Tissue characterization of acute lesions during cardiac magnetic resonance-guided ablation of cavo-tricuspid isthmus-dependent atrial flutter: a feasibility study

Abstract Aims To characterize acute lesions during cardiac magnetic resonance (CMR)-guided radiofrequency (RF) ablation of cavo-tricuspid isthmus (CTI)-dependent atrial flutter by combining T2-weighted imaging (T2WI), T1 mapping, first-pass perfusion, and late gadolinium enhancement (LGE) imaging. CMR-guided catheter ablation offers a unique opportunity to investigate acute ablation lesions. Until present, studies only used T2WI and LGE CMR to assess acute lesions. Methods and results Fifteen patients with CTI-dependent atrial flutter scheduled for CMR-guided RF ablation were prospectively enrolled. Directly after achieving bidirectional block of the CTI line, CMR imaging was performed using: T2WI (n = 15), T1 mapping (n = 10), first-pass perfusion (n = 12), and LGE (n = 12) imaging. In case of acute reconnection, additional RF ablation was performed. In all patients, T2WI demonstrated oedema in the ablation region. Right atrial T1 mapping was feasible and could be analysed with a high inter-observer agreement (r = 0.931, ICC 0.921). The increase in T1 values post-ablation was significantly lower in regions showing acute reconnection compared with regions without reconnection [37 ± 90 ms vs. 115 ± 69 ms (P = 0.014), and 3.9 ± 9.0% vs. 11.1 ± 6.8% (P = 0.022)]. Perfusion defects were present in 12/12 patients. The LGE images demonstrated hyper-enhancement with a central area of hypo-enhancement in 12/12 patients. Conclusion Tissue characterization of acute lesions during CMR-guided CTI-dependent atrial flutter ablation demonstrates oedema, perfusion defects, and necrosis with a core of microvascular damage. Right atrial T1 mapping is feasible, and may identify regions of acute reconnection that require additional RF ablation.


Aims
To characterize acute lesions during cardiac magnetic resonance (CMR)-guided radiofrequency (RF) ablation of cavo-tricuspid isthmus (CTI)-dependent atrial flutter by combining T 2 -weighted imaging (T 2 WI), T 1 mapping, first-pass perfusion, and late gadolinium enhancement (LGE) imaging.CMR-guided catheter ablation offers a unique opportunity to investigate acute ablation lesions.Until present, studies only used T 2 WI and LGE CMR to assess acute lesions.

Methods and results
Fifteen patients with CTI-dependent atrial flutter scheduled for CMR-guided RF ablation were prospectively enrolled.Directly after achieving bidirectional block of the CTI line, CMR imaging was performed using: T 2 WI (n = 15), T 1 mapping (n = 10), first-pass perfusion (n = 12), and LGE (n = 12) imaging.In case of acute reconnection, additional RF ablation was performed.In all patients, T 2 WI demonstrated oedema in the ablation region.Right atrial T 1 mapping was feasible and could be analysed with a high inter-observer agreement (r = 0.931, ICC 0.921).The increase in T 1 values post-ablation was significantly lower in regions showing acute reconnection compared with regions without reconnection [37 ± 90 ms vs. 115 ± 69 ms (P = 0.014), and 3.9 ± 9.0% vs. 11.1 ± 6.8% (P = 0.022)].Perfusion defects were present in 12/12 patients.The LGE images demonstrated hyper-enhancement with a central area of hypo-enhancement in 12/12 patients.

Conclusion
Tissue characterization of acute lesions during CMR-guided CTI-dependent atrial flutter ablation demonstrates oedema, perfusion defects, and necrosis with a core of microvascular damage.Right atrial T 1 mapping is feasible, and may identify regions of acute reconnection that require additional RF ablation.

Introduction
When cardiac arrhythmias are treated with catheter ablation, the longterm success is determined by the ability to create transmural and continuous scar lesions.Although the electrophysiological validation of the ablation line for bidirectional block is the endpoint of most catheter ablation procedures, it remains challenging to distinguish between transient oedema and durable necrosis. 1][4][5][6][7] This innovative procedure can be performed in a hybrid lab combining fluoroscopy and CMR, [8][9][10] or in a fully transformed conventional CMR facility. 65][16] Of note, these human studies only assessed the acute ablation lesion using T 2 -weighted imaging (T 2 WI) and LGE imaging.This approach, nevertheless, failed to identify gaps in ablation line continuity despite a 56% clinical success rate. 168][19] However, until now, acute ablation lesions in patients undergoing CTI-dependent atrial flutter ablation have never been investigated using CMR techniques beyond T 2 WI and LGE.
The aim of this study is to evaluate tissue characteristics of acute ablation lesions in patients undergoing CTI-dependent atrial flutter ablation inside CMR by combining T 2 WI, T 1 mapping, dynamic imaging (first-pass perfusion during contrast infusion), and LGE imaging.

Methods
This clinical trial was approved by the Medical Research Ethical Committee (NL74812.068.20/METC20-064).All patients gave written informed consent.

Patient population and procedural workflow
Fifteen patients scheduled for CMR-guided catheter ablation of CTI-dependent atrial flutter because of symptomatic arrhythmia were prospectively enrolled.CMR-guided catheter ablation of the CTI was performed under general anaesthesia with endotracheal intubation, following a pre-defined procedural workflow. 6For the purpose of these procedures, the electro-anatomical research mapping (EAM) system (iSuite; Philips Healthcare, Best, the Netherlands) was implemented in our institute for the integration of anatomical and electrophysiological data acquired during the procedure.This software package enables active catheter tracking inside the heart to guide the ablation.The 3D whole heart dataset (Figure 1, panel A) is used to construct a 3D mesh model of the heart and extracardiac structures, which is combined with the method of 'active catheter tracking': the ablation catheters are equipped with two miniature MR receiver coils, which measure the co-ordinates of the ablation catheter.The Philips iSuite system can display the model and adjust the real-time slice to contain the catheter tip automatically, so that real-time MRI can be used immediately for device pose confirmation.Post-ablation imaging was performed directly after initially achieving bidirectional block of the CTI line.There was a pre-specified imaging protocol, with the aim of performing all CMR sequences displayed in Figure 1.However, the availability of general anaesthesia limited the total procedural time to strictly 3 h [from preprocedural patient preparation (e.g.intubation, vascular access, and electrical cardioversion when needed) until the patient is awake and detubated after the procedure].This is why in a proportion of patients, we had to curtail this pre-specified protocol, leaving out T 1 mapping and/or contrast-enhanced imaging.After a waiting period of a maximum of 30 min, the durability of the CTI line was re-evaluated.If an acute reconnection was observed, the electrical gap was identified during pacing via the diagnostic catheter in the coronary sinus and additional RF applications were performed to achieve permanent bidirectional block.During this re-evaluation of the ablation line, the electrophysiologists (S.M.C. and D.L.) were blinded to the results of the post-ablation CMR imaging.Importantly, the locations for the additional RF applications were based on the electrical findings alone (i.e.location of acute electrical reconnection) and not based on T 1 values.This blinding allowed us to correlate the periprocedural T 1 values to the regions with acute reconnection.CMR imaging analysis (T 2 W, T 1 mapping, LGE) was performed off-line after the procedure, and not integrated into clinical decision making to guide the procedure.

CMR imaging acquisition
All CMR imaging was performed on a clinical 1.5 T CMR system (Ingenia; Philips Healthcare, Best, the Netherlands).A high-resolution 3D wholeheart electrocardiogram (ECG)-gated respiratory-navigated steady-state free-precession (SSFP) sequence was performed as a roadmap for active catheter tracking throughout the procedure, and to localize the target ablation region.To image the acute ablation lesion, three standard slice orientations were used to visualize the CTI: right anterior oblique (RAO), left anterior oblique (LAO), and transversal view.RAO and LAO correspond to orientations used in conventional fluoroscopy procedures.A stack of three consecutive slices without slice gap was used to reveal all tissue changes within the RF ablation region.For optimal image quality, the T 2 WI, T 1 mapping and cine images were performed during end-expiratory ventilator stops (of the intubated patient under general anaesthesia).To determine the CTI length and thickness in all cardiac phases, an ECG-gated breath-held balanced SSFP cine image was obtained in the RAO view.For T 2 WI, a black-blood T 2 -weighted spectral pre-saturation with inversion recovery (SPIR) sequence was performed in the RAO and transversal view.Native T 1 mapping was acquired in the RAO view using an ECG-gated single-shot 5(3)3 modified Look-Locker inversion-recovery (MOLLI) sequence.Post-ablation imaging prior to contrast infusion consisted of T 2 WI and T 1 mapping with scan parameters and imaging planes identical to pre-ablation imaging.First-pass perfusion imaging was performed during infusion (0.2 mmol/kg) of gadobutrol (Gadovist; Bayer Pharmaceuticals, Berlin, Germany) with a stack of 3-5 consecutive slices in the RAO view (the patient's heart rate determined the available acquisition window for this dynamic imaging).Dark-blood LGE images were acquired 10 min after contrast infusion in the RAO and transversal view using a standard ECG-triggered breath-held phase-sensitive inversion-recovery (PSIR) sequence.The inversion time was set to null the signal of the blood pool for optimal dark-blood contrast to improve detection of small scar regions.The mechanism of this used dark-blood LGE method (blood-nulled PSIR LGE) has been described in detail before. 20,21All LGE images were acquired in mid-diastole during end-expiratory ventilator stops.The given contrast dose reflects local protocol and current international guidelines. 22Typical acquisition parameters of all CMR techniques, the lengths of breath-holds per sequence are listed in Supplementary data online, Table S1, as well as the estimated time-interval of the pre-specified CMR imaging protocol.However, the time duration between each RF application and the CMR assessment of that single lesion will range depending on the number of RF ablation lesions and time to complete the ablation line.

CMR imaging analysis
Two cardiovascular imaging experts (C.M.; radiologist, and G.P.B.; cardiologist, with 13 and 4 years CMR experience, respectively) reviewed the images on a dedicated workstation (Sectra IDS7, Linköping, Sweden).Post-processing was performed using the IntelliSpace Portal (Version 7.0.1;Philips Healthcare, Best, the Netherlands).T 1 mapping was assessed by a third independent reader (B.M.M.; radiologist, with 5 years of CMR experience) in addition to the assessment of the first reader (G.P.B.).A fourth observer (H.M.J.M.N.), blinded to the CMR results, retrospectively reviewed which region(s) had received additional RF lesions because of acute reconnection.T 2 WI was assessed visually and quantified by the myocardial oedema ratio (ER).ER was defined as the ratio between the signal intensity (SI) of the myocardium within the manually delineated ablation region and the SI of reference skeletal muscle. 23An ER above 2.0 is considered indicative of oedema. 24For T 1 mapping, the T 1 times were measured in three regions of the ablation line (i.e.tricuspid annulus, mid region, and near the inferior caval vein) with a standardized approach.As an example, this resulted in 30 regions assessed with T 1 mapping when 10 patients are evaluated.For each region, the change in T 1 value was expressed as the absolute increase (in ms) and the relative increase (in %) post-ablation vs. pre-ablation.For dynamic perfusion imaging during contrast administration, all slice levels (three to five slices in RAO view) were visually assessed by the two independent readers (C.M. and G.P.B.).The acquired perfusion images were compared with the corresponding 2D LGE images (in RAO view), and the reference line on transversal 2D LGE image to compare the extent of the perfusion defect to the LGE area.The morphology of the isthmus was assessed as straight (isthmus depth ≤ 2 mm), concave, or pouch-like (isthmus concave and depth > 5 mm), using the latest atrial diastolic frame (confirmed by the opening of the tricuspid valve in the next frame). 25

Statistical analysis
Continuous variables were expressed as mean ± standard deviation.Categorical variables were expressed as numbers and percentages.T 1 Tissue characterization of acute ablation lesions mapping variables were compared between two groups (regions that did require additional RF applications, and regions that did not) using the Student's t-test.Inter-observer agreement for T 1 mapping was determined using the Pearson's correlation test and Blant-Altman plots.A two-sided P-value of <0.05 was considered significant.Statistical analysis was performed using SPSS Statistics version 28 (IBM, Armonk, NY, USA).

Patient population and procedural characteristics
Baseline patient demographics are demonstrated in Table 1.CMR-guided catheter ablation of CTI-dependent atrial flutter was completed successfully with a bidirectional conduction block in all 15 patients.The mean total procedural time was 169 ± 19 min.The median amount of primary RF applications per patient was 19 [range 12-40].In nine patients, acute reconnection occurred within the waiting period; the median amount of additional RF applications of these patients was 7 [range 3-14].T 1 mapping was performed in 10/15 patients, in whom 30 regions were evaluated (three regions per ablation line).In the subgroup that had T 1 mapping analysis, a total of 11 out of 30 regions showed acute reconnection (see Supplementary data online, Table S2).At initial follow-up, 3 out of 15 patients had a recurrence of CTI-dependent atrial flutter, although the postprocedural interval is still short in some patients (range 2-30 months, median 23 months).

Imaging characteristics of the acute ablation lesions
Post-ablation imaging was performed with T 2 WI (n = 15), T 1 mapping (n = 10), first-pass perfusion imaging during contrast infusion (n = 12), and LGE (n = 12).The reason to curtail the CMR protocol in patients was procedural time restraints in all cases.Figure 1 visualizes the integration of all CMR imaging sequences to assess the tissue characteristics of the ablation region.Table 2 summarizes the imaging results of the study population.The patient-specific CMR results and the distribution of locations with acute reconnection are provided in Supplementary data online, Table S2.In all patients, T 2 WI demonstrated oedema in the ablation region.The mean myocardial oedema ratio (ER; >2.0 indicating oedema) increased from 1.6 ± 0.2 pre-ablation to 3.2 ± 0.6 postablation (P < 0.001).In each patient, the ER was ≤2.0 pre-ablation and increased to >2.0 post-ablation.Right atrial T 1 mapping was feasible in 29/30 regions (in one case artefacts from the adjacent catheter hampered analysis in one out of the three regions).The regions that showed acute reconnection had a significantly lower increase in T 1 value post-ablation in comparison to regions that showed no acute reconnection [37 ± 90 ms vs. 115 ± 69 ms (P = 0.014), and 3.9 ± 9.0% vs. 11.1 ± 6.8% (P = 0.022), respectively].The measured post-ablation T 1 value itself was not significantly different between the regions that did and did not require additional RF ablation [1090 ± 89 ms vs. 1167 ± 166 ms (P = 0.169)].The boxplots are presented in Figure 2. The visual assessment of colour-coded T 1 maps also suggested patchy variation within the ablation line (Figure 3).The T 1 mapping analysis resulted in a high inter-observer agreement (r = 0.931, ICC 0.921).The correlation plot and Bland-Altman plot for T 1 mapping are shown in  The LGE images demonstrated hyper-enhancement at the ablation region with a central area of hypo-enhancement in 12/ 12 patients (Figure 1 and Supplementary data online, Video S2).Hypoperfusion was visible on the slice matching the area of LGE, as well as on slices corresponding to the myocardium outside the hyperenhanced area (see Supplementary data online, Video S1).T 2 WI suggested an increase in wall thickness of 198 ± 38%.The central isthmus was straight in 1/15 patients (7%), concave in 11/15 patients (73%), and pouch-like in 3/15 patients (20%).In every patient, the observed thicker atrial wall post-ablation coincided with a reduction of isthmus depth.In the three patients with pouch-like isthmus morphology, this resulted in an isthmus depth < 5 mm after ablation, i.e. below the threshold of the definition of pouches.

Discussion
This study is the first to assess tissue characteristics of acute cardiac RF ablation lesions using T 1 mapping and first-pass perfusion, in addition to T 2 WI and LGE.
In this study, we demonstrated the feasibility of acute lesion assessment by atrial T 1 mapping after CMR-guided CTI ablation, which could be analysed with a high inter-observer agreement.We showed that the increase in T 1 value after ablation (i.e. the delta between pre-and post-ablation T 1 values of that region) was significantly smaller in regions that showed acute reconnection, when compared with regions that did not show acute reconnection.This suggests that CMR imaging with T 1 mapping has the potential to identify region with acute reconnection, in addition to the current gold standard (i.e.electrical gaps identified during electrophysiological studies).Whether it has additional clinical value to identify areas with late reconnection (and thus might improve long-term ablation success) still has to be investigated.The fact that the post-ablation T 1 value itself (without comparing it to the pre-ablation value of that same region) did not differ between regions that did and did not show acute reconnection means that the imaging sequence needs to be repeated pre-and post-ablation, only acquiring T 1 mapping post-ablation would not suffice.T 1 mapping has been extensively studied in the ventricles, with histological evidence of interstitial myocardial fibrosis and acute oedema, [26][27][28][29] however studies on atrial T 1 mapping are sparse.Only two studies have described native T 1 mapping in the left atrium, 30,31 while no studies have investigated T 1 mapping in the right atrium (RA).One of the main challenges of T 1 mapping when applied to the atria is the partial volume effect.The thin atrial wall risks including the blood pool and/or epicardial structures into the voxel.Furthermore, the complex structure of the atrial wall (muscle, fibrosis, fatty depositions) influences the detected T 1 value within one voxel.The tissue characteristics of fat have been shown to outweigh that of fibrosis in an experimental study of the left atrium. 30Animal studies investigating acute RF ablations in the ventricle showed the ability of T 1 -weighted imaging to separate necrotic  cores from surrounding oedema.They demonstrated T 1 shortening (without gadolinium) correlated with the presence of ferric iron in the ablation core, while bipolar voltages were not distinct. 1,17,19nterestingly, the preliminary results in the present study show predominantly regional increase of the T 1 value after ablation.Our study therefore suggests that extrapolation of these pre-clinical results from the ventricles to the human atria is not suitable.The partial volume effect and complex atrial wall structure play a substantial role in imaging the thin-walled atria, affecting in particular the T 1 value measurements.
The T 1 value is known to be higher in areas of myocardial oedema, and lower in areas of ferric iron.One hypothesis is that the relative abundance of oedema in the thin atrial wall outweighs the small area of ferric iron in the atrial ablation core.Whether the regions with a decrease in T 1 value post-ablation indicate a different ratio between oedema and ferric iron in that part of the ablation line, or are explained to a certain extent by partial volume effect incorporating structures beyond the atrial wall, needs further investigation.
The presence of intacardiac catheters may affect image quality due to artefacts, particularly for parametric mapping (Figure 3C indicates the single catheter-induced artefact, with the catheter placed at the ridge).We encountered no image artefacts during T 1 mapping of the RA CTI when the two catheters resided in the lumen of the RA and coronary sinus.Previous animal studies already showed unaffected T 2 WI during ablation at the site of the catheter. 12Motion artefacts were limited by CMR acquisition during breath-holds of the sedated and intubated patient, and all post-ablation imaging was performed during pacing from the catheter in the coronary sinus, allowing a stable RR-interval.
First-pass perfusion during contrast infusion demonstrated perfusion defects in the ablation region of all patients.The perfusion defect occurs on slices outside the area of LGE, which suggests a border zone of tissue alteration beyond the necrotic area as suggested by a previous animal study. 19wo-dimensional LGE imaging of the acute ablation lesion demonstrated hyper-enhancement with a central area of hypo-enhancement at the site of RF ablation in all patients, which is consistent with previously reported areas of microvascular damage and haemorrhage after RF ablation in animal studies. 4,11,17,32The pathophysiological mechanisms underlying the area of hypo-enhancement are still unclear.The most established hypothesis is that the destroyed microvasculature prevents gadolinium to be present at the central ablation core, while there is a peripheral rim of hyper-enhancement due to contrast entering the ablation lesions via diffusion from the lesion periphery. 11,33Histological analysis of ablation lesions in the left ventricle of pigs demonstrated that the peripheral rim of hyper-enhancement reflects oedema in the border zone of the acute lesion, which would fit studies reporting LGE areas to be larger in the acute phase than in chronic ablation scar. 34In our study, 2D LGE was performed, despite its lower spatial resolution, as a useful alternative to time-consuming 3D LGE.This resonates with a study that compared 2D LGE and 3D LGE to assess chronic CTI ablation scar. 35However, with the introduction of novel strategies, 3D LGE image acquisition can be substantially accelerated, and therefore may become an interesting alternative for 2D LGE imaging.

Study limitations
The study population is small, albeit worldwide the second largest study in this innovative field.][16] The post-ablation imaging was performed during the 'waiting period', the CMR scans were not repeated after the re-application of additional ablation lesions in case of acute reconnection.As such, the electrophysiological findings at long-term (i.e.potential ablation gaps at redo procedures) cannot be directly compared with the imaging characteristics at the end of the procedure.We could not complete the pre-specified imaging protocol in every patient, which may introduce bias, although in all cases, this was due to procedural time restraints set by the availability of general anaesthesia.Repetitive non-contrast imaging (preferably after every (additional) RF application) to assess the tissue changes over time would be desirable, but the clinical procedural workflow did

Conclusion
CMR-guided catheter ablation offers the unique opportunity of therapy guidance and immediate therapy evaluation.T 1 mapping of the acute CTI ablation lesion is feasible, can be analysed with a high interobserver agreement.The distribution of locations with acute reconnection is significantly correlated to the regions with a smaller increase in T 1 value post-ablation (i.e. the delta between pre-and post-ablation T 1 values of that region).When confirmed in larger studies, this would be the first CMR imaging tool to identify regions of acute reconnection directly after ablation.Characterization of the acute lesion during CTI-dependent atrial flutter RF ablation

Figure 1
Figure 1 Overview of CMR sequences to assess the tissue characteristics of the ablation region, in which the arrows indicate the CTI ablation region, extending from the tricuspid annulus towards the inferior caval vein: (A) 3D whole-heart SSFP (one frame from 3D dataset), (B) SSFP cine imaging in RAO view (still end-diastolic frame), (C + G) T 2 WI at the CTI (arrow) demonstrates higher signal intensity and increased wall thickness on post-ablation imaging (G) compared with pre-ablation (C), (E) first-pass perfusion during contrast infusion shows persisting hypo-enhancement in the ablation area (still frame from Supplementary data online, Video S1), (F) 2D dark-blood LGE image with hyper-enhancement in the ablated area and a central area of hypo-enhancement.,(D+ H ) T 1 mapping is performed pre-and post-ablation (detailed figure of a qualitative assessment is shown in Figure 3); CMR, cardio magnetic resonance; CTI, cavo-tricuspid isthmus; LGE, late gadolinium enhancement; SSFP, steady-state free-precession; T 2 WI, T 2 -weighted imaging.

Figure 4 .
Figure4.The estimated bias in measured T 1 times was 12 ms (limits of agreement −102 and 126 ms).First-pass perfusion imaging demonstrated perfusion defects in 12/12 patients (see Supplementary data online, Video S1).The LGE images demonstrated hyper-enhancement at the ablation region with a central area of hypo-enhancement in 12/ 12 patients (Figure1and Supplementary data online, Video S2).Hypoperfusion was visible on the slice matching the area of LGE, as well as on slices corresponding to the myocardium outside the hyperenhanced area (see Supplementary data online, Video S1).T 2 WI suggested an increase in wall thickness of 198 ± 38%.The central isthmus was straight in 1/15 patients (7%), concave in 11/15 patients (73%), and pouch-like in 3/15 patients (20%).In every patient, the observed thicker atrial wall post-ablation coincided with a reduction of isthmus depth.In the three patients with pouch-like isthmus morphology, this resulted in an isthmus depth < 5 mm after ablation, i.e. below the threshold of the definition of pouches.

Figure 2
Figure 2 Boxplots indicating the distribution of T 1 mapping values of regions with and regions without additional radiofrequency lesions.Left: absolute T 1 value of the region post-ablation.Middle: increase in absolute T 1 value after ablation.Right: percentage increase in T 1 value after ablation.

Figure 3
Figure 3 Qualitative assessment of T 1 mapping, examples of three patients (A-C).Anatomical images (a, b) and the corresponding colour-coded T 1 maps (c, d) pre-ablation and post-ablation.The outlined target ablation area in image a-d corresponds to the zoom view in the bottom row e-h, respectively.In the target region, the blue areas (low T 1 values) suggest fatty structures within the right atrial wall, and green areas the thin endocardium.On the post-ablation colour-coded image (d), there are also areas of yellow and red within the target ablation area suggestive of higher T 1 values.The arrow in C indicates the artefacts from the adjacent catheter.

Figure 4
Figure 4 Inter-observer agreement of T 1 mapping.Blant-Altman plot (above) with the mean T 1 value of the two observers and the inter-observer difference, bias = 12 ms, the limits of agreement −102 and 126 ms.Scatter plot (below) of the measured T 1 values by both observers, with the line of perfect agreement.

Table 1 Baseline patient demographics
DS 2 -Vasc score = clinical risk score to evaluate thrombo-embolic risk in atrial flutter patients, where a score ≥ 2 indicates high risk.
a CHA 2