MRI pattern characterization of cerebral cardioembolic lesions following atrial fibrillation ablation

Background Recognizing etiology is essential for treatment and secondary prevention of cerebral ischemic events. A magnetic resonance imaging (MRI) pattern suggestive of an embolic etiology has been described but, to date, there are no uniformly accepted criteria. Aim The purpose of the study is to describe MRI features of ischemic cerebral lesions occurring after transcatheter ablation of atrial fibrillation (AF). Methods A systematic review and meta-analysis of studies performing brain imaging investigations before and after AF transcatheter ablation was performed. The incidence of cerebral ischemic lesions after AF transcatheter ablation was the primary endpoint. The co-primary endpoints were the prevalence of the different neuroimaging features regarding the embolic cerebral ischemic lesions. Results A total of 25 studies, encompassing 3,304 patients, were included in the final analysis. The incidence of ischemic cerebral lesions following AF transcatheter ablation was 17.2% [95% confidence interval (CI) 12.2%–23.8%], of which a minimal fraction was symptomatic [0.60% (95% CI 0.09%–3.9%)]. Only 1.6% of the lesions (95% CI 0.9%–3.0%) had a diameter >10 mm, and in 20.5% of the cases the lesions were multiple (95% CI 17.1%–24.4%). Brain lesions were equally distributed across the two hemispheres and the different lobes; cortical location was more frequent [64.0% (95% CI 42.9%–80.8%)] while the middle cerebral artery territory was the most involved 37.0% (95% CI 27.3–48.0). Conclusions The prevailing MRI pattern comprises a predominance of small (<10 mm) cortical lesions, more prevalent in the territory of the middle cerebral artery.


Introduction
Stroke is a leading cause of mortality and long-term disability (1); recognizing the underlying cause of stroke is relevant for treatment, prognosis, and secondary prevention (2).However, about 25% of strokes are defined "cryptogenic" because the etiology remains unknown in spite of exhaustive investigations (3).Hart et al. recently proposed to call these types of lesions "embolic strokes of undetermined source" (ESUS) (4), considering that the vast majority recognizes an embolic etiology.Detecting a potential mechanism would, indeed, be relevant, when therapeutic options aiming at preventing recurrences are available, such as the initiation of oral anticoagulation in case of atrial fibrillation (AF) or percutaneous closure in case of suspected paradoxical embolism through a patent foramen oval (PFO).
With the development of computerized tomography (CT) and magnetic resonance imaging (MRI) techniques, a specific imaging pattern of brain lesions for every etiologic type of acute stroke has been hypothesized.A number of studies have concluded that in cardioembolic strokes, lesions are mostly multiple and cortical in location (5)(6)(7), but a systematic description of the neuroimaging features associated with a cardioembolic genesis of ischemic strokes is presently lacking.In the past two decades, silent cerebral lesions have been described at MRI following AF transcatheter ablation (8).These lesions have been carefully identified by comparing imaging before and after the procedure with MRI diffusion-weighted imaging (DWI) sequences, the most sensitive technique for the detection of acute cerebral ischemia (9).
Because during an uncomplicated transcatheter AF ablation the patient does not experience hemodynamic compromise, we hypothesize that the pathophysiological mechanism underlying brain lesions is embolic, and not related to hypoperfusive genesis.The major mechanisms involved include endothelium damage at transseptal puncture, conventional clotting (e.g., from the groin) and crossed emboli in the iatrogenic interatrial septum defect, and thermal thrombus formation in the left atrium (charring and gas embolism).In addition, during a transcatheter ablation the introduction of bulky devices, such as multielectrode catheters, and balloons carries a risk of air embolism (10).
In the present study we conducted a systematic review and meta-analysis to describe the neuroimaging features of newly formed cardioembolic lesions after AF transcatheter ablation.

Methods
This work was conducted following the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (11).

Search strategy
Pertinent articles were searched in MEDLINE/PubMed with MeSH strategy, using the following terms: ((cerebral lesion* OR stroke OR silent cerebral lesion OR SCI embolism) AND cerebral MRI AND (Atrial fibrillation ablation OR Fib ablation OR AF ablation)).The search was ended on 30 January 2023.

Study selection and data extraction
Two independent reviewers (EB, MB) screened the retrieved citations through title and/or abstract.When potentially pertinent, the studies were appraised as complete reports according to the following inclusion/exclusion criteria: • They reported the absolute number of new ischemic cerebral lesions that occurred after AF transcatheter ablation; this implies that the included studies were designed to perform cerebral MRI scan prior to and immediately after the ablation procedure.• They reported at least one of the evaluated characteristics of the ischemic cerebral lesions (please refer to the following section regarding the specific study endpoints for detailed description of evaluated features).
Exclusion criteria were non-human setting; duplicate reporting (in which case, the manuscript reporting the largest sample of patients was selected).Two independent, unblinded reviewers (EB and MB) abstracted the following data on prespecified forms: authors, journal, year of publication, baseline clinical and interventional features, cerebral MRI protocol, and neuroimaging features.Data collection was conducted by mutual agreement and all potential disagreement was resolved by a third reviewer (AS).

Study endpoints
The incidence of new cerebral ischemic lesions after AF transcatheter ablation was the primary endpoint.The co-primary endpoints were the prevalence of the following neuroimaging features regarding new cerebral ischemic lesions: • Lesion dimension: size > 10 mm; • Multiple lesions; • Left vs. right hemisphere location; • Cortical, subcortical (white matter and/or basal ganglia), or cerebellar location; • Involved anatomical lobe: frontal, parietal, occipital, temporal; • Involved vascular territory: anterior cerebral artery (ACA), middle cerebral artery (MCA), posterior cerebral artery (PCA), and border zone (BZ).
The secondary endpoint was the incidence of symptomatic cerebral ischemic events (stroke/transient ischemic attack).

Statistical analysis
The baseline characteristics of the pooled study populations were reported as median values and their interquartile ranges (IQRs).The meta-analysis of the proportions (crude incidence of new cerebral ischemic lesions and prevalence of the prespecified neuroimaging features) was performed using a generalized linear mixed models (12) under a random-effect framework and the results were reported together with the corresponding 95% confidence interval (CI).Cochrane I 2 test was used to investigate heterogeneity, with I 2 values of 25%, 50%, and 75% representing, respectively, mild, moderate, and extensive heterogeneity.Statistical analyses were performed with R version 4.0.0 (R Foundation for Statistical Computing, Vienna, Austria) and p-values less than 0.05 were considered statistically significant.

Results
The initial search retrieved 125 studies.Among these, three studies were removed because of duplication, and one study was not assessed for eligibility because of the unavailability of the English translation.Given that 39 studies were not pertinent to the topic of the meta-analyses, 82 studies were assessed for eligibility, and 25 eventually included (13-37).Figure 1 reports the detailed PRISMA flowchart of the selection process.The resulting meta-analytic population encompassed 3,304 patients who had undergone AF transcatheter ablation and pre-and post-procedure cerebral MRI scans to detect new ischemic lesions.More details regarding the included studies, in particular concerning the adopted cerebral MRI protocol, the ablation technique, and the energy source are reported in FIGURE 1 PRISMA flowchart of the study selection process.The incidence of new ischemic cerebral lesions in patients that had undergone AF transcatheter ablation was 17.2% (95% CI 12.2%-23.8%;I 2 : 93%) (Figure 2A).Among these lesions, only a minimal proportion had ischemic symptoms [0.60% (95% CI 0.09%-3.9%;I 2 : 56%)] (Figure 2B).

Discussion
The main findings of the present analysis are as follows (Figure 4-Graphical Abstract): • Among the proportion of patients presenting a cerebral ischemic lesion at cerebral MRI after AF transcatheter ablation (17%), 0.6% are symptomatic from a neurological standpoint.• These cardioembolic cerebral lesions are generally balanced between the right and the left sides and are ubiquitously detected in all cerebral lobes; the lesions are typically small (diameter less  Determining the underlying cause of an acute stroke is important not only to guide patient's immediate management but also to prevent new events, given that stroke recurrence is strongly Frontiers in Cardiovascular Medicine 05 frontiersin.orgrelated to its specific etiology (38-40).Imaging has a primary role in early diagnosis of strokes: in fact, some patterns of brain infarction may suggest a specific cause.In particular, regarding ESUS, early identification of a potential cause underlying the ischemic event might be of paramount importance, since it could expedite the clinical decision process leading to the adoption of therapeutic strategies, such as initiation of oral anticoagulation in case of an AF-related genesis.Moreover, considering that patients with AF frequently present asymptomatic cerebral lesions (41), keeping in mind that AF-related subclinical lesions might be due to several mechanisms (42)(43)(44)(45)(46)(47), recognizing a certified neuroimaging pattern suggestive of subclinical AF-related lesions might even help prevent the occurrence of clinically relevant events.Despite these potential benefits, a systematic description of the neuroimaging features associated with a cardioembolic genesis of ischemic strokes is presently lacking.For this purpose, AF transcatheter ablation can be regarded as an in vivo model of cardioembolic lesions, used to derive a neuroimaging The present analysis suggests that typical cardioembolic lesions, such as occurring during an AF ablation procedure, tend to be located at a cortical level and, particularly, in the vascular territory of the middle cerebral artery.Cerebral lesions related to a hypoperfusive genesis, instead, more commonly are subcortical and determine watershed infarcts (50-52).Watershed infarcts occur at the border between cerebral vascular territories where the tissue is furthest from arterial supply and thus most vulnerable to hypotension and hypoperfusion, or might exacerbate embolism-related damage (e.g., delayed embolism "washout," impaired perfusion of ischemic penumbra) (53).In any case, this neuroimaging "fingerprint" differs quite clearly from that of cerebral lesions related to a cardioembolic genesis.

Limitations
The analysis is limited by the inherent limitations of a meta-analysis.In particular, although all studies performed diffusion-weighted imaging sequences to detect new cerebral lesions, hidden technicalities across the different studies cannot be excluded.Moreover, the present results might apply prevalently to a population of paroxysmal AF patients undergoing catheter ablation.In addition, although we selected a cut-off value of 10 mm to distinguish between smaller and larger lesions, we cannot exclude that different cut-offs (e.g., 3 or 5 mm) might be more appropriate and provide different trends.Finally, we cannot directly exclude that the populations included in studies from the same Research Groups might present partial overlaps.

Conclusions
In conclusion, by thoroughly assessing incidence and neuroimaging features of cerebral ischemic lesions following AF transcatheter ablation, it emerges that in predominance they consist of small (<10 mm) cortical lesions, almost ubiquitous in all cerebral lobes and in both hemispheres, prevalent in the territory of the middle cerebral artery.
The present analysis supports the existence of an, at least, suspicious neuroimaging "fingerprint" of cardioembolic brain lesions.If confirmed in specifically designed studies, specific neuroimaging features in de novo cerebral lesions would rapidly prompt a tailored clinical management, shortening time to diagnosis of the underlying etiology, and, potentially, preventing recurrences.
than 10 mm), generally single (multiple only in about 20% of the cases), and they preferentially affect the cerebral cortex (in nearly two-third of the cases) of the MCA vascular territory.

FIGURE 2 Forest
FIGURE 2Forest plot for (A) overall and (B) symptomatic periprocedural cerebral ischemic lesions incidence.

FIGURE 3 Forest
FIGURE 3Forest plot of the different neuroimaging features: (A) diameter more than 10 mm, (B) multiple lesions, (C) left hemisphere location, (D) cortical location, (E) frontal lobe location, and (F) middle cerebral artery territory location.

TABLE 1
General characteristics of the included studies.