endoscopic mucosal resection
full-thickness resection device
interstitial cells of Cajal
Gastroparesis is a syndrome of delayed gastric emptying in the absence of mechanical obstruction. Common symptoms of the disease include early satiety, postprandial fullness, nausea, vomiting, bloating and upper abdominal pain . The most common etiologies for gastroparesis are idiopathic, diabetic and surgical, whereas less common causes include parkinsonism, amyloidosis, paraneoplastic and scleroderma . Gastroparesis is primarily a neuropathic condition, whether it is autonomic neuropathy (idiopathic or diabetic) or loss of extrinsic neural control (vagal nerve injury, parkinsonism, amyloidosis). Less commonly the underlying pathophysiology could be infiltration or atrophy of the muscle layer as seen in malignancies, scleroderma as well as amyloidosis . Gastroparesis is associated with decreased patients’ quality of life, increased health-care costs; including office visits, emergency room visits, hospitalizations; as well as negative health outcomes . The number of gastroparesis-related hospitalizations has been increasing in the United States. Moreover, economic impact of gastroparesis-related hospitalizations has been raising with estimated of more than 200 million dollars in 2004 .
The pathophysiology of gastroparesis is poorly understood. Data suggest that there are distinct histologic subsets of gastroparesis. The gastric nervous system cannot be evaluated with traditional endoscopic punch biopsy as it usually misses submucosal neuromuscular structures and does not capture muscularis propria (MP). Traditionally, diagnostic histopathological evaluation of gastrointestinal (GI) motility disorders was only possible with a surgical biopsy. However, a surgical full-thickness bowel biopsy is associated with some significant morbidity. In a study by Knowles et al, 124 patients with clinico-physiological diagnoses of chronic intestinal pseudo-obstruction underwent laparoscopy and full-thickness bowel biopsy. Surgery-related morbidity such as bleeding and obstructive symptoms occurred in 11% of patients. Conversion to an open operation occurred in 2% . Apart from short-term adverse events, any surgical procedure is a risk factor for long-term complications, such as intraabdominal adhesions with consequent pain and/or bowel obstruction. Although specific diagnostic yield was high, particularly for jejunal biopsies (89%), due to the invasive nature of the surgical full-thickness biopsy and its potential complications, it has not gained broad utilization .
Recent advances in endoscopic technology have now made acquisition of deep tissue or full-thickness biopsies possible. This chapter will review the current techniques for full-thickness or deep gastric biopsies, and discuss practicalities regarding whether this approach should be implemented in clinical practice.
Gastric nervous system
The enteric nervous system contains 200–600 million neurons that are distributed in small ganglia and are mainly found in two plexuses: the myenteric (or Auerbach’s) and submucosal (or Meissner’s) plexus. The myenteric plexus lies between the longitudinal and circular muscle and extends the entire length of the gut. The submucosal plexus is located between the circular muscle layer and the muscularis mucosa, and is best developed in the small intestine, but not in the stomach or esophagus .
One important participant of gastrointestinal autonomy are the nonneural cells called interstitial cells of Cajal (ICC), which serve as electrical pacemakers and mediators of enteric neurotransmission . A gastric pacemaker area is localized at the proximal part of the great curvature and generates a slow-wave basal rhythm with a frequency of three cycles per minute, which spreads distally and determines the frequency of gastric contractions . This area was not significantly affected by a proximal gastric vagotomy and when transected from the gastric corpus, a new pacemaker appeared in the distal stomach with a slower frequency .
Gastrointestinal autonomic rhythmicity is thought to be comprised from ICC as well as the myenteric plexus and submucosal plexus. These structures are widely distributed in the gastrointestinal tract. ICC generate spontaneous pacemaker currents that drive the mechanical activities of smooth muscle cells . These cells have been referred to as pacemakers in gastrointestinal muscles . Based on their morphological features and location, ICC are divided into intermuscular (between the circular and longitudinal muscle layers), intramuscular (within the muscle layers), deep muscular plexus (between inner and outer parts of the circular muscle), submucosal surface of the circular muscle and subserosal layer . Different types of ICC are thought to carry out different roles, however regardless of their function type, a decrease in ICC density has been found in patients with gastrointestinal motility disorders including gastroparesis . In addition, preservation of ICC has been associated with a better response to gastric electrical stimulation in patients with gastroparesis and their depletion has been seen in patients with greater symptomatology even after three months of gastric electrical stimulation .
Pathophysiology of gastroparesis
Pathophysiological changes in gastroparesis are a complex group of changes incited by alterations in various mechanisms. Gastric emptying is affected by less frequent contractions, antroduodenal incoordination and pyloric spasm . Along with discordance of function, disturbances of gastric compliance (increased or decreased ) have been reported as potential contributors to the symptomatology. Such disturbance of synchrony is attributed to loss of coordination between smooth muscle, autonomic nerves and ICC, all of which separately or in combination, are affected by an underlying condition. ICC have since gained broad recognition and subjects with diabetic and non-diabetic gastroparesis have been found to have a reduced number . In addition, several studies have demonstrated subcellular changes in animals with a decrease in nitric oxide-containing nerves , which was later seen in human studies along with a deficiency of ICC and substance P .
Initial studies of gastric tissue conducted on cadaveric stomachs using silver staining demonstrated normal histology with a normal myenteric plexus, abdominal vagus and branches in both affected subjects and the control group . Since then, more specific and sophisticated immunohistochemical staining techniques have been developed and implemented for the enteric nervous system.
Full-thickness gastric biopsies obtained by the Gastroparesis Clinical Research Consortium, revealed that both idiopathic and diabetic gastroparetic stomachs had cellular abnormalities in 83% of cases . The most common abnormality was a decrease in ICC. Residual ICC were often observed as not in contact with nerve endings and only rarely with smooth muscle cells, or with other ICC. The second most common abnormality was phenotypic changes in immune cells in the muscularis propria or an increase in the number of immune cells. Notably these changes were seen in both idiopathic and diabetic patients. Other findings in this study included decreased expression of neuronal marker PGP9.5 suggesting a decrease in nerve fibers. Nitric oxide synthase expression, however was found to be decreased more in idiopathic gastroparesis compared to diabetic gastroparesis (40% vs 20%, respectively).
There are three diagnostic tests objectively demonstrating delayed gastric emptying: scintigraphy, wireless motility capsule, and breath testing . The most commonly used test is solid-phase scintigraphy , in which a 99m Tc sulfur colloid-marked egg sandwich is ingested by the patient along with a piece of toast and jelly. Thereafter, the percent of food remaining in the stomach is reported at key time points: 1 h (37–90%), 2 h (30–60%), and 4 h (0–10%). The most reliable parameter is gastric retention at 4 hours, although measurement at 1 and 2 hours is also important, since prolongation of early phases may be associated with symptoms of gastroparesis . The above-mentioned diagnostic methods can help establish a diagnosis, but they provide no insight into the pathophysiology of the disease, the underlying cause (idiopathic, diabetic, or infiltrative process) or individual therapy.
In order to find answers to these questions, full-thickness gastric biopsies have been viewed as a promising modality. It provides information on ganglia abnormality, reduction in ICC count and inflammatory processes, all of which can be potential markers for grading the disease, predicting response to treatment and prognosis. Examination of deep gastric tissue may lead to a better understanding of pathophysiology of the disorder and bring the prospect of new therapeutic options.
Endoscopic deep tissue biopsy
Acquisition of deep gastrointestinal wall layers requires deep tissue or a full-thickness biopsy, i.e. full thickness implies creation of a wall defect, or, alternatively, perforation. More recently endoscopic methods of full-thickness tissue acquisition have been studied and implemented in clinical practice. Several endoscopic full-thickness biopsy techniques have been described such as endoscopic ultrasound-guided fine needle biopsy of the muscularis propria of the gastric wall, percutaneous endoscopically assisted full-thickness biopsy and endoscopic full thickness biopsy.
Endoscopic ultrasound-guided fine needle biopsy of the muscularis propria of the gastric wall by using a 19-gauge core needle in patients with gastroparesis has been reported by Othman et al. . Eleven patients underwent radial and linear endoscopic ultrasound (EUS) before undergoing gastric neurostimulator placement. Linear echoendoscope was used to obtain a core biopsy of the stomach wall in the antrum of the posterior wall with a 19-gauge Procore biopsy needle (Cook Endoscopy, Winston-Salem, NC). The needle was allowed to completely pass through the wall of the stomach beyond the serosa. Ten uniform back and forth movements of the needle were made for each of the passes with a 10-ml syringe suction applied. No more than 5 passes with the FNA needle were performed. The patients were taken to surgery within 24 hours for gastric neurostimulator placement and to obtain a surgical full-thickness stomach biopsy. The site of endoscopic puncture was identified by a previously placed tattoo. Endoscopic and surgical specimens were evaluated for ICC count using immunohistochemical staining with antibodies to CD117. Success rates of the EUS-guided core biopsy was 81% and 54.5% for ICC and myenteric plexus assessment, respectively. Surgical success was reported to be 90% for histologic assessment of both ICCs and myenteric plexus. No statistically significant difference was found in the success rate for obtaining sufficient tissue for histologic evaluation between the endoscopic and surgical approaches. The method’s safety profile was remarkable for absence of any after-procedure adverse events. The limitations of the study included a single-center design, small number of patients, fragmented endoscopically obtained tissue, lack of correlation between the number of myenteric plexus nerve bundles per high-power field (HPF) in endoscopic and surgical biopsies as well as nonblinded pathologist analysis. In addition, the rate of success in obtaining sufficient tissue for histologic assessment of ICCs (81%) and myenteric plexus (54.5%) is low compared to the more recently emerging methods of endoscopic full-thickness biopsy (EFTB).
Percutaneous endoscopically assisted full-thickness biopsy of the stomach has been reported in animal and human studies . Fraser et al. made a 3 mm incision through the abdominal skin, through which a spring-loaded, 14-gauge biopsy needle was inserted to obtain 4 adequate antral biopsy specimens from three dogs. The needle passages were performed under constant endoscopic surveillance. There were no significant adverse events, no morbidity or mortality in the post-procedure period. Presence of an intact gastric wall with the enteric nervous system, ICC and glia was confirmed. Limitations of this approach were the small sample size, small specimen size (2 mm by 4 mm) as well as fragmented specimens. The same group of authors performed a similar technique on ten patients with gastroparesis (diabetic or idiopathic) and severe dyspepsia. A 4 mm incision was made through the abdominal skin and a spring-loaded biopsy gun with a 14-gauge coaxial needle was inserted. After penetrating the stomach under direct visualization of endoscopy, an internal notched needle was extended. The same internal needle was then withdrawn allowing the notch to straddle the stomach wall. The needle assembly was subsequently removed and tissue was submitted for analysis. Up to four passes were performed on each patient. No routine closure was performed. There was one episode of minor bleeding successfully treated with an Endoclip. There was transient mild abdominal wall pain in most patients. Sufficient tissue for histological review was obtained in 9 of 10 patients. Mean surface area of the biopsy specimen was 8.5 mm 2 . The technique had an attractive safety profile with no significant morbidity or mortality. However, the number of cases was small as was the size of the obtained tissue sample. In addition, in one case of histologically confirmed inflammatory leiomyopathy, the patient was referred for laparoscopic full-thickness biopsy for confirmation prior to initiation of immunosuppression therapy. Hence the suggested technique did not prove to be adequate in substituting surgical intervention. Furthermore, this technique could also be challenging in obese patients due to increased amount of tissue that the biopsy needle has to penetrate before reaching the lumen of the stomach.
Endoscopic full-thickness biopsy
Endoscopic full-thickness biopsy (EFTB) methods have been divided into two different approaches: full thickness resection followed by closure of the wall defect (cut then close); and creating wall duplication (with serosa-serosa opposition) before resection (close then cut).
Cut-then-close techniques for endoscopic full thickness biopsy
Animal studies of cut-then-close techniques
Rajan et al. and Fritscher-Ravens et al. performed EFTB on pig models with tunneling technique and a novel endoscopic prototype device, respectively. In the first study, a submucosal tunnel was created via a tunneling balloon and subsequent deep tissue resection was performed underneath the mucosal flap, which served as protection for peritoneal cavity from contamination. In the second study, a metal anchor was pushed through the endoscope and through the gastric wall to the peritoneal side. Once on the other side, the anchor retracted the tissue which was then cut by an endoscopic metal cutter from the luminal side. Limitations of these studies were the need for additional steps of submucosal tunnel creation as well as suture placement and the need for a novel circular cutter and anchor devices .
Application of cut-then-close techniques in human
Direct endoscopic full thickness resection has been implemented for resection of gastric submucosal tumors . With this technique, needle knife dissection is performed into the muscularis propria to the level of the submucosal tumor, and the tumor is then dissected from the gastric wall. Perforation occurred as anticipated. Once the tumor is removed, the wall defect is closed with through-the-scope clips or other endoscopic closure devices. There were no reported significant complications with these techniques, however creation of a large wall defect and loss of air/CO 2 may potentially lead to collapse of the gastrointestinal lumen and inability to maintain a reasonable operative field. Additionally, creation of a significant size perforation could lead to peritoneal cavity contamination.
Due to the above-mentioned limitations, the cut-then-close technique has not gained popularity for the diagnosis of gastrointestinal neuromuscular diseases. Instead close-then-cut approach has become a more viable option.
Animal studies for close-then-cut techniques
In 2008 Rajan et al. performed a “no hole” double endoscopic mucosal resection (EMR) method for muscle biopsy on the stomachs of six pigs . The procedure started with a cap-assisted EMR to expose the gastric muscle layer. In the next step, a pseudopolyp was created by suctioning the muscle layer. Two endoloops were placed at the base of the pseudopolyp, after which the latter was resected via snare electrocautery above the endoloop level. All six samples confirmed the presence of a full-thickness gastric wall. There was a high rate of delayed perforation with resultant peritonitis, proving the method to be unsafe.
In a more recent study, Rajan et al. performed full-thickness biopsies of the duodenum and rectum on five pigs using an over-the-scope Padlock Clip (11 mm diameter; Aponos Medical, Kingston, NH). The Padlock clip is a star-shaped nitinol ring with six inner needles preassembled on an applicator cap. In both duodenal and rectal locations, and after the clip was deployed, resection was performed above the clip level. In the rectum, “double-resection” was used, wherein mucosa was first resected using band ligation. There were no procedural adverse events and the muscularis propria was confirmed in all resected specimens. This was the first study to show feasibility of EFTB in the duodenum and colorectum in the animal model. Nevertheless, appropriate cautions need to be taken to avoid potential adverse events, including perforation (immediate or delayed), bleeding, stricture formation and adjacent organ entrapment.
Clinical studies of close-then-cut techniques for EFTB
Clip assisted EFTB with a non-dedicated device
Rajan et al. have applied the “no-hole” gastric muscle biopsy using over-the-scope clip (OTSC) in three patients with symptomatic refractory idiopathic gastroparesis. The first step was endoscopic mucosal resection (EMR) using the band ligation approach (Duette; Cook Endoscopy, Bloomington, Inc) to unroof and expose underlying MP. The MP was then retracted into the cap of an over the scope clip (Padlock Clip with diameter of 11 mm; Aponos, Kingston, NH). The clip was deployed and the pseudopolyp of MP was resected via hot snare. There was no immediate or delayed perforation. All three patients developed upper abdominal pain without fever or leukocytosis. On follow-up endoscopy, a well-healed scar was seen in two patients and minimal residual ulceration with a retained clip in 1 patient. Mean length of the MP was 10.3±1.5 mm. Myenteric ganglia and interstitial cells of Cajal were obtained in all patients.
Ngamruengphong et al. described their experience in 13 patients with diffusely delayed colonic transit using the close-then-resect EFTB technique with a modified OTSC. A 4 mm distal clear straight cap (Olympus America, Center Valley, PA) was placed at the tip of an upper endoscope followed by a 14 mm OTSC (14/6 t) mounted over the cap in order to increase total cap length and thus increase the amount of tissue captured. Rectal muscle resection was performed 5 cm above the anal verge, which projected into the infraperitoneal rectum. Posterior wall was chosen to avoid injury to surrounding genitourinary structures. A rat-tooth forceps (Olympus America) was passed through the working channel of the endoscope and grasped the targeted area at the posterior rectal wall. The tissue was retracted without suction into the OTSC cap, after which OTSC was deployed. The created pseudopolyp was then resected 5 mm above the deployed OTSC using a dual knife (Olympus America) or a snare. Technical success was achieved in 100% of cases. In the post-procedure period, only two mild adverse events were observed: self-limited bleeding with the stable hemoglobin and moderate rectal pain without fever or leukocytosis. Presence of both inner circular and outer longitudinal muscle layers as well as multiple myenteric ganglia and submucosal ganglia was seen on microscopic examination. Two patients were found to have decreased ICC. Results of the EFTB impacted diagnosis and management in four patients: diagnosis of idiopathic slow transit constipation was made in two patients and Hirschsprung’s disease was excluded in another two patients. The authors demonstrated safety, feasibility and effectiveness of EFTB using a modified OTSC. Although in this study, the described technique for a full-thickness biopsy is for the rectum, it is likely that this technique could be feasible for the gastric location as well, as increasing total cap length would increase the amount of tissue captured.
Clip-assisted EFTB with a dedicated device
The idea of a dedicated flexible endoscopic full-thickness resection device (FTRD) was introduced in 2001 by Schurr et al. based on a stapling approach. The authors later developed a new device based on over-the-scope-clip mounted on an applicator cap thus developing a new FTRD . The cap was enlarged to a depth of 25 mm with a diameter of 14 mm. A snare was added to the frontal section of the cap and the clip was modified from the original OTSC to improve lateral tissue capture by means of additional teeth on the lateral aspect of the clip. As in conventional OTSC system, the clip was released by means of a thread running through the working channel of the endoscope. The electrical snare was closed by means of a separate handle. An instrument could be used through the working channel of the endoscope to facilitate tissue invagination into the cap. The device was mounted on the tip of the endoscope and could be used in the colon or upper digestive tract. Once at the site of the lesion, the tissue was gradually manipulated into the cap by means of a forceps and endoscopic suction. Subsequently the clip was released compressing the double layers of the wall. A pseudopolyp was created and resected by activating the integrated snare above the clip level . The device has since been used by several authors for various therapeutic purposes and locations including the colorectum .
Valli et al. described colorectal endoscopic full-thickness resection using FTRD in four patients with suspected severe neuromuscular gut disorders. Technical and histologic success rate was 100%. The mean diameter of the resected specimen was 21 mm (range, 20–22). No adverse events connected to the procedure occurred. In all four patients, the procedure helped establish a diagnosis either by demonstrating specific findings of a condition, or by ruling out the alternative diagnosis. In this small size study, FTRD was a safe and an effective method of full-thickness tissue acquisition ( Fig. 35.1 ) .