Targeted ablation of gastric pacemaker sites to modulate patterns of bioelectrical slow wave activation and propagation in an anesthetized pig model

Abstract

Gastric motility is coordinated by underlying bioelectrical slow waves. Gastric dysrhythmias occur in gastrointestinal (GI) motility disorders, but there are no validated methods for eliminating dysrhythmias. We hypothesized that targeted ablation could eliminate pacemaker sites in the stomach, including dysrhythmic ectopic pacemaker sites. In vivo high-resolution serosal electrical mapping (16 × 16 electrodes; 6 × 6 cm) was applied to localize normal and ectopic gastric pacemaker sites in 13 anesthetized pigs. Radiofrequency ablation was performed in a square formation surrounding the pacemaker site. Postablation high-resolution mapping revealed that ablation successfully induced localized conduction blocks after 18 min (SD 5). Normal gastric pacemaker sites were eliminated by ablation (n = 6), resulting in the emergence of a new pacemaker site immediately distal to the original site in all cases. Ectopic pacemaker sites were similarly eliminated by ablation in all cases (n = 7), and the surrounding mapped area was then entrained by normal antegrade activity in five of those cases. Histological analysis showed that ablation lesions extended through the entire depth of the muscle layer. Immunohistochemical staining confirmed localized interruption of the interstitial cell of Cajal (ICC) network through the ablation lesions. This study demonstrates that targeted gastric ablation can effectively modulate gastric electrical activation, including eliminating ectopic sites of slow wave activation underlying gastric dysrhythmias, without disrupting surrounding conduction capability or tissue structure. Gastric ablation presents a powerful new research tool for modulating gastric electrical activation and may likely hold therapeutic potential for disorders of gastric function.

NEW & NOTEWORTHY This study presents gastric ablation as a novel tool for modulating gastric bioelectrical activation, including eliminating the normal gastric pacemaker site as well as abnormal ectopic pacemaker sites underlying gastric dysrhythmias. Targeted application of radiofrequency ablation was able to eliminate these pacemaker sites without disrupting surrounding conduction capability or tissue structure. Gastric ablation presents a powerful new research tool for modulating gastric electrical activation and may likely hold therapeutic potential for disorders of gastric function.

INTRODUCTION

The rhythmic bioelectrical activity of the stomach, called slow waves, is a central physiological mechanism governing gastric motility. Slow waves are initiated and coordinated by a network of interstitial cells of Cajal (ICC) (1, 2). In the healthy stomach, ICC form a syncytium that entrains to initiate slow wave activity from a single dominant pacemaker site located on the greater curvature of the upper corpus, where ICC have the highest intrinsic frequency (3, 4). Slow waves then propagate distally as rings that terminate at the pylorus (5). Conversely, abnormal slow wave activation patterns (termed dysrhythmias), including retrograde propagation from ectopic pacemaker sites, have been implicated in several gastrointestinal (GI) disorders including gastroparesis (6), chronic unexplained nausea and vomiting (7), and functional dyspepsia (8).

Therapeutic options remain limited for gastric dysfunction, and the ability to correct electrical dysrhythmias is a potential avenue toward novel therapies (9, 10). Emerging bioelectric interventions for modulating gastric electrophysiology have typically focused on gastric pacing and high-frequency gastric electrical stimulation (11). High-frequency gastric stimulation does not correct dysrhythmias (12, 13) but does appear to offer limited efficacy for symptom improvement, although the mechanism of action remains unclear (14, 15). Lower-frequency gastric pacing aims to control slow wave activation and has shown promising improvement in slow wave frequency and symptoms in some patients (16, 17) but awaits confirmation of effectiveness for correcting dysrhythmic activation patterns (11).

Radiofrequency (RF) ablation is a common therapy used to disrupt abnormal electrical activation in the heart. It functions by delivering targeted energy to the tissue at the site of an abnormality to form nonconducting scar tissue, and thereby preventing initiation from, and/or conduction through, that region (18). RF ablation techniques have recently been translated to the stomach, where validation studies have shown it is an effective method for inducing conduction blocks in gastric tissue (19, 20). Gastric ablation therefore offers a potential method for eliminating dysrhythmias in the stomach, such as ectopic pacemaker sites (i.e., sites of aberrant slow wave initiation).

In this study, we hypothesized that targeted RF ablation could be used to induce localized conduction blocks at the stable normal gastric pacemaker site and dysrhythmic ectopic pacemaker sites, to inhibit slow wave initiation and/or entrainment from those regions. RF ablation was targeted to sites of normal and ectopic pacemaker sites in the in vivo pig stomach, identified by baseline high-resolution electrical mapping, and postablation mapping was used to confirm that slow wave initiation and/or entrainment were eliminated from pacemaker sites.

METHODS

In Vivo Experimental Methods

Ethical approval was provided by the University of Auckland Animal Ethics Committee. Experiments were performed in vivo in 13 female cross-breed weaner pigs. Animal care was as previously described (21). In brief, the pigs were fasted overnight and then subjected to general anesthesia, induced with Zoletil and maintained with isoflurane. Vital signs were continuously monitored and maintained within normal physiological ranges. A midline laparotomy was performed, and the gastric serosal surface was exposed with minimal gastric handling. At the conclusion of the experiments, the animals were euthanized with a lethal bolus injection of pentobarbital sodium while still under anesthesia.

High-Resolution Electrical Mapping

High-resolution electrical mapping was performed using validated flexible-printed-circuit (FPC) electrode arrays (256 electrodes, 16 × 16 array, 4 mm spacing; FlexiMap, Auckland, New Zealand) (22, 23). The FPC array was gently positioned over the serosal region of interest and covered in warm (39°C) saline-soaked gauze packs to keep the organ moist and maintain gentle pressure. The wound edges were approximated with surgical clamps to limit cooling and/or drying of the abdominal cavity, and baseline slow wave activity was recorded for at least 5 min to find the position of a pacemaker site (either the normal pacemaker site or an ectopic site).

A pacemaker site was identified as the location where slow wave activity was initiated within the high-resolution electrode array and was further characterized with high-amplitude and rapid circumferential activity (Fig. 1A) (24). To identify the location of a pacemaker site, the FPC array was initially positioned at the midcorpus, adjacent to the greater curvature, on either the anterior or posterior surface of the stomach. After being recorded for at least 5 min, the likely location of slow wave initiation (i.e., pacemaker site) was determined based on the slow wave propagation direction, and the FPC array was repositioned toward the direction of slow wave initiation for the subsequent recording(s) until the location of slow wave initiation was within the area covered by the FPC array, thereby identifying the pacemaker site.

Figure 1.

High-resolution electrical mapping and targeted gastric ablation methods applied in female pigs (n = 13). A: electrograms from the electrode positions labeled in B, with slow wave ATs marked as red dots. The initiation point of slow waves marked with a star as the position of the gastric pacemaker site. B: FPC electrode arrays used for high-resolution electrical mapping (256 electrodes; 16 × 16 array, 4 mm spacing). C: RF ablation catheter performing ablation on the serosal surface of the stomach. D: placing the FPC array over the ablation lesion. E: gross image of an ablation lesion on the serosal surface of the stomach. The right border of the tissue section is the greater curvature of the stomach. F: schematic diagram of ablation shape (black rectangle), along the greater curvature or lesser curvature and their effective lesion area (red dashed area) for closed-box ablation (i), open-box against greater curvature (ii), and open-box against lesser curvature (iii). ATs, activation times; FPC, flexible printed circuit; RF, radiofrequency.

The pacemaker site was marked with a suture and then the FPC array was removed. Targeted gastric ablation was performed around the pacemaker site (Fig. 1C), as outlined below (Gastric Ablation). Immediately after ablation, the FPC array was again placed over the same region of interest and slow wave activity was recorded for at least 30 min. High-resolution mapping was repeated over the ablation site ∼2 h after ablation, for ∼15 min in duration, to confirm the postablation slow wave activity after a stabilization period (19).

Gastric Ablation

Gastric ablation was performed at the serosal surface using a Stokert-70 RF generator with a working frequency of 500 Hz (Biosense Webster, Irvine, CA). Ablation was applied using the temperature-control setting at 70°C without irrigation, with an ablation exposure time of 5 s per point, as previously described and validated to result in a gastric conduction block (19).

Ablations were performed in two different geometries (Fig. 1F): 1) in the shape of a “closed-box” (n = 7) where the pacemaker site was completely enclosed within a box-shaped ablation lesion, with the greater curvature as the line of symmetry (Fig. 1Fi) and 2) the shape of an “open-box” (n = 6) where the pacemaker site was partially enclosed within a three-sided box-shaped ablation lesion, with the fourth side of the box remaining nonablated and positioned along the greater curvature (Fig. 1Fii) or lesser curvature (Fig. 1Fiii). These two geometries were employed to investigate if total isolation is critical to eliminate the gastric pacemaker site (i.e., closed-box geometry), or if partial isolation is sufficient (i.e., open-box geometry). Ablations were also performed with two different “box” sizes: 1) a smaller effective size of ∼5 cm² for n = 7 cases and 2) a larger size of ∼9 cm² for n = 6 cases.

Electrophysiological Signal Acquisition and Analysis

Signals from the high-resolution electrical mapping were collected at 512 Hz using an ActiveTwo system (BioSemi, Amsterdam, The Netherlands) modified for passive electrodes. Signal analysis was performed in the validated Gastrointestinal Electrical Mapping Suite (GEMS; FlexiMap, Auckland, New Zealand) (25). Signals were first downsampled to 30 Hz and then filtered using a Gaussian moving median filter to remove baseline drift and a Savitzky-Golay filter (“low-pass”; effective cut-off of ∼2 Hz) to remove high-frequency noise (26). Slow wave activation times (AT; Fig. 1A) were identified and clustered using validated algorithms (25), followed by comprehensive manual review to ensure accuracy. Slow wave propagation was visualized using isochronal AT mapping (25). Amplitude, velocity, and frequency of slow waves were calculated and mapped using validated algorithms (25, 27). To calculate the amplitude and velocity in the region surrounding the pacemaker site, a subset of ∼15–25 electrodes were manually selected around the initiation point of slow wave activation (i.e., the pacemaker site).

Gross Image Evaluation

Immediately before euthanasia, the full visible ablation lesion was excised for gross image analysis and histological evaluation, including ∼5 mm of undamaged surrounding tissue. Excised tissue sections were photographed (Fig. 1E). The region of interest within the ablation lesion “box” was calculated using ImageJ software (28), in which the lesion border was manually segmented, delineating the change of color from pale pink (healthy) to white on the serosa.

To allow comparison of the size of ablation for all cases of both open and closed-box geometries, the effective lesion area per one side of the stomach (lesion “box” on anterior or posterior) was calculated (Fig. 1F). The “effective lesion size” was defined as the area of disconnected tissue on only the anterior or posterior surface of the stomach. In cases of closed-box geometries, which extended across both surfaces of the stomach, the larger of the two values was used.

Histological Analysis: H&E Stain

Excised tissue samples were cross-sectioned perpendicular to the ablation lesion and were fixed in 10% neutral buffered formalin and embedded in paraffin blocks. Tissue sections (5-μm thick) were stained with hematoxylin and eosin (H&E) to evaluate the degree of damage to the muscularis propria and mucosal layers. The cross-sectional area of the ablation lesion on the muscle and mucosal layers of H&E slides was calculated using ImageJ software (28). The lesion border was manually segmented, delineating along the change of tissue structure.

Immunohistochemistry

Paraffin-embedded tissue sections (5-μm thick) were deparaffinized and immersed in ethylenediaminetetraacetic acid (EDTA) buffer (pH 9.0) and placed in an antigen retriever. After the slides were washed in tris-buffered saline (TBS) for 10 min, endogenous peroxidase was neutralized with 3% hydrogen peroxide for 5 min and then washed in TBS. Subsequently, slides were blocked with bovine serum albumin (BSA; MP Biomedicals, Auckland, New Zealand) to block nonspecific protein binding, followed by application of c-kit-Cell Marque primary rabbit monoclonal antibody diluted at 1:50 (DAKO, Carpinteria, CA) for 1 h. Thereafter, Novolink polymer (Novocastra, Leica Biosystems, Buffalo Grove, IL) was applied for 30 min. Slides were treated with DAB (Novocastra, Leica Biosystems, Buffalo Grove, IL) for 5 min and with water. Slides were counterstained with CAT hematoxylin (Biocare Ltd, Birmingham, UK) for 2 min. Sections were dehydrated and coverslipped using Leica Biosystems-Surgipath Micromount mounting medium.

This antibody has been validated as a specific marker for ICC (29). In addition, a positive control slide of pig stomach was used to validate the specificity. A serial dilution of the antibody was tested and an optimal dilution of 1:50 was established. The positive control antigen expression was specific to the ICC and negative in the smooth muscle, validating the specificity, as seen in the literature. A negative control slide was also included without applying the antibody, but underwent all other steps of the immunohistochemical protocol.

Slides were imaged using the Metasystems VSlide scanner (Metasystems, Altlussheim, Germany) at the Biomedical Imaging Research Unit (BIRU), University of Auckland.

Statistical Analysis

Quantitative data are presented as mean and standard deviation (SD). Statistical differences before and after ablation were compared using paired Student’s t test, with a significance threshold of P < 0.05.

RESULTS

Experimental Data Set

We performed 13 ablation sequences in 13 pigs [46.4 kg (SD 9)], with over 900 min of total high-resolution mapping recordings, including baseline [n = 13; 13 min (SD 7)], immediately postablation [n = 13; 27 min (SD 8)], and 2 h postablation [n = 13; 16 min (SD 10)]. Both normal (n = 6) and ectopic (n = 7) pacemaker sites were identified and recorded during baseline mapping, as is typical in the weaner pig model (21, 30).

Elimination of Normal Pacemakers

Baseline.

In six cases, normal activation was observed during the initial baseline recordings. The normal antegrade slow wave activity had a frequency, amplitude, and velocity in the mapped area of 3.5 cpm (SD 0.6), 1.3 mV (SD 0.4), and 6.5 mm/s (SD 1.1), respectively, consistent with previous porcine studies (30, 31). High-resolution mapping identified a single pacemaker site in these cases, located on the greater curvature of the upper corpus (Fig. 2, A–E), consistent with the typical location of the normal gastric pacemaker site (3, 4, 30). Slow-wave activity in the region surrounding the pacemaker site displayed the expected anisotropic propagation pattern, with greater amplitude [2.4 mV (SD 1.1)] and higher velocity [8.9 mm/s (SD 2.2)] compared with antegrade propagation in the corpus [1.3 mV (SD 0.4), P = 0.02 and 6.5 mm/s (SD 1.1), P = 0.01]. Retrograde propagation from the normal gastric pacemaker site toward the fundus occurred for only a short distance (∼3 cm or less), with no propagation into the fundus, also consistent with previous descriptions of the normal gastric pacemaker site (3, 4, 30).

Figure 2.

Example high-resolution mapping results of targeted ablation of the normal pacemaker site, in a female pig. Ablation resulted in the elimination of the normal pacemaker site and activation of a new pacemaker site distal to the ablation region. A–E: baseline mapping showed normal pacemaker site in the high corpus, greater curvature. F–J: postablation mapping 60 min after gastric ablation showed that a conduction block was established at the location of the ablation lesion, which resulted in elimination of entrainment from the ablated normal pacemaker site. A: position of the FPC electrode array on the stomach. B: electrograms from the 8 electrode positions labeled in C, with slow wave ATs marked as red dots. C: isochronal AT map of slow wave propagation. Each color band indicates the area of slow wave propagating per 1 s from red (early) to blue (late). D: velocity map of the same slow wave, showing the speed (color spectrum) and direction (arrows) of the wave at each point on the array. E: amplitude map of the same wave. F: the electrode array position over the ablation region across the greater curvature (black box). G: electrograms from the electrode positions labeled in H, showing slow wave activity inside and outside of the ablation box with different frequency, and no discernible ATs in the electrodes at the ablation lesion. H–J: isochronal AT, velocity, and amplitude field maps, as explained above in C–E. Conduction block was present at the ablation lesion (black box), which inhibited entrainment from the normal pacemaker site inside the box. A new pacemaker site distal to the ablation region entrained the stomach, showing rapid, high-amplitude conduction. ATs, activation times; FPC, flexible printed circuit.

Postablation.

For all six cases of ablation of the normal pacemaker sites, ablation successfully eliminated the slow wave activity arising from the ablated site, confirmed by high-resolution mapping (Fig. 2, F–J, Supplemental Fig. S1; see https://doi.org/10.6084/m9.figshare.17064707.v1). Ablation of normal pacemaker sites exhibited an initial unstable transition phase, with activity initiating near or from the ablation site (Fig. 3, ii–iv), before a subsequent stable phase. After 4 min (SD 2), the ablated lesion formed a conduction block and after 15 min (SD 4) reached stable phase. Our results showed the establishment of a conduction block, where slow wave amplitude at the ablated tissue decreased to being too small to detect (Fig. 2G). Slow wave propagation inside the ablation box was blocked by the ablation, rendering that previous pacemaker site unable to activate and entrain the surrounding stomach, effectively creating an exit block at the original pacemaker site (Figs. 2 and 3) (32). A new pacemaker site emerged distal to the ablation sequence [1.8 cm (SD 1.0) distal], adjacent to the greater curvature, and entrained antegrade propagation down the stomach (Figs. 2 and 3v; Supplemental Fig. S1; Supplemental Fig. S2, see https://doi.org/10.6084/m9.figshare.17064755.v1; Supplemental Fig. S3, see https://doi.org/10.6084/m9.figshare.17064830.v1; and Supplemental Fig. S4, see https://doi.org/10.6084/m9.figshare.17064845.v1).

Figure 3.

Example high-resolution mapping results of targeted gastric ablation of a normal pacemaker site that resulted in the elimination of activation from that pacemaker site and a new pacemaker site distal to the ablation region, in a female pig. A: position of the FPC array on the stomach at baseline, 5 min, 15 min, 45 min, and 2 h after ablation. B: isochronal AT map as explained in Fig. 2 (isochronal interval = 1 s). C: velocity map of the same slow wave, as explained in Fig. 2. D: amplitude map of the same slow wave. i: baseline, normal pacemaker site with the rapid, high-amplitude circumferential activation initiating from high corpus, greater curvature. ii: postablation 5 min, slow wave activity initiating from the region close to the ablation, whereas conduction block is not yet established. iii and iv: postablation 15 and 45 min, respectively, showing time dynamic slow wave activity, with the pacemaker site switching position. v: postablation 2 h, stable pacemaker site distal to the ablation region activating the stomach. Black arrows indicate the direction of slow wave propagation. The black box indicates a conduction block at the ablation lesion, and translucent box indicates an unestablished conduction block. AT, activation time; FPC, flexible printed circuit.

The new pacemaker site showed significantly lower frequency compared with baseline [baseline 3.5 cpm (SD 0.6) vs. postablation 2.8 cpm (SD 0.8), P = 0.01; Fig. 4i], consistent with the known frequency gradient along the stomach (4). Across all six pigs, the average amplitude in the region surrounding the new pacemaker site (region with rapid, high-amplitude circumferential activity) was significantly lower than the amplitude of the region surrounding the pacemaker site in baseline [baseline 2.4 mV (SD 1.1) versus postablation 1.6 mV (SD 0.5), P = 0.04; Fig. 4i]. The velocity of the region surrounding the pacemaker site in baseline and postablation was not significantly different [baseline 8.9 mm/s (SD 2.2) versus postablation 8.5 mm/s (SD 2.7), P = 0.78; Fig. 4i]. The average amplitude and velocity of the antegrade propagation mapped away from the pacemaker site in baseline and postablation were not significantly different [1.3 mV (SD 0.4) versus 1.0 mV (SD 0.2), P = 0.16 and 6.5 mm/s (SD 1.1) versus 5.9 mm/s (SD 1.3), P = 0.49].

Figure 4.

Slow wave characteristics at baseline versus postablation, in female pigs (n = 13). i: measurements of cases of normal pacemaker sites (n = 6): frequency (A), amplitude (B), and velocity (C) in the region surrounding the pacemaker site. ii: measurements of cases of ectopic pacemaker sites (n = 7): frequency (D), amplitude (E), and velocity (F) in the region surrounding the pacemaker site. *P < 0.05, paired Student’s t test.

Elimination of Ectopic Pacemakers

Baseline.

In seven cases, baseline activity showed dysrhythmic slow wave patterns, initiating from a nonprimary or ectopic pacemaker site on the low corpus from either the greater curvature (n = 5) or lesser curvature (n = 2). An ectopic pacemaker site was defined as any location of slow wave activation outside of the established normal pacemaker site on the greater curvature of the mid-upper corpus (3, 33) and is also consistent with the use of the term “ectopic” by others in the gastrointestinal (34–36), and cardiac fields (37). The frequency, amplitude, and velocity in the mapped area were 3.6 cpm (SD 0.3), 6.5 mm/s (SD 0.9), and 1.3 mV (SD 0.3), respectively. Slow wave activity at the region surrounding the pacemaker site showed significantly higher velocity and amplitude than the rest of the mapped area [8.9 mm/s (SD 1.2) and 3.1 mV (SD 0.7), P < 0.01; Fig. 5, A–E] and was consistent with previous observations of rapid, high-amplitude circumferential activation near pacemaker sites in pigs and humans (6, 7, 24).

Figure 5.

Example high-resolution mapping results of targeted ablation of an ectopic pacemaker site, in a female pig. Ablation resulted in the elimination of the ectopic pacemaker site and restoration of normal antegrade slow wave activity. A–E: baseline mapping showed ectopic pacemaker site in the mid-low corpus, greater curvature. F–J: postablation mapping 60 min after gastric ablation showed that a conduction block was established at the location of the ablation lesion, which resulted in elimination of activation from the previous ectopic pacemaker site. A: position of the FPC electrode array on the stomach. B: electrograms from the 8 electrode positions labeled in C, with slow wave ATs marked as red dots. C: isochronal AT map as explained in Fig. 2 (isochronal interval = 1 s). D: velocity map of the same slow wave, as explained in Fig. 2. E: amplitude map of the same slow wave. F: the electrode array position, with the ablation lesion near the middle of the array (black box against greater curvature). G: electrograms from the electrode positions labeled in H, showing antegrade activation entraining around the ablation lesion, and no discernible ATs in the electrodes at the ablation lesion. H–J: isochronal AT, velocity, and amplitude field maps, as explained for baseline. Conduction block was present at the ablation lesion (black box), which inhibited activation from the ectopic pacemaker site. A new pacemaker site in the proximal region of the stomach entrained antegrade propagation down the stomach. Immediately distal to the block, rapid circumferential activity activated the excitable tissue to restore a transverse wavefront. ATs, activation times; FPC, flexible printed circuit.

Postablation.

Across all seven cases of ablation of an ectopic pacemaker site, high-resolution mapping successfully confirmed the establishment of a conduction block, where slow wave amplitude at the ablated tissue decreased to being too small to detect. Slow wave propagation inside the ablation box was blocked by the ablation, rendering that previous ectopic pacemaker site unable to activate and entrain the surrounding stomach, effectively creating an exit block at the ectopic pacemaker site (Fig. 5, F–J) (19, 32). However, the elimination of activation inside the box did not occur immediately, and instead occurred as a two-stage process of 1) an unstable transition phase immediately after ablation (Fig. 6, ii and iii), then 2) a stable phase where the conduction block was formed and the pacemaker site was eliminated (Figs. 5 and 6iv), or isolated within the ablation box with no entrainment of the surrounding region (Fig. 7). In the unstable transition phase immediately after ablation, slow wave activity was observed to be initiated from regions adjacent to the ablated tissue, exhibiting variable slow wave activation patterns (Fig. 6; Supplemental Fig. S5; see https://doi.org/10.6084/m9.figshare.17064851.v1). The ablated lesion formed a conduction block at a mean time of 5 min (SD 3) after the ablation was performed, and the activity reached a stable postablation phase at a mean time of 20 min (SD 6; n = 7). Over the transition period that occurred before the stable phase, slow wave activity switched between antegrade and retrograde propagation around the ablation box multiple times.

Figure 6.

Example high-resolution mapping results of targeted gastric ablation of an ectopic pacemaker site that resulted in the elimination of that ectopic pacemaker site and restoration of normal activity, in a female pig. A: position of the FPC array on the stomach at baseline, 5 min, 15 min, 45 min, and 2 h after ablation. B: isochronal AT map as explained in Fig. 2 (isochronal interval = 1 s). C: velocity map of the same slow wave, as explained in Fig. 2. D: amplitude map of the same slow wave. i: baseline, ectopic pacemaker site with the rapid, circumferential, high-amplitude activity initiating from low corpus, greater curvature. ii: postablation 5 min, slow wave activity initiating from the region close to the ablation, whereas conduction block is not yet established. iii: postablation 15 min, dynamic unstable phase with two pacemaker sites, one proximal to the mapping area and one close to the ablation region activating at the same time. iv: postablation 45 min, the proximal pacemaker site becomes the dominant pacemaker. v: postablation 2 h, repositioning the FPC array to the proximal region of the stomach confirmed the position of the dominant normal pacemaker site after ablation. Black arrows indicate the direction of slow wave propagation and the double black lines at the end of an arrow represent colliding waves. The black box indicates a conduction block at the ablation lesion, and translucency indicates an unestablished conduction block. AT, activation time; FPC, flexible printed circuit.

Figure 7.

Example high-resolution mapping results of targeted gastric ablation of an ectopic pacemaker site with a large box, that resulted in the elimination of entrainment from that ectopic pacemaker site and competing activation in antegrade and retrograde directions proximal and distal to the ablation region, in a female pig. A: position of the FPC array on the stomach. B: isochronal AT map as explained in Fig. 2 (isochronal interval = 1 s). C: velocity map of the same slow wave, as explained in Fig. 2. D: amplitude map of the same slow wave. i: baseline, ectopic pacemaker site with the rapid, high-amplitude circumferential activation initiating from low corpus, greater curvature. ii–iv: postablation slow wave activity of three consecutive waves (waves 1–3). Black arrows indicate the direction of slow wave propagation, the double black lines represent colliding waves, and the black box indicates a conduction block at the ablation lesion. AT, activation time; FPC, flexible printed circuit.

In the stable postablation phase, for 5/7 cases, a new pacemaker site activated the stomach and was located proximal to the ablation region (previous ectopic pacemaker site) on the greater curvature (Fig. 6; Supplemental Fig. S5). The new proximal pacemaker site and characteristics were consistent with that of the normal gastric pacemaker site on the greater curvature of the upper corpus, therefore suggesting that normal slow wave propagation was restored in these cases (3, 4, 30). The new proximal pacemaker site operated at significantly lower frequency compared with baseline [baseline 3.6 cpm (SD 0.3) versus postablation 2.7 cpm (SD 0.5), P = 0.0; Fig. 4ii] and initiated slow wave activity with rapid, high-amplitude circumferential activity in the region surrounding the pacemaker site, an established pattern of gastric pacemaker sites (6, 7, 24). The average amplitude in the region surrounding the new pacemaker site (region with rapid, high-amplitude circumferential activity) across five animals was 2.2 mV (SD 0.9), which was significantly lower than the amplitude of the region surrounding the ectopic pacemaker site in baseline [3.1 mV (SD 0.7), P = 0.01; Fig. 4ii]. Slow wave velocity in the region surrounding the pacemaker site in baseline and postablation was not significantly different [baseline 8.6 mm/s (SD 1.4) versus postablation 8.7 mm/s (SD 1.4), P = 0.97; Fig. 4ii]. The average amplitude and velocity of the longitudinal propagation mapped away from the pacemaker site were not significantly different [1.2 mV (SD 0.5) versus 1.0 mV (SD 0.3), P = 0.14; and 6.5 mm/s (SD 0.9) versus 6.1 mm/s (SD 1.0), P = 0.35].

For 2/7 cases, stable postablation recordings showed two pacemaker sites activating the stomach, one pacemaker site proximal to the ablation sequence at a frequency of 2.1 cpm (SD 0.1) and a second competing pacemaker site distal to the ablation sequence at a frequency of 1.3 cpm (SD 0.2). The slow waves from these two “competing” pacemaker sites collided adjacent to the ablation box (Fig. 7). These two cases had ablation boxes that were larger in effective size than the other five ablation boxes that resulted in a single proximal pacemaker site after ablation, suggesting that the effective size of ablation could be important to restoration of a normal proximal pacemaker site. The role of effective ablation size is discussed further below (see Size of ablated region).

Impact of Shape and Size of Ablated Region

Shape of ablation.

In this study, two ablation geometries were employed to eliminate activation and/or entrainment from pacemaker sites: 1) open-box (n = 6), with the open side of the box on either the greater curvature or lesser curvature and 2) closed-box ablation (n = 7), with the greater curvature as the line of symmetry (Figs. 1F and 8). Ablation resulted in elimination of activation and/or entrainment from pacemaker sites across all cases (n = 13), including both the open-box and closed-box geometries.

Figure 8.

Summary schematic diagram of postablation slow wave activation patterns resulting from the targeted ablation of gastric pacemaker sites (A: normal; B and C: ectopic), with different ablation shape and size. Bii and Cii represent restoration of normal antegrade conduction after elimination of the ectopic pacemaker by ablation, with minimal disruption to the antegrade propagation caused by the ablation box. Ciii showed the similar result, with additional damage.

After the transition phase, for cases with a closed-box ablation, the pacemaker site was isolated from the rest of the stomach. Slow wave activity from the new pacemaker site propagated around the ablation box to entrain the stomach, whereas the isolated tissue in the ablation box showed either 1) activity with different frequency (Figs. 2 and 7) or 2) no activity (Fig. 5). These results confirmed that ablation had effectively eliminated the initial pacemaker site (Figs. 2, 5, 7, and 8, ii and iii; Supplemental Figs. S1, S4, and Supplemental Fig. S6, see https://doi.org/10.6084/m9.figshare.17065031.v1).

The open-box geometry (n = 6) similarly resulted in effective elimination of the activation and/or entrainment from the pacemaker site within the box, even though tissue within the open-box was not completely isolated from the rest of the stomach (Figs. 3, 6, and 8, ii and iii; Supplemental Figs. S2, S3, S5).

Altogether, the stable postablation slow wave activation patterns of both open-box and closed-box ablation sequences were similar, with successful elimination of the initial pacemaker site and activity from the new pacemaker site entraining around the ablation region.

Size of ablated region.

To evaluate the impact of the surface area of the effective ablated region on the resultant slow wave activity, ablations for both ectopic and normal pacemaker sites were performed with an average effective ablated region area of two sizes: 1) a smaller area of 5.4 cm² (SD 0.5) for n = 7 cases, and 2) a larger area of 8.8 cm² (SD 1.2) for n = 6 cases [63% (SD 20) increase in size].

For ablation of normal pacemaker sites, the increase of the ablated region size from 5.5 cm² (SD 0.7; n = 3) to 8.5 cm² (SD 1.2; n = 3) did not show a significant effect on postablation measurements [3.6 cpm (SD 0.6) versus 2.3 cpm (SD 0.6), P = 0.16; 8.0 mm/s (SD 1.5) versus 8.6 mm/s(SD 2.0), P = 0.81; 1.4 mV (SD 0.1) versus 1.7 mV (SD 0.8), P = 0.57]. For all cases, postablation slow wave activity resulted in the emergence of a new pacemaker site immediately distal to the ablation box, with some retrograde propagation toward the upper corpus and fundus (Fig. 8A; Supplemental Fig. S4). However, an increase of the size of the ablation box resulted in retrograde propagation in a larger portion of the stomach (Fig. 8Aiii).

For ablation of ectopic pacemaker sites, the effective size of the ablated region had a significant effect on the resultant postablation slow wave propagation. For n = 4 cases with the effective size of 5.2 cm² (SD 0.3), elimination of the initial pacemaker site was successful and postablation activity resulted in antegrade propagation from a new pacemaker site proximal to the ablation box, consistent with the normal gastric pacemaker site (Figs. 5H, 6v, and 8, Bii and Cii, Supplemental Figs. S5 and S6). For the remaining cases (n = 3), the size of ablation was 9.4 cm² (SD 0.9). Although the elimination of the initial pacemaker site was successful for all ablations, for 2/3 cases, postablation slow wave activation occurred from two pacemaker sites, proximal and distal to the ablated region (Figs. 7 and 8Biii), suggesting that a larger ablated area could promote postablation dysrhythmias.

Histology

H&E and immunohistochemical analysis.

The qualitative evaluation of H&E-stained tissue sections confirmed the formation of a lesion that extended through the entire muscle layer across all cases with an average cross-sectional area of 17.7 mm² (SD 7.2) in the transmural plane (i.e., through the depth of the tissue; Fig. 9i). The ablation lesion remained localized to the ablation sequence with coagulation, necrosis, hemorrhage, and edema. A discernible border with a width of 117 µm (SD 56) separated tissue with structural damage and the adjacent normal tissue (Fig. 9i). The nonablated tissue within the ablation box had normal cellular integrity of muscle and mucosal structures (Fig. 9E). In 6/13 cases, ablation induced mucosal damage, including deformation of glands and extravasation of blood (Fig. 9).

Figure 9.

Example evaluation of acute tissue damage caused by ablation in a female pig stomach, using H&E stain for structural analysis and immunohistochemical (c-Kit) stain for ICC analysis at the ablation lesions. A and B: full-thickness view of the gastric tissue with ablation lesion. The region on the right side of the tissue section is the tissue that was inside the ablation box. The lesion region can be distinguished by the deformation of tissue structure. C and D: ablation lesion border on the muscle layer (marked with the dashed line). C: the lesion border is discernible by disruption of tissue structure, edema, elongated and swollen nuclei, extravasation. D: the cellularity of ICC across the ablation border in the muscle layer. E and G: tissue inside the ablation box showed properties of normal tissue (E) and ICCs exhibited consistent morphology (G). F and H: ablation resulted in full-thickness muscle lesion (F) and ICCs deformed with their cell bodies reduced compare with nonablated tissue (H). I: high-magnification view of a full-thickness section from the serosa (right) to mucosa (left). ICC, interstitial cells of Cajal.

ICC (c-kit⁺ cells) in the ablated tissue were altered, which resulted in the interruption of the ICC network through the ablation lesion (Fig. 9ii). Comparison of H&E and c-kit⁺ slides showed that the region of structural damage and disruption of ICC network corresponded directly with the similar discernible border at the ablation region (Fig. 9).

DISCUSSION

In this study, we investigated the potential of targeting gastric ablation to sites of slow wave initiation (i.e., pacemaker sites) in the in vivo porcine stomach. The major findings were: 1) targeted ablation successfully eliminated both normal and ectopic pacemaker sites by isolating and disrupting activation and/or entrainment from the pacemaker sites; 2) targeted ablation of the normal gastric pacemaker site resulted in the emergence of a new pacemaker site distal to the ablation area, adjacent to the greater curvature, occurring at a lower frequency; and 3) successful ablation of ectopic pacemaker sites resulted in activation of a new pacemaker site in the mid to high corpus, adjacent to the greater curvature, indicating restoration of the normal gastric pacemaker site. In addition, successful restoration of normal antegrade propagation after ablation of ectopic pacemaker sites depended on the effective area of the ablated region. Together, these results demonstrate the potential for gastric ablation to serve as a novel technique for eliminating sites of aberrant initiation occurring in gastric dysrhythmias.

The ablation of the normal dominant pacemaker site in the upper corpus of the greater curvature was an important aspect of this study because the pacemaker site in that region is known to be stable over time (3, 4, 30), whereas ectopic pacemaker sites can be stable or unstable (6, 7, 21). Ablation of the normal gastric pacemaker site resulted in disruption of the ICC network and effective elimination of activation and/or entrainment from the pacemaker site. Subsequently, a new pacemaker site emerged immediately distal to that location, showing the rapid, high-amplitude circumferential activity that is characteristic of activation patterns at gastric pacemaker sites (3, 7, 24). The frequency and amplitude of slow waves from the new pacemaker sites were significantly lower than the initial normal pacemaker sites at baseline, clearly revealing the intrinsic frequency gradient of gastric ICC networks (4, 33).

The ablation of ectopic pacemakers was shown to lead to restoration of normal antegrade slow wave activation, with return of the dominant normal gastric pacemaker site to the upper corpus of the greater curvature (3, 4, 30). After ablation of the ectopic pacemaker site, the new dominant pacemaker site had a lower frequency, according with previous considerations that ectopic pacemaker sites dominate the stomach by operating at the highest effective frequency in the syncytium (21, 35, 38). It is notable that the frequency differences between these ectopic and normal pacemaker sites can be less than 1 cpm, such that ectopic sources do not need to be in the significantly higher range of classically “tachygastric” frequency to induce large-scale retrograde propagation in the stomach.

This study establishes ablation as a potential therapeutic candidate for gastric dysfunction. Gastric electrical abnormalities have been observed in several disorders, including chronic nausea and vomiting syndromes and gastroparesis (39–41), and effective therapies are lacking with current approaches typically focusing on symptom control rather than targeting specific mechanisms (42). Retrograde slow wave activation, arising from ectopic sources of initiation and often at “normal” frequencies, is increasingly being recognized as a feature in a subset of patients suffering chronic gastric symptoms (7, 10, 43), making ectopic sources a legitimate focus of current therapeutic interest. The in vivo animal model employed in this study was a healthy pig, which are prone to show dysrhythmic slow wave activity in the baseline (21). These dysrhythmias allowed us to investigate the feasibility of altering gastric activation using ablation. The ability we have shown for gastric ablation to eliminate ectopic pacemaker sites is a promising advance because there are no other existing techniques for correcting these abnormalities (9). Gastric pacing has been explored as a potential therapy (16, 17, 44), but its effectiveness for correcting dysrhythmic patterns has not yet been confirmed (11). Further research is now required to investigate the potential effectiveness and respective roles of both gastric ablation and pacing, likely offering complementary strategies.

An interesting observation in this study was that the size of the ablated region impacted the resultant slow wave activity. For ablations of ectopic pacemaker sites, a larger area of disconnected tissue could result in two new competing pacemaker sites, proximal and distal to the ablation box, activating the stomach with different frequencies. This result indicates that a large ablation could induce uncoupling of the frequency gradients between these regions (4), likely due to the weaker ICC coupling in the region of the vicinity of the lesser curvature leading to termination of slow wave conduction at the sides of a large ablation field (4, 33). These observations are also consistent with feasibility studies of linear ablations across the stomach, where long circumferential ablations resulted in emergence of a distal ectopic pacemaker site rather than wavelet rotation of the advancing antegrade-propagating waves around the lesser-curvature side of the block (19). In comparison, a smaller ablation area in this present study was still effective in eliminating the pacemaker site but did not induce uncoupling, while also leaving room for coupled activation on the lesser-curvature side of the lesion. A smaller ablation box is therefore a preferable candidate option for therapy, enabling restoration of normal conduction pathways.

Another interesting finding was that both shapes of ablation employed in this study, open box and closed box, successfully eliminated activation and/or entrainment from pacemaker sites, and no significant difference in the final mapping results was observed. Ablations in the form of a closed-box resulted in complete isolation of the pacemaker site from the rest of the stomach, thereby eliminating it as an effective pacemaker site. However, cases with an open-box also resulted in termination of the activation and/or entrainment from the ablated pacemaker site even though that tissue region remained connected to the rest of the stomach through the “open” side of the box. This observation is currently unexplained, but one possible explanation for this termination of the pacemaker site via the open-box ablation geometry could be due to disruption or isolation of a critical mass of bioelectrical networks contributing to both normal and ectopic pacemaker sites (6, 7, 24), or the disturbance of the critical boundary conduction pathways that trigger the ectopic activation. Hence, future investigation is required to define the optimal ablation configuration with minimal damage to tissue.

This study provides valuable foundation for the feasibility of ablation to eliminate ectopic pacemaker sites in the stomach, but extensive further research is now needed to progress toward a viable therapy. The current study was limited to an acute trial with mapping of the postablation slow wave activity for a duration of up to 4 h after ablation. One important next step will be to conduct a series of recovery trials to establish the longer-term efficacy of ablation on gastric slow wave activity after a period of healing. Recovery studies are also needed to investigate the safety profile of gastric ablation, to address the risk of perforation, and to determine the impact of ablation on motility. Simultaneous high-resolution mapping on the serosal and mucosal surfaces has demonstrated transmural synchrony of conduction pattern across the gastric wall (31), but potential transmural synchrony after ablation should be investigated as this technique progresses toward clinical application. The anesthetized pig model used in this study exhibited both normal and dysrhythmic activity (21, 30), which is useful for investigations of dysrhythmia monitoring and intervention. However, anesthetic drugs are known to have a minor effect on slow wave frequency (45) and may have been responsible for a slight decrease in frequency observed during the studies. The effective area and shape of gastric ablation should also be optimized in future studies, to target dysrhythmias while minimizing unnecessary damage. In parallel, further work is also required to develop companion minimally invasive mapping techniques, such as body surface (46) and endoscopic mapping (31), to define and localize dysrhythmias and guide ablation. In addition, research is needed toward translation of the current method of serosal ablation to minimally invasive mucosal application.

Many parallels can be made between gastric and cardiac electrophysiology (47), and many similarities now exist between gastric and cardiac ablation as well. The terminology used to describe gastric ablation in this study has been chosen based on standard gastrointestinal terminology to date. The term “ectopic pacemaker” is commonly used in gastrointestinal electrophysiology to describe a site of electrical activation away from the primary pacemaker site (34–36, 48), translated from the terms “ectopic foci” or “ectopic beat” in cardiac electrophysiology (37, 49, 50). However, terminology like “nonprimary pacemaker” could be a useful alternative as this area of gastrointestinal research continues to develop in future. Similarly, cardiac terminology of “overdrive suppression” has potential application to describe entrainment dynamics in the stomach (51), “exit block” could be a useful term to describe the results of the ablation box procedure that we performed in this preliminary study of pacemaker ablation in the stomach (32), and concepts like wave front curvature that have been comprehensively developed in cardiac electrophysiology could usefully inform the continued development of gastric ablation (52).

In conclusion, this study establishes the feasibility of gastric ablation as a novel technique for eliminating sources of gastric slow wave initiation, including ectopic pacemaker sites underlying gastric dysrhythmias. This work could hold therapeutic potential in the management of gastric dysfunction.

SUPPLEMENTAL DATA

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.17064707.v1.

Supplemental Fig. S2: https://doi.org/10.6084/m9.figshare.17064755.v1.

Supplemental Fig. S3: https://doi.org/10.6084/m9.figshare.17064830.v1.

Supplemental Fig. S4: https://doi.org/10.6084/m9.figshare.17064845.v1.

Supplemental Fig. S5: https://doi.org/10.6084/m9.figshare.17064851.v1.

Supplemental Fig. S6: https://doi.org/10.6084/m9.figshare.17065031.v1.

GRANTS

These studies and/or authors were supported by the New Zealand Health Research Council, Royal Society Te Apārangi, and the National Institutes of Health DK057061, DK052766, DK123549, and AT010875.

DISCLOSURES

No commercial financial support was received for any material presented in this paper. L.K.C., N.P., S.J.A., G.F., G.O., and T.R.A-G. hold intellectual property and/or patent applications on gastrointestinal electrophysiology. L.K.C., N.P., and T.R.A-G. are shareholders in FlexiMap Ltd. G.O. is a director and shareholder in the Insides Company and Alimetry Ltd. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

Z.A., L.K.C., N.P., S.J.A., G.F., A.B., G.O., and T.R.A-G. conceived and designed research; Z.A., L.K.C., N.P., C-H.A.C., and T.R.A-G. performed experiments; Z.A., L.K.C., N.P., R.A., C-H.A.C., S.A., and T.R.A-G. analyzed data; Z.A., L.K.C., N.P., R.A., C-H.A.C., A.M., S.A., S.J.A., G.F., A.B., G.O., and T.R.A-G. interpreted results of experiments; Z.A., C-H.A.C., and T.R.A-G. prepared figures; Z.A., G.O., and T.R.A-G. drafted manuscript; Z.A., L.K.C., N.P., R.A., C-H.A.C., A.M., S.A., S.J.A., G.F., A.B., G.O., and T.R.A-G. edited and revised manuscript; Z.A., L.K.C., N.P., R.A., C-H.A.C., A.M., S.A., S.J.A., G.F., A.B., G.O., and T.R.A-G. approved final version of manuscript.