Innovative gastric endoscopic muscle biopsy to identify all cell types, including myenteric neurons and interstitial cells of Cajal in patients with idiopathic gastroparesis: a feasibility study

Background and Aims: The pathophysiology of some GI neuromuscular diseases remains largely unknown. This is in part due to the inability to obtain ample deep gastric wall biopsies that include the intermuscular layer of the muscularis propria (MP) to evaluate the enteric nervous system, interstitial cells of Cajal (ICCs), and related cells. We report on a novel technique for gastric endoscopic muscle biopsy (gEMB). Methods: Patients with idiopathic gastroparesis were prospectively enrolled in a feasibility study by using a novel no hole gEMB. Main outcome measures were technical success, adverse events, and histologic confirmation of the intermuscular layer, including myenteric neurons and ICC. The gEMB was a double resection clip-assist technique. A site was identified on the anterior wall of the gastric body as recommended by the International Working Group on histologic techniques. EMR was performed to unroof and expose the underlying MP. The exposed MP was then retracted into the cap of an over-the-scope clip. The clip was deployed, and the pseudopolyp of MP created was resected. This resulted in a no-hole gEMB. Results: Three patients with idiopathic gastroparesis underwent gEMB. Patients had severe delayed gastric emptying with a mean (+/- standard deviation [SD]) of 49 +/- 16.8% of retained gastric contents at 4 hours. They had no history of gastric or small-bowel surgery and did not use steroids or other immunosuppressive drugs. The gEMB procedure was successfully performed, with no procedural adverse events. Postprocedural abdominal pain was controlled with nonsteroidal anti-inflammatory agents and opioid analgesics. Mean length of resected MP was 10.3 +/- 1.5mm. Mean procedure time was 25.7 +/- 6 minutes. Hematoxylin and eosin (H&E) staining of tissue samples confirmed the presence of both inner circular and outer longitudinal muscle, as well as the intermuscular layer. H&E staining showed reduced myenteric ganglia in 1 patient. In 2 patients, specialized immunohistochemistry was performed, which showed a marked decrease in myenteric neurons as delineated by an antibody to protein gene product 9.5 and a severe decrease in ICC levels across the muscle layers. At 1 month follow-up, upper endoscopy showed a well-healed scar in 2 patients and minimal ulceration with a retained clip in 1 patient. CT of the abdomen confirmed the integrity of the gastric wall in all patients. Because of lack of an immune infiltrate in the resected samples, patients were not considered suitable for immunosuppressive or steroid therapy. Conclusions: gEMB is feasible and easy to perform, with acquisition of tissue close to surgical samples to identify myenteric ganglia, ICCs, and multiple cell types. The ability to perform gEMB represents a paradigm shift in endoscopic tissue diagnosis of gastric neuromuscular pathologies.