Endoscopic muscle biopsy sampling of the duodenum and rectum: a pilot survival study in a porcine model to detect myenteric neurons.

BACKGROUND AND AIMS

Small bowel and colorectal muscle biopsy sampling requires a surgical approach. Advancing our understanding of the pathophysiology of motility disorders, such as functional bowel disorders, intestinal pseudo-obstruction, and slow-transit constipation, is hindered by our inability to noninvasively obtain muscularis propria (MP) for evaluation of multiple cell types, including myenteric neurons. The aims of this study were to determine (1) technical feasibility, reproducibility, and safety of performing duodenal endoscopic muscle biopsy sampling (dEMB) and rectal endoscopic muscle biopsy sampling (rEMB) using a clip-assist technique and (2) the presence of myenteric neurons in tissue samples.

METHODS

Five 40-kg pigs were studied. Each animal underwent a dEMB and rEMB procedure. dEMB was performed using a single resection clip-assist technique. An over-the-scope clip was advanced to the duodenum. Tissue was suctioned into the cap and the clip deployed. The pseudopolyp of the duodenal wall created was then resected using snare electrocautery. rEMB was performed using a double resection clip-assist technique. EMR was initially performed to uncover the underlying MP using a band ligation technique. An over-the-scope clip was then advanced to the exposed MP. The MP was retracted and suctioned into the cap and the clip deployed. The pseudopolyp of the MP was resected using snare electrocautery. An antibody to protein gene product 9.5 was used to determine the presence of myenteric neurons in the samples. Animals were kept alive for 2 weeks, at which time an upper endoscopy and necropsy were performed.

RESULTS

dEMB and rEMB were successfully performed in all animals with no procedural adverse events using this "no hole" (close then cut) approach. Mean procedure times for dEMB and rEMB were 23.7 ± 2.5 minutes and 13.25 ± 2.8 minutes, respectively. Mean length of resected full-thickness duodenal wall was 13.25 ± 4.3 mm and rectal MP was 12.5 ± 1.7 mm. Hematoxylin and eosin stain and antibody to protein gene product 9.5 confirmed the presence of MP with inner circular, outer longitudinal, and intermuscular layers, including myenteric neurons, in all samples. Clinical course was uneventful in all animals. Repeat upper endoscopy at 2 weeks showed well-healed dEMB sites. Necropsy in all animals showed no perforation, fluid collection, or abscess at the dEMB and rEMB sites.

CONCLUSIONS

Based on this preclinical study, dEMB and rEMB appear to be technically feasible, reproducible, and safe. Sufficient MP tissue was obtained to identify myenteric neurons. These promising results are a step toward successful and safe implementation of these techniques into clinical practice for tissue diagnosis of muscle-based pathologies.