Localized bioelectrical conduction block from radiofrequency gastric ablation persists after healing: safety and feasibility in a recovery model

Abstract

Gastric ablation has demonstrated potential to induce conduction blocks and correct abnormal electrical activity (i.e., ectopic slow-wave propagation) in acute, intraoperative in vivo studies. This study aimed to evaluate the safety and feasibility of gastric ablation to modulate slow-wave conduction after 2 wk of healing. Chronic in vivo experiments were performed in weaner pigs (n = 6). Animals were randomly divided into two groups: sham-ablation (n = 3, control group; no power delivery, room temperature, 5 s/point) and radiofrequency (RF) ablation (n = 3; temperature-control mode, 65°C, 5 s/point). In the initial surgery, high-resolution serosal electrical mapping (16 × 16 electrodes; 6 × 6 cm) was performed to define the baseline slow-wave activation profile. Ablation (sham/RF) was then performed in the mid-corpus, in a line around the circumferential axis of the stomach, followed by acute postablation mapping. All animals recovered from the procedure, with no sign of perforation or other complications. Two weeks later, intraoperative high-resolution mapping was repeated. High-resolution mapping showed that ablation successfully induced sustained conduction blocks in all cases in the RF-ablation group at both the acute and 2 wk time points, whereas all sham-controls had no conduction block. Histological and immunohistochemical evaluation showed that after 2 wk of healing, the lesions were in the inflammation and early proliferation phase, and interstitial cells of Cajal (ICC) were depleted and/or deformed within the ablation lesions. This safety and feasibility study demonstrates that gastric ablation can safely and effectively induce a sustained localized conduction block in the stomach without disrupting the surrounding slow-wave conduction capability.

NEW & NOTEWORTHY Ablation has recently emerged as a tool for modulating gastric electrical activation and may hold interventional potential for disorders of gastric function. However, previous studies have been limited to the acute intraoperative setting. This study now presents the safety of gastric ablation after postsurgical recovery and healing. Localized electrical conduction blocks created by ablation remained after 2 wk of healing, and no perforation or other complications were observed over the postsurgical period.

INTRODUCTION

Gastrointestinal (GI) bioelectrical activity, called slow waves, is an essential mechanism for the coordination of GI motility. Slow waves are generated by interstitial cells of Cajal (ICC), which lie within and between smooth muscle layers of the GI tract (1). Gastric ICC generate slow waves at specific intrinsic frequencies, with the frequency gradient decreasing in the oral-anal direction and greater curvature to lesser curvature (2, 3). In the healthy stomach, slow waves initiate from a dominant pacemaker region on the greater curvature of the upper corpus, where the ICC network has the highest intrinsic frequency, and propagate as rings of activation that terminate at the pylorus (3–5).

Conversely, disordered slow-wave activation patterns, termed dysrhythmias, are associated with a spectrum of gastric disorders including gastroparesis (6, 7), chronic unexplained nausea and vomiting (8, 9), and functional dyspepsia (7, 10). Limited interventional options exist for gastric dysfunction, with therapies primarily focusing on pharmaceutical interventions to control symptoms, with variable success. The potential of correcting electrical dysrhythmias has emerged as a possible interventional avenue, inspired by electrophysiological advances in cardiology. Emerging techniques of gastric pacing and gastric ablation aim to treat disorders by altering and manipulating the underlying electrical activity, although pacing is yet to demonstrate sustained efficacy in eliminating dysrhythmias (11–13).

Gastric ablation is an emerging technology in the GI field that has demonstrated efficacy in altering slow-wave activity and eliminating dysrhythmias in the in vivo stomach (14, 15). Gastric ablation induces localized lesions in the gastric musculature by delivering radiofrequency (RF) energy to form nonconducting scar tissue. That scar tissue serves as a localized conduction block of slow-wave activation, which can be targeted to disrupt sites of dysrhythmia in the stomach by high-resolution electrical mapping after ablation, as is routinely performed clinically in the heart (14). Gastric ablation has previously demonstrated the capability to target regions of ectopic gastric pacemakers and eliminate abnormal initiation of slow waves by inducing localized conduction blocks in acute in vivo studies (15). However, the longer-term feasibility of ablation to modulate gastric slow-wave activity after healing, and the safety profile of gastric ablation, is unknown.

Two concerns for the translatability of gastric ablation toward further research and clinical trials relate to the recoverability of the procedure, and more specifically that: 1) ablation may result in perforation of the stomach in the postsurgical period and 2) healing of the ablation lesion may result in reconnectivity of the induced conduction block, for example, through reconnectivity of ICC through the ablation scar. These potential risks of gastric ablation will only manifest in the postsurgical healing period and were therefore unable to be investigated in the acute porcine model used for bioelectrical studies of gastric ablation to date (14, 15). Previous studies of thermal and nonthermal GI wall injuries, including perforation, suggest that diagnosis of such injuries occurs within 7 days postsurgery (16, 17). Studies of surgical gastric transection and GI anastomosis have shown that temporary disruptions to slow-wave rhythmicity recover within 24 h (18) and bioelectrical reconnectivity occurs across the healed gastrotomy or anastomosis scar within 2 wk (4, 19, 20). Therefore, a duration of 2 wk was selected for an initial series of gastric ablation recovery studies.

We hypothesized that gastric ablation can be performed without resulting in perforation in the postsurgical recovery period, and unlike surgical gastrotomy or anastomosis, the wider thermal scar formed by ablation can create a sustained conduction block through which slow-wave propagation does not recover. We used gastric ablation at validated settings (14, 15), in a 2 wk recovery model, to investigate the risk of perforation caused by ablation and the sustainability of localized conduction blocks after this initial healing period. The results of this study help to define the safety and feasibility of gastric ablation as a translational technology, now enabling long-term, targeted intervention of gastric slow-wave propagation.

MATERIALS AND METHODS

Animal Care and Surgery

Ethical approval was obtained from the University of Auckland Animal Ethics Committee. Weaner, cross-bred, female pigs arrived in groups of two, to allow for a paired group of control (sham ablation) and experimental animal (RF-ablation). They were housed in adjacent floor pens for 7–8 days to allow for acclimation to the housing, feeding, and research personnel (Fig. 1A). Throughout the study, animals were weighed weekly to monitor their growth and fed standard pellet pig feed (twice daily) according to their weight (500 g/10 kg/day), with free access to water. The pigs were then randomly allocated into the control versus experimental group.

Figure 1.

Methods for in vivo gastric ablation recovery study. A: time line of in vivo experiments. B: illustration of the mapping and ablation protocol in initial and terminal surgery. The position of ablation (RF or sham) on the serosal surface of the stomach marked as pink line. C: illustration of FPC electrode arrays used for high-resolution electrical mapping (256 electrodes; 16 × 16 array, 4 mm spacing) placed over the ablation. D: example of electrograms from 32 electrode proximal and distal to RF-ablation in the initial recovery surgery. The time interval between slow waves in the proximal and distal regions confirms the formation of the conduction block. In addition, the direction of slow waves in the proximal region is antegrade and in the distal region is retrograde. E: RF-ablation on the anterior surface of the stomach immediately after performing the ablation in the initial recovery surgery (marked by dashed line). F: gross image of an RF-ablation lesion on the serosal surface of the stomach after the terminal surgery. The lesion is hardly visible; however, scar tissue is forming, as characterized by histology. G: gross image of a transverse frozen-section of an RF-ablation lesion after the terminal surgery. The tissue color changed from pale pink (healthy) to red at the RF-ablation lesion. FPC, flexible-printed-circuit; RF, radio frequency.

An initial recovery surgery was performed on days 7–8 for the control and ablation group, respectively. A fentanyl transdermal patch (50 µg/h, Novartis, NZ) was attached to the chest of the animal 24 h before the surgery to control pain. The pigs were fasted overnight before surgery. General anesthesia was induced with Zoletil (0.1 mL/kg, Virbac, NZ) and maintained with propofol (Diprivan 2%, 0.2–0.4 mg/kg/min, AstraZeneca, UK) and prophylactic antibiotics were given by injection (200 mg/mL, 5 mg/kg, Excede LA, Zoetis, NZ). Vital signs were continuously monitored and maintained within normal physiological ranges (14). A midline laparotomy (∼10 cm) was performed, and the gastric serosal surface was exposed with minimal gastric handling.

Baseline high-resolution electrical mapping was then performed, as described in High-Resolution Electrical Mapping. For animals in the experimental ablation group, RF-ablation was performed, as described in Gastric Ablation; for animals in the control group, a sham-ablation was performed instead, where the gastric tissue was pressed with the tip of forceps to mimic the ablation catheter tip, for the same duration and arrangement as the RF-ablation but with no energy delivered to the tissue. After ablation (RF or sham), a period of postablation high-resolution mapping was repeated over the ablation site, as described in High-Resolution Electrical Mapping.

At the end of the experiments, all recording equipment was removed from the abdomen and the laparotomy was closed via a mass closure with 0 PDS (Polydioxanone) sutures (Ethicon, Johnson & Johnson). Marcain (0.25% with epinephrine, AstraZeneca, 1–2 mg/kg) was injected into the rectus sheath and subcutaneously surrounding the incision, and the skin was then closed with 3/0 monocryl sutures (Ethicon, Johnson & Johnson). After the surgery, paracetamol (250 mg/5 mL, Aspen Pharmacare, NSW, Australia) and Metacam (0.5-1 mg/kg/day, Boehringer Ingelheim) were given for 3 days to control inflammation and pain, and omeprazole (20 mg/day, Teva Pharma) was given for 7 days to minimize the potential of stomach ulcers. Animals had immediate access to water at the time of recovery and were fed a meal of standard solid pellets 2 h after the initial surgery. They were fed their normal diet (water, solid food, fruit/vegetable treats) for the rest of the recovery duration. If diarrhea was observed, animals were treated with antidiarrhea powder for 2–3 days (225 g, Phoenix Pharm, NZ), which was mixed with the standard pellet meal.

Two weeks after the initial surgery, a terminal surgery was performed with the same surgical methods as the initial surgery. Because no recovery data for gastric ablation currently exist, the 2 wk follow-up duration was informed by previous surgical studies of gastric transection and anastomosis that reported restoration of slow-wave conduction across that scar within 2 wk for all subjects (4, 19, 20). After the laparotomy, high-resolution electrical mapping was performed at multiple locations around the stomach, including across the ablation lesion, as described in High-Resolution Electrical Mapping. The animals were then euthanized with a lethal bolus injection of pentobarbital sodium while still under anesthesia.

High-Resolution Electrical Mapping

Intraoperative high-resolution electrical mapping was performed using validated flexible-printed-circuit (FPC) electrode arrays (256 electrodes, 16 × 16 array, 4 mm spacing) (FlexiMap, Auckland, NZ) (21, 22). The FPC array was gently positioned over the serosal region of interest and overlain with warm (37°C) saline-soaked gauze to maintain moisture and gentle pressure of the electrodes onto the serosa. The wound edges were approximated with surgical clamps. In the initial surgery immediately after laparotomy, slow-wave activity was initially recorded for at least 5 min to define the baseline electrical activity. The midpoint of the FPC array along the greater curvature was marked with a suture, and the FPC was then removed. Gastric ablation was performed across the middle of the previously mapped area, as marked by the suture (Fig. 1B). The specific ablation methods and parameters are detailed in Gastric Ablation.

After the ablation, the FPC electrode array was replaced in the same location of the stomach, now spanning the ablation lesion. High-resolution mapping was performed for ∼15–20 min, until dissociated propagation and/or frequency of slow waves was observed in the proximal versus distal regions across the ablation, signifying that a conduction block had been established (Fig. 1, C and D) (14). The electrodes were then removed, and the animal was recovered, as detailed in Animal Care and Surgery.

Two weeks later, intraoperative high-resolution mapping was repeated. The position of ablation was identified by the suture. The FPC electrode array was placed over the ablation region, as well as proximal and distal regions on the anterior and posterior surfaces of the stomach, to collate a comprehensive picture of the slow-wave activity across the stomach.

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). Ablations were performed across the midline of the previously mapped region and marked by the suture to ensure that the ablation was positioned within the region of the baseline electrical recording.

For the RF-ablation group, ablations were applied using the temperature-control setting without irrigation at 65°C and an ablation exposure time of 5 s per point, as previously described and validated to result in a conduction block in acute trials (14). For the control group, sham ablations were performed using the tip of a pair of forceps to mimic the tip of the ablation probe, with no power delivery at room temperature and equivalent exposure time of 5 s per point. For both the ablation and control groups, each ablation sequence consisted of 43 ± 4 adjacent points, forming a continuous line perpendicular to the greater curvature around the circumference of the stomach (i.e., across the anterior and posterior surfaces) (Fig. 1B). This linear circumferential “ring” geometry of ablation was designed to enable clear recognition of whether a conduction block was present at the site of ablation (14, 23) but was not designed to have a specific functional or symptomatic impact on the pigs, beneficial or detrimental.

Electrophysiological Signal Acquisition and Analysis

Signals were acquired at 512 Hz using an ActiveTwo system (BioSemi, The Netherlands) modified for passive use. Signal processing was performed in the validated Gastrointestinal Electrical Mapping Suite (GEMS; FlexiMap, Auckland, NZ) (24). Data were first downsampled to 30 Hz and then filtered using a Gaussian moving median to remove baseline drift and a Savitzky-Golay filter (“low-pass” with an effective cut-off of ∼2 Hz) to remove high-frequency noise (25). Slow-wave activation times (ATs) were marked and clustered using validated algorithms (24) followed by comprehensive manual review to ensure accuracy. Slow-wave propagation was visualized using isochronal AT mapping to visualize the area of propagation per unit of time (24). Slow-wave amplitude, velocity, and frequency were calculated and mapped using validated algorithms (24).

Gross Image Evaluation and Tissue Preparation

At the final surgery, immediately before euthanasia, the tissue across the circumference of the stomach that included the lesion (marked with the suture, Fig. 1E) was excised for gross and histological evaluation, ∼3 cm in width (Fig. 1, F and G). The tissue was cross-sectioned parallel to the greater curvature (i.e., transverse to the linear ablation lesion), and samples from different positions across the stomach were prepared (i.e., greater curvature, body, lesser curvature). The tissue orientation was consistent for all cases. Excised tissue samples were fixed in 10% neutral-buffered formalin and embedded in paraffin blocks.

Histological Analysis

Tissue sections were cut from paraffin blocks (5 μm thick) and stained with hematoxylin and eosin (H&E) and Masson’s trichrome stain (MTS) to evaluate the degree of damage to the muscularis propria and mucosal layers, using standard protocols for each stain (26, 27). Samples were assessed qualitatively based on the rate of inflammation, collagen deposition, angiogenesis, and edema (plenty: 5–4, moderate: 3–2, a few: 1).

Immunohistochemistry

ICCs were stained using a c-kit marque primary rabbit monoclonal antibody (c-kit-YR145; Cell Marque, CA), followed by secondary antibody (Novolink polymer; Novocastra, MA), using a validated protocol (14, 15). The immunohistochemistry protocol was standardized in the histology laboratory for specificity and sensitivity. In the known positive control using pig stomach, the antibody was serially tested and an optimal dilution of 1:30 was established. The positive control antigen expression was specific to the ICC cells and negative in the smooth muscle. Slides were imaged using the Metasystems VSlide scanner (Metasystems, Germany) at the Biomedical Imaging Research Unit (BIRU), University of Auckland.

ICCs (c-kit+ cells) were manually segmented in subsections of high-powered fields (×200 magnification) using Photoshop CS (Adobe, CA). The area of ICC was then calculated for each high-powered field and compared between the muscle layers within the lesion versus outside of the lesion, with each value obtained as a mean of three different sites.

Statistical Analysis

Quantitative data are presented as means ± standard deviation. Statistical differences in slow-wave characteristics before and after ablation, and ICC area in different regions of tissue, were compared using paired Student’s t test, with a significance threshold of P < 0.05.

RESULTS

Animal Care

This study comprised a total of six animals, with an average weight of 28 ± 0.7 kg on the arrival day, 33.3 ± 2.0 kg on the day of initial surgery, and 44.0 ± 1.4 kg at the time of the final surgery (i.e., 2 wk after the initial recovery surgery). Weight gained by the ablation group and control groups were similar between the initial and final surgery (gain of 11.2 ± 1.7 kg vs. 10.2 ± 1.1 kg, P = 0.065; Supplemental Table S1; see https://doi.org/10.6084/m9.figshare.19689124).

There were no differences observed in the postsurgical recovery of the ablation versus control groups across the 2-wk postsurgical duration, including body temperature, food consumption, animal behavior, and animal demeanor. After the initial surgery, all animals woke, became alert, responded to the study personnel, stood up on their feet, and drank water within 1–2 h after anesthesia.

Bowel function recovery was variable. Four animals had soft stools in the days following surgery (2/3 in each group). In two cases of the ablation group, the condition extended to diarrhea and was treated with antidiarrhea powder, which returned the stool consistency to normal within 3 days.

Postoperative surgical adhesions were identified at the 2-wk terminal surgery in 2 of 6 pigs. In one case of the ablation group, the inferior surface of the liver had adhered to the anterior surface of the stomach, and in one case of the control group, the liver had adhered to the abdominal wall. In these two cases, the animal did not show any complications regarding their growth or overall health.

High-Resolution Mapping Recordings

The dataset comprised a total of 610 min of high-resolution electrical mapping, including baseline, acute postablation, and 2 wk postablation (Fig. 1).

Baseline

For all cases, normal antegrade activation was exclusively observed during baseline recordings (9.5 ± 2.0 min), and the mean frequency, amplitude, and velocity in the mapped area were 4.3 ± 0.2 cpm, 2.1 ± 0.7 mV, and 6.4 ± 0.8 mm/s, respectively (Figs. 2i and 3i).

Figure 2.

Example high-resolution mapping results of RF ablation at baseline, acute postablation, and 2 wk postablation (n = 1 pig). A: position of FPC electrode array on the stomach. B: isochronal AT map of slow-wave propagation. Each color band indicates the area of slow-wave propagation per 1 s from red (early) to blue (late). Black arrows indicate the direction of slow-wave propagation and the double black lines at the end of an arrow represent termination of a slow wave (i.e., at a conduction block). C: electrograms from the 9 electrode positions labeled in B, with slow-wave ATs marked as red dots. i: baseline mapping showed normal antegrade slow-wave activity. ii: acute postablation mapping showed that a conduction block (black bar) was established at the location of the RF-ablation lesion. Electrograms from the electrode positions labeled in B showed antegrade activation proximal and distal to the ablation site and no discernible ATs in the electrodes at the RF-ablation site. iii: 2 wk postablation showed a conduction block at the location of the RF-ablation lesion persisted after the 2 wk of healing. A secondary pacemaker activated the distal region to the RF-ablation lesion with rapid, circumferential conduction. Electrograms from the electrode positions labeled in B, showing antegrade activation proximal and distal to the ablation site, and low-amplitude signals in the electrodes at the RF-ablation site. The frequency of slow waves in the proximal region to the RF-ablation sequence was higher than in the distal region. AT, activation time; FPC, flexible-printed-circuit; RF, radio frequency.

Figure 3.

Example high-resolution mapping results of sham ablation at baseline, acute postablation, and 2 wk postsurgery (n = 1 pig). A: position of FPC electrode array on the stomach. B: isochronal AT map of slow-wave propagation as explained in Fig. 2. C: electrograms from the 9 electrode positions labeled in B, with slow-wave ATs marked as red dots. i: baseline mapping showed normal antegrade slow-wave activity. ii: acute postablation mapping showed normal antegrade slow-wave propagation in the stable phase with no sign of a conduction block at the sham-ablation region, as expected. Electrograms from the electrode positions labeled in B, showing antegrade activation proximal and distal to the ablation site, and no trace of a conduction block. iii: 2 wk postablation showed normal antegrade slow-wave propagation in the stable phase, similar to the acute postablation results. AT, activation time; FPC, flexible-printed-circuit.

Acute Postablation

Slow-wave recordings were recommenced ∼5 min after ablation (allowing for removal of the catheter and replacement of the mapping electrodes), and slow-wave activity was recorded for 12.3 ± 3.7 min.

RF-ablation group.

The formation of a conduction block at the ablation region was confirmed after 5.0 ± 0.3 min, verified by slow-wave amplitude at the ablation sequence decreasing to being too small to detect (Fig. 2ii; Supplemental Fig. S1ii; see https://doi.org/10.6084/m9.figshare.19689127) (14). Slow-wave activity distal to the ablation sequence showed higher amplitude and velocity with propagation in the circumferential direction compared with the rest of the mapped area that had velocity primarily in the longitudinal direction (4.8 ± 0.7 vs. 1.7 ± 0.1 mV, P = 0.029; 11.5 ± 2.0 vs. 6.9 ± 1.0 mm/s, P = 0.035). This rapid, high-amplitude circumferential activation was another marker of effective gastric conduction block (Supplemental Fig. S1, I and J) (14, 28). Slow-wave activity in the regions proximal and distal to the ablation showed dissociated slow-wave activation and/or frequencies across the ablation line (Figs. 1, C and D, and 2ii), although the quantitative frequency difference was not statistically significant, highlighting the need for high-resolution mapping to assess spatial profiles (3.8 ± 0.4 vs. 3.5 ± 0.0 cpm, P = 0.35).

Control group (sham-ablation).

For the initial 2.0 ± 0.2 min of the postsham mapping period, slow-wave activity showed patterns of dysrhythmias, likely due to the gastric handling from the sham procedure, consistent with previous studies (29, 30). Afterward, slow-wave activity in the mapped area showed stable activity with continuous slow-wave propagation through the sham-ablation line (Fig. 3ii). There was no observation of an amplitude decrease at the sham-ablation line or rapid high-amplitude activity in the surrounding area. Frequency, amplitude, and velocity in the proximal versus distal regions of the sham ablation showed no difference (4.2 ± 0.5 vs. 4.2 ± 0.6 cpm, P = 1.000; 1.9 ± 0.5 vs. 2.0 ± 0.4 mV, P = 0.287; 7.2 ± 1.3 vs. 6.3 ± 0.7 mm/s, P = 0.486), further demonstrating lack of block or defect at the site of sham ablation, as expected.

Two Weeks Postablation

At the 2-wk recovery surgery, slow-wave activity was recorded for a total duration of 79.8 ± 13.0 min from various positions on the stomach to obtain a comprehensive evaluation of the electrical activation and propagation after ablation and gastric healing.

RF-ablation group.

Data showed that conduction blocks remained consistent and effective at the site of ablation (Figs. 2iii and 4H, and Supplemental Fig. S2; see https://doi.org/10.6084/m9.figshare.19689130). In all three ablation cases, the conduction block at 2 wk postprocedure was verified by established metrics (8, 14, 23), including undetectable activity at the ablation line, inhibited propagation at the ablation line, and rapid high-amplitude circumferential activity distal to the ablation (5.6 ± 1.6 vs. 2.0 ± 1.2 mV, P = 0.006; 13.5 ± 0.24 vs. 7.8 ± 0.9 mm/s, P = 0.007). Slow-wave activity occurred at significantly lower frequencies distal to the ablation compared with proximal activity (3.8 ± 0.7 vs. 4.4 ± 0.3 cpm, P = 0.039).

Figure 4.

Example high-resolution mapping results of RF-ablation at 2 wk postablation with an incomplete conduction block on the greater curvature of the stomach, and liver adhesion to the anterior surface of the stomach (n = 1 pig). High-resolution mapping over the greater curvature of the stomach (i: A–E), and posterior surface (ii: F–J). A and F: position of the FPC electrode array on the stomach. B and G: electrograms from the 8-electrode positions labeled in C and H (respectively), with slow-wave ATs marked as red dots. C and H: isochronal AT map of slow-wave propagation as explained in Fig. 2. D and I: velocity map of the same slow wave, showing the speed (color spectrum) and direction (arrows) of the wave at each electrode position on the array. E and J: amplitude map of the same wave. A complete conduction block was present at the ablation lesion on the posterior surface of the stomach (black bar), which inhibited normal antegrade propagation. However, on the greater curvature, an incomplete conduction block (translucent bar) caused a “leak” of slow-wave activation through the ablation lesion, with slow velocity propagation through the lesion (dashed lines) and blocked activation for one in every five cycles (as shown in B). This activity activated the distal region to the RF-ablation lesion with rapid, high amplitude, circumferential conduction. AT, activation time; FPC, flexible-printed-circuit; RF, radio frequency.

Across all three cases, in the proximal region, slow waves were initiated from the greater curvature in the upper corpus, which is the normal gastric pacemaker site (3, 31). The slow-wave propagation pattern in the region distal to the ablation varied across each individual case (Figs. 2iii and 4, C and H; Supplemental Figs. S2, C and H, and S3, iii and iv; see https://doi.org/10.6084/m9.figshare.19689133), including a rotating wavelet around the lesser-curvature edge of the ablation lesion and/or a secondary pacemaker on the greater curvature distal to the ablation.

For two of three cases, a complete conduction block was sustained across the ablation line, verified with established conduction block markers (Fig. 2iii; Supplemental Fig. S2). In one case, a partial incomplete block was detected with a small localized gap in the block at the greater curvature (i.e., in the tissue located under the gastro-omental artery; Fig. 4i). Recordings of the posterior surface of the stomach showed a complete conduction block across the corpus in that region (Fig. 4ii); anterior recordings could not be achieved in this case due to adhesions between the liver and the anterior surface of the stomach. Figure 5 shows the variation of slow-wave frequency in the proximal and distal regions (recorded simultaneously at 2 wk postablation) for each case of ablation over time. In the two cases with a complete block across the stomach (Fig. 5, i and ii), proximal and distal regions showed different and/or uncoupled frequencies.

Figure 5.

Comparison of frequency at 2 wk postablation in the proximal versus distal regions to the ablation sequence (RF or sham). i–iii: frequency analysis example of each of the RF-ablation cases (n = 1 pig each). For the first two cases (i and ii), the frequency was uncoupled between the proximal and distal regions, confirming the persistence of the conduction block at the ablation lesion after 2 wk of recovery. The third RF-ablation case (iii) shows the frequency analysis of the case with an incomplete conduction block over the greater curvature. For the initial ∼200 s, the frequency in proximal and distal regions to the ablation sequence was coupled, suggesting that there was no conduction block present. After the initial ∼200 s, the frequency in the distal region became sporadically uncoupled from the frequency in the proximal region, demonstrating an incomplete and inconsistent block at the ablation lesion. iv: example of frequency analysis of a sham-ablation case (n = 1 pig). The initial 400 s of the recording showed frequency variability in the proximal versus distal regions due to unstable slow-wave activity, presumably caused by the gastric handling during the procedure. Once the slow-wave activity became stable (at ∼400 s), the frequency in the proximal and distal regions was coupled. RF, radio frequency.

The one case of partial, localized incomplete conduction block presented a more complex case of variable frequency coupling and uncoupling. The antegrade-propagating wavefront in the proximal stomach typically activated the distal stomach through the small gap in the conduction block, although was occasionally slowed through that gap, or blocked (Fig. 4i). The first 5 min of recording showed similar frequencies in the proximal and distal regions (Fig. 5iii), with the frequency coupling caused by “escape” of the proximal activation through the gap in the conduction block (Fig. 4i). After the initial 5 min, the frequency in the distal region became sporadically uncoupled from the frequency in the proximal region (Fig. 5iii), demonstrating an incomplete and inconsistent block at the ablation lesion and correlating with the high-resolution mapping data showing the sporadic “escape activity” through the localized gap in the conduction block. Analysis of the time interval of two adjacent electrodes on opposite sides of the ablation lesion and located along the greater curvature (electrodes 4 and 5 in Fig. 4B), showed a gradual increase of time delay over the lesion (Figs. 4B and 5iii). When the time interval reached ∼5 s, the next slow wave failed to propagate through the lesion for a single wave cycle, representing a conduction block for that single wave.

Control group (sham-ablation).

For all 3 cases, slow-wave activity showed normal antegrade propagation with no amplitude drop at the location of the sham ablation. Slow-wave activity in the proximal versus distal regions to the sham ablation showed no significant difference (frequency: 4.2 ± 0.3 vs. 4.2 ± 0.3 cpm, P = 1; amplitude: 1.8 ± 0.3 vs. 1.9 ± 0.4 mV, P = 0.603; velocity: 8.2 ± 0.4 vs. 8.4 ± 0.7 mm/s, P = 0.855) (Fig. 3iii and 5iv; Supplemental Fig. S4; see https://doi.org/10.6084/m9.figshare.19689136). In addition, a comparison of the 2-wk recovery data versus baseline data showed no significant difference in frequency (P = 0.560), amplitude (P = 0.573), or velocity (P = 0.065). However, throughout the experiments, we observed that handling the stomach close to the suture area caused dysrhythmic activity for a few minutes, before the stomach returned to stable normal electrical activity, again highlighting the sensitivity of the stomach to handling (Fig. 5iv) (29, 30, 32).

Histology and Gross Image Evaluation

There was no sign of perforation in any of the ablation cases. In the initial surgery, the ablation lesion was visible by a change of color and texture of the tissue (Fig. 1E). In the 2-wk recovery surgery, the ablation lesion was barely visible (Fig. 1F). The lesion was detected by the position of the suture placed in the initial surgery to mark the position of the ablation. Qualitatively, the tissue of the ablation lesion was stiff and less flexible than the normal tissue. On the cross-sectional view of the frozen section, the lesion was readily differentiated by the pink-red color of the lesion compared with the pale color of the normal tissue due to angiogenesis in the lesion area (Fig. 1G).

Figure 6, A–D, shows the qualitative evaluation of H&E and MTS tissue sections from various positions of the lesion on the stomach (58 total, 29 of each stain; samples from greater curvature, body, and lesser curvature). On average, the histological evaluation showed a medium-to-high level of inflammation within the ablation lesion (Figs. 6, A–D, and 7i). The number of collagen bundles was low-to-medium within the lesion. The collagen bundles within the lesion were stained light blue (Fig. 6, B and D), hence the collagen was newly formed and the lesions were in the initial stages of proliferation and remodeling (33). The rate of angiogenesis was medium to high, along with a medium level of edema.

Figure 6.

H&E and MTS-stained sections of gastric tissue, 2 wk postablation (RF and sham). A and B: example of RF-ablation sample with typical damage and lesion depth showing a medium level of inflammation and collagen deposition. C and D: example of RF-ablation tissue sample with a high level of damage and lesion depth showing a high level of inflammation, low rate of collagen deposition, and a high rate of angiogenesis. In this rare example, although the ablation caused a transmural lesion, it did not cause perforation. E and F: example of sham-ablation sample with properties of healthy tissue. H&E, hematoxylin and eosin; MTS, Masson’s trichrome stain; RF, radio frequency.

Figure 7.

Histological and immunohistochemical evaluation of RF-ablation lesion after 2 wk of recovery. A: qualitative evaluation of RF-ablation lesion after 2 wk of recovery for inflammation, collagen deposition, angiogenesis, and edema (plenty: 5–4, moderate: 3–2, a few: 1), and the transmural depth of the lesion relative to the various tissue layers (n = 24 samples from 3 pigs). B: quantitative measurement of ICC area per field (×200) in RF-ablation lesion, healthy tissue, and the lesion border in longitudinal (ICC-LM) and circular (ICC-CM) muscle layers (n = 19 samples from 3 pigs). ICC, interstitial cells of Cajal; RF, radio frequency.

The depth of the lesions was mainly full muscle thickness, although a high rate of variation was observed in some tissue samples from partial damage to the muscle layer (intact circular muscle bundles) to full-thickness lesion (granulation tissue extending into mucosa) (Fig. 7).

Tissue samples from the sham-ablation group had normal tissue properties (Fig. 6, E and F), and as expected, no significant difference was observed in the tissue samples from the sham-ablation site. Histological evaluation of tissue adjacent to the lesion in the RF-ablation cases showed no difference when compared with the control tissue from the sham-ablations, confirming that damage caused by RF-ablation was localized to the site of ablation and did not affect the wider surrounding gastric tissue.

Immunohistochemistry stain-quantification of ICC as area per field.

The area of c-kit+ cells at high-power magnification (×200) within the RF-ablation lesions, indicating ICC, was 117 ± 123 µm² in the longitudinal muscle, and 56 ± 65 µm² in the circular muscle layer (Figs. 8, A–C, and 7B). The area of ICC within the RF-ablation lesions showed a significant decrease compared with the area of ICC per field in healthy, nonablated longitudinal (563 ± 275 µm², P < 0.001) and circular muscle layers (198 ± 73 µm², P < 0.001) (Fig. 8, C and D). ICC within the RF-ablation lesion had lower intensity of c-kit+ stain (i.e., fainter brown appearance) and deformed shape (Fig. 8, A and B). In the circular muscle layer, c-kit+ cells had significantly higher density in the healthy tissue along the border with the ablation lesions, compared with the normal healthy tissue away from the lesion (429 ± 196 vs. 198 ± 73 µm², P < 0.001) (Figs. 8E and 7B). In the longitudinal muscle layer, c-kit+ cells had higher density along the lesion border compared with the normal healthy tissue away from the lesion, but it was not statistically significant (628 ± 231 vs. 563 ± 275 µm², P = 0.169). These areas of high-density ICC were variably distributed throughout the lesion border (Figs. 8E and 7B). In cases of sham ablation, ICC showed consistent population and morphology (Fig. 8, F and G).

Figure 8.

Immunohistochemistry (c-Kit) stain of ICC in gastric tissue, 2 wk postablation (RF and sham). i: A–E: ICC within the lesion have lower density, with a lower intensity of brown (c-kit positive), whereas the ICC near the border of lesion have higher density. Properties of ICC in RF-ablation lesion (×200 magnification) showing: ICC cellularity in the longitudinal muscle layer on the right side and circular muscle layer on the left side of the image (A) and ICC cellularity in circular muscle layer (B). C: cellularity of ICC across the RF-ablation lesion, the lesion border, and healthy tissue in the muscle layer (×20 magnification). ICC cellularity in healthy tissue section (D) and ICC across the lesion border (E) (longitudinal muscle layer on the right side and circular muscle layer left side of the image; ×50 magnification). ii: F and G: properties of ICC in sham-ablation tissue sections were similar to healthy tissue (×200 magnification). ICC cellularity in sham-ablation tissue sections (F) (longitudinal muscle layer on the right side and circular muscle layer left side of the image) and in circular muscle layer (G). ICC, interstitial cells of Cajal; RF, radio frequency.

DISCUSSION

This study aimed to investigate the sustained efficacy of gastric ablation on modulating slow-wave conduction in an animal recovery model and to provide initial safety data as a critical advance toward clinical translation. Importantly, ablation did not induce gastric perforation, and none of the animals showed any detrimental health impacts. Mapping demonstrated that conduction blocks induced by ablation persisted after the 2-wk healing period, verified by standard conduction block markers (14, 23). Histological data showed disrupted ICC connectivity in early scar and lack of ICC regeneration into the lesion at the current stage of healing. Altogether, these findings provide the first evidence that gastric ablation is feasible and safe for modulating gastric slow-wave activity through the postprocedural healing phase.

The finding that the conduction block caused by RF ablation persists after 2 wk of healing is novel. Classic studies previously demonstrated that gastric transection and immediate reanastomosis had only a temporary effect on slow-wave propagation, with entrainment of slow-wave frequency across the suture line recovering fully within 2 wk (4, 19). The current study duration of 14 days was therefore selected as an appropriate timeframe to confirm persistent conduction block from ablation, ruling out rapid restoration of conduction across an induced lesion. It remains to be confirmed if ablation achieves a long-term physiological conduction block, but this is expected given that ablation achieves a complete tissue injury (14).

In all cases of ablation, the frequency of slow waves proximal to the circumferential ablation was higher than distal regions, revealing the underlying proximal-to-distal intrinsic frequency of the ICC network in the stomach (4, 31). Recordings at 2 wk postablation showed that slow waves were initiated from the normal gastric pacemaker site in the high corpus (3, 31), at baseline frequency, whereas distal regions showed a variety of activity patterns. When the block was complete, slow waves were initiated from either a secondary pacemaker or a rotating wavelet that propagated around the lesser-curvature side of the lesion. Both patterns activated regions distal to the block with a rapid high-amplitude circumferential activity, which is a marker of effective conduction block (28). These results agree with the mapping results from our previous acute gastric ablation study (14).

In the case with the incomplete block on the greater curvature, the incomplete block showed reduced slow-wave conduction through the ablation line, with a lengthening conduction interval and block of every fourth to fifth wave (Fig. 4). Histological evaluation confirmed that the lesion in this site only covered partial muscle depth, explaining the incomplete block. This incomplete lesion was likely due to a small gap in ablation continuity rather than inadequacy of the ablation settings (65°C for 5 s per point), as these settings have previously been validated to routinely result in full-thickness lesions inducing a complete block (14). This finding emphasizes the need for achieving optimized and standardized methods for gastric ablation, including optimal settings, spacing between ablation points and ablation depths, while also remaining flexible enough to create different ablation geometries.

None of the ablation cases resulted in gastric perforation, nor other serious adverse events compared with the sham group. However, in a few samples, RF-ablation lesions were extended to the mucosal layer of the stomach, reiterating the importance of using ablation parameters that deliver the minimal energy required to achieve functional efficacy in creating a conduction block (14). Perforation is the most serious theoretical complication of gastric ablation, as it would result in life-threatening sepsis requiring emergency surgery. Based on previous studies of thermal GI wall injuries during laparoscopic or endoscopic procedures, perforation would be most likely to arise within a 3- to 5-day postprocedural window (16, 17), well within our 2-wk study duration. It is worth mentioning that 2 h after the initial surgery, the animals were fed solid food, and digestion of solid food did not cause complications. Histology results showed that ablated tissue was still in the inflammatory phase (granulation tissue), with early collagen formation indicating an emerging scar. Although most of the lesions studied had a depth completely through the tunica muscularis and/or into the submucosa, histological analysis showed substantial depth variation, likely reflecting the nonirrigated temperature-control mode of a commercial RF-ablation system. Future studies could evaluate alternative ablation control methods, such as power-controlled RF-ablation with irrigation to control the impedance rise (34), or electroporation (35, 36), to achieve greater consistency and to avoid full thickness issues.

In one case of sham ablation, the liver adhered to the abdominal wall, and in one case of RF-ablation, the liver adhered to a small region on the stomach. Adhesions are a common complication in open surgery but can impact quality of life (37, 38). A single adhesion occurred in both the RF- and sham-ablation groups in our study, thus no conclusion can be made whether RF ablation had an impact on the risk of adhesion. In addition, extensive research has been done on prevention of tissue adhesion (39, 40). Currently, pharmacological treatments and physical barriers are used to lower the risk of postsurgical adhesions, and these methods may be useful if ablation is found to increase the risk of adhesions in the future (41).

Immunohistochemistry evaluation showed that ICC (c-Kit+ cells) in the ablation lesion were scarce, with no sign of acute repair or regrowth of ICC in the lesion and with both longitudinal and circular muscle layers affected. In some regions of healthy tissue that bordered the ablation lesion, ICC numbers and/or concentration were significantly higher than the normal healthy tissue, which may be an acute response to inflammation (2). This effect was patchy and inconsistent, and variable orientation of ICC within the sectioned tissue fields may also be partly responsible (42, 43). The ablation lesions were still in the healing phase at the 2-wk time point, and future studies with longer postsurgical recovery periods would therefore be beneficial to determine the longer-term impact of healing on lesion structure.

This study was limited to a small feasibility cohort to provide the first evidence for safety and efficacy. Ablation functions by creating localized conduction blocks that can be targeted to eliminate sites of abnormal electrical activation (15, 44), hence, why the formation and maintenance of a conduction block was a critical focus of this initial recovery study. The circumferential ring geometry of the ablation was designed to provide a clear decoupling of the proximal and distal stomach to easily identify the presence of a conduction block and was not designed to have therapeutic impact. However, we have previously shown that ablation can eliminate abnormal ectopic pacemakers in the acute period (15), and follow-up recovery studies may now be designed to evaluate ablation geometries with greater translational value. To move toward clinical translation, several technical advances are needed, including longer-term studies and less invasive approaches (e.g., via laparoscopic or endoscopic instruments) (29, 45). It would also be beneficial to monitor the electrophysiology of the stomach throughout the study, using implanted electrodes or emerging noninvasive body-surface modalities (46). Future studies should include functional measurements of gastric motility in the postablation period, for example, gastric emptying and/or manometry measurements, to elucidate the functional impacts of this intervention.

Although this study indicates that sustained gastric slow-wave ablation is feasible, clear therapeutic targets must be defined before clinical translation can begin. It is now possible to investigate avenues of increased clinical relevance by targeting ablation at correcting dysrhythmias. One promising target is sustained retrograde propagation, which is an emerging mechanism of dysmotility (3, 9). Retrograde propagation mostly originates from entrainment by an ectopic pacemaking region distal to the natural pacemaker, and our separate recent preclinical study has shown that ectopic gastric sources can be effectively and reliably identified in real-time and destroyed by ablation, as is achieved in therapeutic cardiology (15). Foci and pathways responsible for retrograde gastric propagation have recently been mapped in patients with severe postoperative dysmotility, including ectopic pacemaking after sleeve gastrectomy, and a retrograde “gastric aberrant pathway” arising after distal gastrectomy, providing examples where ablation might offer an appealing alternative to major revisional surgery (3, 47). Gastric dysrhythmias have also been associated with functional dyspepsia, reflux, chronic nausea, vomiting syndromes, and gastroparesis (3, 7–9, 48). It remains unclear if eliminating specific conduction pathways could yet be impactful in these conditions, but gastric ablation now provides a technique to enable such investigations.

Conclusions

This study demonstrates the sustained ability of gastric ablation to block slow-wave propagation after 2 wk of healing without perforation, demonstrating the safety and efficacy of ablation as a potential long-term intervention for gastric dysrhythmia.

SUPPLEMENTAL DATA

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.19689127.

Supplemental Fig. S2: https://doi.org/10.6084/m9.figshare.19689130.

Supplemental Fig. S3: https://doi.org/10.6084/m9.figshare.19689133.

Supplemental Fig. S4: https://doi.org/10.6084/m9.figshare.19689136.

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.19689124.

GRANTS

These studies and/or authors were supported by the New Zealand Health Research Council (to T. R. Angeli-Gordon), Royal Society Te Apārangi (to T. R. Angeli-Gordon), and the National Institutes of Health DK057061 (to G. Farrugia), DK052766 (to G. Farrugia), DK123549 (to A. Beyder), and AT010875 (to A. Beyder).

DISCLOSURES

No commercial financial support was received for any material presented in this paper. G. O’Grady and L. K. Cheng hold intellectual property on gastrointestinal electrophysiology. L. K. Cheng, N. Paskaranandavadivel, and T. R. Angeli-Gordon are shareholders in FlexiMap Ltd. G. O’Grady 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., G.O., L.A.N., A.B., G.F., S.J.A., N.P., L.K.C., and T.R.A-G. conceived and designed research; Z.A., G.O., L.A.N., A.E.M., N.P., and T.R.A-G. performed experiments; Z.A., C-H.A.C., A.M., S.A., G.B.S., and N.P. analyzed data; Z.A., G.O., L.A.N., A.E.M., C-H.A.C., A.M., S.A., A.B., G.F., S.J.A., G.B.S., N.P., L.K.C., and T.R.A-G. interpreted results of experiments; Z.A. prepared figures; Z.A. and T.R.A-G. drafted manuscript; Z.A., G.O., L.A.N., A.E.M., C-H.A.C., A.M., S.A., A.B., G.F., S.J.A., G.B.S., N.P., L.K.C., and T.R.A-G. edited and revised manuscript; Z.A., G.O., L.A.N., A.E.M., C-H.A.C., A.M., S.A., A.B., G.F., S.J.A., G.B.S., N.P., L.K.C., and T.R.A-G. approved final version of manuscript.