showed only a non-significant overall association, with a hazard ratio of 2.28 and a p-value of 0.09, and its predictive power appeared confined to events occurring within the first twelve months. The comparison suggests that atrial mechanical performance, not merely chamber size, carries the prognostic signal that matters for unmasking occult AF.

Perhaps the most clinically actionable finding emerged when the researchers combined LACI-VM with an assessment of left ventricular filling pressures, expressed as the E/e′ ratio—the ratio of early transmitral flow velocity to early diastolic mitral annular velocity. Patients whose LACI-VM remained below 2.5 and whose E/e′ was 10 or lower experienced an AF incidence of only about 2 percent, identifying a genuinely low-risk subgroup in whom intensive, invasive rhythm monitoring may be safely deferred. At the opposite extreme, when both the coupling index was elevated at 2.5 or above and filling pressures were elevated above 10, the AF risk climbed to roughly 20 percent—a tenfold difference between the two phenotypes. This layered approach transforms a single ultrasound number into a practical decision tool: one echo study, performed during the index admission, can separate stroke survivors who need implantable loop recorders from those who can be managed with conventional, less burdensome surveillance.

The implications reach into one of the most contested corners of stroke medicine. Randomized trials such as CRYSTAL-AF and FIND-AF, which used implantable or prolonged monitoring, established that occult atrial fibrillation is common after cryptogenic stroke, detecting it in a meaningful fraction of patients over months to years. Yet prolonged monitoring is costly, and debates persist about how intensively every ESUS patient should be screened, whether oral anticoagulation benefits all detected subgroups, and how to weigh the burden of devices against the risk of recurrent embolism. European Society of Cardiology consensus statements have repeatedly called for better tools to stratify patients. An index that costs nothing beyond an echocardiogram already being performed fits that need precisely, potentially concentrating expensive monitoring resources on the patients most likely to benefit.

The study also builds on a growing literature. LACI-VM was first proposed as a predictor of outcomes in heart failure with reduced ejection fraction, and subsequent work has linked atrioventricular coupling indices to cardiovascular events in the Multi-Ethnic Study of Atherosclerosis, to new-onset AF in hypertrophic cardiomyopathy, and to heart failure and AF in the general population. Earlier investigations specifically in ESUS populations reported that the volumetric-mechanical coupling index associated with AF detection, and the Milan group itself had previously shown that peak atrial longitudinal strain can catch atrial fibrillation in sinus-rhythm patients admitted for cerebral ischemic events. The present analysis extends this line by demonstrating independent prediction in a dedicated ESUS and TIA cohort and by pairing the index with filling pressures to define a truly low-risk stratum.

Certain caveats temper the enthusiasm. The investigation was retrospective and conducted at a single center, and the confidence interval around the hazard ratio is wide, reflecting the modest number of endpoint events; larger, prospective, multicenter validation will be needed before the index enters guidelines. The follow-up duration of roughly eleven months captures the early window in which most events occurred but may miss later arrhythmia emergence. Still, the direction of the evidence is coherent with atrial physiology, the measurement is reproducible with standard techniques endorsed by the American Society of Echocardiography and the European Association of Cardiovascular Imaging, and the effect sizes are large enough to be clinically meaningful even in a small cohort.

For patients and clinicians alike, the message is compelling. A stroke whose cause remains invisible may not be invisible to the heart itself: the left atrium often carries the fingerprints of the arrhythmia that launched the clot. By reading those fingerprints through a simple ratio of atrial size to atrial squeeze, obtained from an ultrasound exam already part of routine care, physicians may soon be able to tailor rhythm monitoring with far greater precision—intensifying surveillance for the one in five patients at highest risk while sparing the many at genuinely low risk from unnecessary devices and prolonged uncertainty. In a field where every undetected episode of atrial fibrillation represents a potential second, potentially devastating stroke, a cheap, fast, and widely available predictor is exactly the kind of advance that could change practice.

Subject of Research: Prediction of early new-onset atrial fibrillation after embolic stroke or TIA of undetermined source using echocardiographic left atrial volumetric–mechanical coupling indices

Article Title: Echocardiography-derived Left Atrial Volumetric–Mechanical Coupling Index predicts early new-onset atrial fibrillation after embolic stroke/transient ischemic attack of undetermined source

Article References: Giordano, C., Gamberini, G., Faggiano, A., Garau, L., Conti, G., D’Andria, P. F., Moda, F., Mallardi, G., Saviano, M., Bonato, S., Ghione, I., Lanfranconi, S., Gherbesi, E., Santangelo, G., Ruscica, M., Bursi, F., & Carugo, S. (2026). Echocardiography-derived Left Atrial Volumetric–Mechanical Coupling Index predicts early new-onset atrial fibrillation after embolic stroke/transient ischemic attack of undetermined source. Clinical Research in Cardiology. https://doi.org/10.1007/s00392-026-03041-y

Image Credits: AI Generated

DOI: 10.1007/s00392-026-03041-y

Keywords: atrial fibrillation, embolic stroke of undetermined source, echocardiography, left atrium, LACI-VM, left atrioventricular coupling index, transient ischemic attack, rhythm monitoring, implantable loop recorder, atrial cardiomyopathy, E/e′ ratio, stroke prevention

The coronary angiographic findings deserve particular emphasis. Coronary air embolism is a recognized but uncommon event, usually described in the context of cardiac surgery, percutaneous procedures, or right-to-left shunts, and it is exceptionally rare as the presenting manifestation of an atrioesophageal fistula. In this patient, the operator could see gas as a mobile, lucid filling defect in the mid right coronary artery on the very first contrast injection, with TIMI grade 0 flow distally. Aspiration with the catheter removed a large volume of gas without any thrombus, and antegrade flow immediately returned to TIMI grade 3. That sequence is diagnostically powerful: the absence of thrombus or atherosclerotic occlusion, the instantaneous restoration of flow after aspiration, and the simultaneous presence of intravascular gas elsewhere on imaging together rule in gas embolism as the mechanism of the inferior ST elevation rather than plaque rupture or coronary spasm. It also demonstrates a practical point for interventional cardiologists, namely that aspiration can be both diagnostic and therapeutic in suspected coronary air embolism, restoring perfusion even when the underlying source of the gas cannot be corrected at the same sitting.

The electrocardiographic distribution of injury adds another layer of plausibility. The right coronary artery supplies the inferior wall in most individuals, and the ST-segment elevation in leads II, III, and aVF matched the angiographic location of the gas. Transient ST elevation from air embolism has been reported before, but it is easily mistaken for acute plaque rupture, which can lead to inappropriate delays while anticoagulation, loading with antiplatelet agents, or plans for percutaneous intervention are considered. In a patient weeks after left atrial ablation, the combination of abrupt neurologic depression, fever, and ST elevation should immediately raise suspicion for systemic air embolism from a fistula, and emergency angiography in this setting can be justified both to treat the ischemic event and to document its mechanism.

The microbiology in this case reinforces the pathophysiologic model. Blood cultures grew Streptococcus constellatus together with small quantities of other streptococci. S. constellatus belongs to the anginosus group, organisms that colonize the oropharynx, gastrointestinal tract, and genitourinary tract and are well known for abscess formation and bacteremia arising from mucosal disruptions. Isolation of such organisms from the bloodstream shortly after the abrupt clinical deterioration is difficult to explain by any mechanism other than translocation of esophageal or oral contents into the left atrium. It also helps distinguish fistula-related sepsis from the low-grade fevers that can accompany uncomplicated pericarditis or esophageal ulceration after ablation. In practice, when blood cultures from a post-ablation patient with neurologic findings grow oral or enteric streptococci, the pretest probability of an atrioesophageal fistula rises substantially, and empiric broad-spectrum antibiotic coverage covering streptococci and anaerobes should be started without waiting for definitive imaging.

The imaging strategy in this case reflects the delicate balance between diagnostic certainty and patient safety. Non-contrast computed tomography of the head and chest was performed rapidly and showed gas in the cortical sulci, along the superior sagittal sinus, in all cardiac chambers, and in the aorta, coronary arteries, pulmonary arteries, and cervical vessels. Retrospective review identified a suspected tract between the posterior left atrial wall and the adjacent esophagus. Contrast-enhanced computed tomography, which would have provided the strongest direct evidence of a fistulous connection, was not performed because the patient was hemodynamically unstable. Upper gastrointestinal endoscopy and transesophageal echocardiography were deliberately avoided, consistent with the widely held view that endoscopic insufflation can propel additional gas and esophageal contents through the fistula into the arterial circulation and worsen neurologic injury. This is a recurring dilemma in the literature: the tests that would confirm the diagnosis are the same tests that may accelerate the disease. Most experts therefore recommend that when the clinical syndrome and non-invasive imaging are sufficiently suggestive, treatment should proceed on the presumptive diagnosis, with surgical consultation obtained immediately, rather than waiting for endoscopic or contrast confirmation that may never be safely obtainable.

Timing is another consistent theme. The fistula in this patient declared itself approximately one month after ablation, which falls squarely within the typical window. Published series describe symptom onset most often in the second to fourth week after the procedure, with a mean interval of roughly two to three weeks, and the esophageal ulceration that precedes fistulization has been documented endoscopically within days of radiofrequency delivery. The delay occurs because the thermal injury must deepen through the esophageal mucosa and muscularis before the wall is breached. This latent period has two practical consequences. First, patients and clinicians should not consider the danger period over at hospital discharge; symptom education must extend for at least a month, and any new fever, chest pain, dysphagia, odynophagia, hematemesis, seizure, stroke, or altered mental status in that window warrants urgent evaluation. Second, the delayed onset means the causal connection to the ablation is easy to miss, particularly when the patient presents to a facility that did not perform the procedure, as happened here. Procedural details were unavailable to the treating team, which likely contributed to diagnostic uncertainty during the first hours.

Prevention remains the most effective intervention, because survival once the fistula is established is poor even with prompt surgery. Radiofrequency energy is not tissue selective, and the esophagus lies immediately posterior to the left atrium in many patients, sometimes separated by only a few millimeters of tissue. The distance between the esophagus and the posterior left atrial wall has been identified as an important risk factor, and esophageal ulcers have been found in a small but meaningful fraction of patients screened endoscopically after ablation. Strategies to reduce risk include limiting posterior wall radiofrequency power and duration, monitoring esophageal temperature during ablation, and using active esophageal cooling devices, which have been proposed to lower the risk of fistula formation. Mechanical esophageal displacement, luminal esophageal temperature probes, and the choice of energy source, including cryoablation or pulsed field ablation with its theoretical tissue selectivity, are areas of ongoing study. Post-procedural acid suppression and treatment of gastroesophageal reflux may also matter, because reflux is thought to promote the progression from superficial ulcer to deep fistula.

The hemodynamic and laboratory evolution in this patient tracks the expanding injury. Early point-of-care testing showed only modest myoglobin elevation with near-normal troponin, but by ten hours the high-sensitivity troponin had risen to more than 6,000 ng/L and ultimately exceeded 25,000 ng/L, reflecting extensive myocardial necrosis from coronary embolism and prolonged shock. Interleukin-6 above 5,000 pg/mL signaled a profound systemic inflammatory response driven by bacteremia and tissue injury. Recurrent bradycardia and hypotension preceded cardiac arrest, and death occurred approximately twenty-seven hours after admission despite antibiotics, anticoagulation, antiplatelet therapy, statin, and vasopressor support. Surgical repair was considered but deferred because refractory shock made perioperative survival unlikely. This outcome mirrors the published experience, in which mortality for atrioesophageal fistula remains high even in surgically treated patients, and delays of more than twenty-four hours between symptom onset and repair are associated with markedly worse survival. The lesson is that the window for salvage is measured in hours, and the decision to operate must often be made on clinical suspicion alone.

Finally, the authors’ caution about diagnostic certainty is worth underscoring. The case demonstrates direct angiographic visualization of intracoronary gas and successful catheter aspiration, and it shows multisite intravascular gas on non-contrast imaging together with bacteremia from organisms characteristic of the oropharyngeal and gastrointestinal flora. These findings together make the diagnosis of an atrioesophageal fistula complicated by systemic air embolism highly probable, but the fistulous tract itself was only a retrospective suspicion on non-contrast imaging, never confirmed by contrast-enhanced cross-sectional imaging, endoscopy, or surgery. Reporting the case with that limitation intact is scientifically honest and clinically useful, because it models the epistemic situation clinicians actually face: the evidence is often strong but incomplete, the confirmatory tests are hazardous, and the cost of waiting is death. For clinicians, the practical synthesis is straightforward. In any patient within several weeks of atrial fibrillation ablation who presents with fever, new neurologic deficits, altered consciousness, seizures, or acute ST-segment elevation, atrioesophageal fistula with systemic air embolism belongs at the top of the differential diagnosis. Immediate computed tomography of the head and chest, blood cultures, empiric antibiotics, avoidance of endoscopy and esophageal instrumentation, urgent cardiothoracic surgical consultation, and consideration of hyperbaric oxygen therapy where available constitute the appropriate initial response, and coronary angiography with aspiration can play a defined role when myocardial ischemia is the dominant manifestation.

Subject of Research: Cardiovascular Medicine

Article Title: Suspected atrioesophageal fistula complicated by coronary air embolism after catheter ablation for atrial fibrillation: direct angiographic visualization and catheter aspiration in a case report

Article References: Yang, C., Zeng, X., Wang, S., Wang, C., Lan, J., Wen, Y., & Zheng, Z. (2026). Suspected atrioesophageal fistula complicated by coronary air embolism after catheter ablation for atrial fibrillation: direct angiographic visualization and catheter aspiration in a case report. Frontiers in Cardiovascular Medicine, 13, Article 1985470. https://doi.org/10.3389/fcvm.2026.1985470

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1985470

Keywords: atrial fibrillation, catheter ablation, atrioesophageal fistula, coronary air embolism, case report, radiofrequency ablation, Streptococcus constellatus, coronary angiography, left atrium, esophageal injury, cardiac arrest, Frontiers in Cardiovascular Medicine

The research, led by Mengyuan Jing, Haoxiang Lu, and Qing Liu with colleagues at Lanzhou University Second Hospital, the Guangdong Cardiovascular Institute, and partner institutions, set out to build an explainable machine learning model that could predict functional improvement after ablation in patients with atrial fibrillation combined with heart failure. Rather than relying on a single measurement, the team fused three complementary streams of information: the geometry of the left atrium and pulmonary veins as seen on cardiac CT angiography, quantitative radiomics features extracted from the left atrial wall, and a handful of routine clinical variables. The result was a combined model, dubbed COMB, that achieved an area under the receiver operating characteristic curve of 0.866 in the training set, 0.803 in the validation set, and 0.845 in an independent testing set drawn from other hospitals.

Those numbers deserve unpacking, because the area under the curve, or AUC, is the workhorse metric for judging how well a model separates patients who will improve from those who will not. A value of 0.5 would mean the model performs no better than a coin flip, while 1.0 would indicate perfect discrimination. Scores consistently above 0.80 across three separate patient cohorts, including an external test set the model had never encountered during development, represent solid, clinically meaningful performance. Equally important is how the team arrived at that performance. Patients from the primary institution were randomly divided into training and validation sets in a seven-to-three ratio, while two other organizations contributed entirely separate testing cohorts, a design that guards against the model simply memorizing the quirks of one hospital’s scanner or patient population.

The technical heart of the study lies in what the researchers chose to measure. Cardiac CT angiography, already a routine part of pre-ablation planning because it maps the pulmonary veins before catheter insertion, provides exquisitely detailed three-dimensional images of the left atrium, the chamber where fibrillation typically originates. From these images the team extracted morphological features describing the shape of the left atrium and the pulmonary veins, capturing subtle geometric signatures such as chamber distortion and remodeling that a human reader might overlook. In parallel, they applied radiomics, a computational approach that converts medical images into hundreds of quantitative descriptors of texture, intensity, and spatial pattern, to the left atrial wall itself, the thin muscular sleeve where ablation lesions are created and where fibrotic change often determines whether the procedure succeeds.

From this high-dimensional feature space, the researchers distilled two focused models. Two shape features and three left atrial wall radiomics features survived rigorous screening and were used to construct what they called the Shape model and the Wall model, respectively. Each model produced a continuous score, the Shape_score and the Wall_score, which quantified how strongly a given patient’s cardiac anatomy resembled that of patients who went on to improve. These scores were then integrated with the clinical variables that proved most informative, namely gender, hyperlipidemia, blood urea level, and the type of atrial fibrillation, to yield the final combined model. The parsimony is striking: out of the vast number of features that could have been included, the final predictor rests on just a handful of imaging and clinical inputs, all obtainable from examinations and blood tests that are already standard practice.

What elevates the study beyond a typical prediction exercise is its commitment to explainability. Black-box algorithms have long been a stumbling block for clinical adoption, because physicians are understandably reluctant to act on a probability they cannot interrogate. The researchers addressed this by applying SHAP, or Shapley additive explanations, a technique borrowed from cooperative game theory that assigns each input feature a precise contribution to every individual prediction. In effect, SHAP reveals which factors pushed a particular patient’s predicted probability up or down, and by aggregating these contributions across the cohort, the team identified the features with the greatest impact on outcomes in both the Shape and Wall models. This transparency allows a cardiologist to see, for example, whether an unfavorable prediction stems from pronounced atrial remodeling, abnormal wall texture suggesting fibrosis, or a clinical factor such as persistent rather than paroxysmal fibrillation.

The clinical stakes of this kind of tool are considerable. Patients with atrial fibrillation and concomitant heart failure represent a particularly vulnerable group, and ablation in these individuals carries procedural risk, substantial cost, and a demanding recovery. If a model can reliably flag patients unlikely to experience functional improvement, clinicians could counsel them more honestly, weigh alternative management strategies such as optimized medical therapy, or reserve ablation for those with the greatest expected benefit. Conversely, identifying patients with a high predicted probability of improvement could support earlier referral and shared decision-making grounded in quantitative evidence rather than clinical intuition alone. Because the model’s inputs come from CT angiography and routine laboratory work, it could in principle be deployed without any additional testing beyond what pre-ablation workups already include.

The multicenter architecture of the study strengthens its claims in another important way. Models trained and tested within a single institution often flatter themselves, absorbing center-specific artifacts in scanner protocol, image reconstruction, and patient mix. By including 240 patients in the training set, 101 in the validation set, and 75 in external testing sets from two other organizations, the team demonstrated that the model’s discrimination held up across different hospitals and imaging environments. The ethics committees of all three participating institutions approved the retrospective study, and the requirement for individual informed consent was waived because of its retrospective design, with the work conducted in accordance with the Declaration of Helsinki.

The authors themselves are careful about the limits of what they have shown. The study was retrospective, meaning it looked backward at patients who had already undergone ablation rather than prospectively assigning the model to guide care in real time. The team explicitly states that further prospective evaluation is required before the model can be implemented clinically, a caveat that applies to nearly every prediction algorithm now emerging in cardiovascular medicine. Prospective validation would involve using the COMB model to generate predictions before ablation and then tracking whether those predictions match observed outcomes, ideally across diverse populations and healthcare systems. Questions about calibration, the agreement between predicted probabilities and actual event rates, and about how the model behaves in subgroups underrepresented in the data, would also need answers.

Even with those caveats, the study offers a compelling glimpse of where cardiac imaging and artificial intelligence are converging. The same CT scan obtained to map a patient’s pulmonary veins becomes, through shape analysis and radiomics, a quantitative portrait of atrial health, and machine learning converts that portrait, together with a few clinical facts, into an individualized forecast of recovery. The SHAP framework keeps the forecast auditable, showing clinicians exactly which anatomical and biological signals drove the conclusion. For the millions of patients whose fibrillation and heart failure travel together, and for the physicians deciding whether to offer them ablation, a tool that turns pre-procedural images into evidence-based expectations could reshape the conversation, replacing uncertainty with a number that both doctor and patient can understand, question, and trust.

Subject of Research: Explainable machine learning on cardiac CT angiography to predict functional improvement after atrial fibrillation ablation in patients with heart failure

Article Title: Explainable machine learning model based on cardiac CT angiography for predicting functional improvement after atrial fibrillation ablation: a multicenter study

Article References: Jing, M., Lu, H., Liu, Q., Jing, Y., Lei, F., Yang, X., Chen, G., Xi, H., Xin, W., Zhu, H., Sun, Q., Zhang, Y., Ren, J., Ren, W., Liu, Z., Wang, G., & Zhou, J. (2026). Explainable machine learning model based on cardiac CT angiography for predicting functional improvement after atrial fibrillation ablation: a multicenter study. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02822-1

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02822-1

Keywords: atrial fibrillation, heart failure, catheter ablation, cardiac CT angiography, machine learning, radiomics, left atrium, SHAP explainability, predictive modeling, multicenter study, BMC Medical Imaging, cardiac imaging