Resecting a small gastrointestinal stromal tumor: feasible, but are you feeling lucky today?

Gastrointestinal stromal tumors (GISTs) have the potential to undergo malignant transformation. Given that the endoscopic ultrasound (EUS) appearance of GISTs is not specific, and that EUS-guided tissue sampling may not be sufficient to distinguish a GIST from a leiomyoma, or a malignant GIST from a benign one, an argument can be made for resection of any lesion suspected of being a GIST, regardless of its size. Various endoscopic techniques have been developed to resect or ablate suspected GISTs. These include endoscopic submucosal dissection (ESD), endoscopic enucleation, endoscopic ligation, endoscopic ligation and enucleation (ELE), endoscopic full-thickness resection, and submucosal tunneling endoscopic resection (STER) [1].

“While perforation may be endoscopically treatable and tolerated in most cases, an ideal technique of GIST removal would render this complication unnecessary and avoidable.”

In this issue of Endoscopy, Zhang et al. from Shenzhen present their technique of ELE for small (< 2 cm) subepithelial tumors (SETs) originating from the muscularis propria [2]. All SETs reported in this case series had purely intraluminal growth on EUS. 

The technique consisted of first injecting the submucosa to raise the mucosa away from the SET, then incising the overlying mucosa with a hook knife to create a large mucosal defect. A cap attachment was then used to generate and apply suction, allowing the SET to be aspirated into the cap lumen through the mucosal defect and ligated at its base with an endoloop. The SET was then transected above the loop ligature with a snare. After resection, if residual tumor was suspected on endoscopic inspection, repeat loop ligation was performed, leaving the ligated tissue to necrose and slough off. The mucosal defect was closed with clips.

In total, 14 SETs arising from the muscularis propria with a mean diameter of 11 mm were resected, of which nine were GISTs and four were leiomyomas. The technical success rate was 100 %. Four perforations occurred, but these were able to be closed with clips without further clinical sequelae.

Ligation-assisted endoscopic enucleation of small intraluminal subepithelial tumors originating from the muscularis propria is not a new technique. Several variations have been reported. In 1996, Chang et al. described the adjunctive use of band ligation for the safe resection of a 1.2-cm leiomyoma [3]. In 2010, a study from Taiwan reported band ligation of small gastric stromal tumors using bands and detachable 30-mm loops, followed by snare resection, with no complications [4].

Incision of the mucosa overlying the SET, coupled with ligation using bands or loops has also been previously described. The Tianjin group injected the submucosa with a methylene blue solution and then performed a hemicircumferential to circumferential mucosal incision using a triangle knife or hook knife to expose the underlying SET [5]. Cap-assisted ligation (band or loop) was then performed to enucleate the SET using an ESD technique. Of 22 resections, there were three perforations (14 %), all of which were successfully closed with clips. The Shengjing group performed combined incision and loop ligation in the reverse order, first deploying a band, then incising the mucosa without submucosal injection [6]. A hook knife was used to dissect out the SET, sometimes assisted by a snare.

What is the rationale for ligation prior to enucleation? Ligation alters the lesion morphology so it becomes more polypoid, ideally creating a pseudostalk, which makes instrument access and resection easier. Ligation may also demarcate the resection plane below the SET. Whether ligation prior to resection improves safety is questionable. A reduction in the risk of delayed bleeding or perforation would only be expected if the ligature remained in place after resection for a sustained period. It is more likely that the ligature will fall off prior to complete wound healing, though a very tight ligature may hold secure. A ligature may also be more likely to hold if full-thickness ligation is achieved, with resultant serosa-to-serosa apposition.

Given the limitation of the current tools for intended full-thickness ligation, the absence of perforation after ELE should at least raise concerns about incomplete resection. The perforation rate in the Shenzhen group’s study was 29 % (4 of 14 resected SETS); this may be viewed as “high” if perforation is regarded as an undesirable outcome, but also as “low” if perforation is viewed as an inherent consequence of the full-thickness resection required for achieving and verifying R0 resection.

Enucleation of a SET can be performed without ligation. Zhang et al. from the Taizhou hospital treated a population of small SETs originating from the muscularis propria by enucleation alone [7]. The submucosa was first injected with a saline solution mixed with indigo carmine and adrenaline, and this was followed by a mucosal incision to expose the underlying SET. The SET was then enucleated using a hook knife or needle knife. The perforation rate was 33 % (23 /69), similar to the rate reported in the Shenzhen study; however, four patients with perforations required laparoscopic repair.

When ligation is performed for small lesions, should we use a band or a loop? No study has provided a head-to-head comparison. In contrast to a band, which has an inherent residual diameter after deployment, a loop can theoretically be more tightly and fully closed until all potential space between the loop’s opposing arcs is obliterated. Though loops are known to be prone to slippage, the ability for tighter closure should generate a superior strangulation effect compared to band ligation. Tighter closure may also define the ligation plane more clearly when subsequent resection is pursued. On a practical note, the Shenzhen study used loops that were 20 mm in diameter, deployed in a 15-mm cap (D-201-11802; Olympus). Some centers may only have access to 30-mm loops, which are more difficult to pre-loop along the circumferential ridge of the cap. 

The main limitation of ELE is that the technique may visibly remove the SET, thus lending the impression of curative resection, but cannot guarantee R0 resection without production of a full-thickness resection specimen. The authors claim “complete removal” in all patients, but provide no data regarding en bloc resection, tumor margins, or capsule integrity [2]. Microscopic residual tumor may be left embedded at the resection base or within the resection scar. The author’s technique of ligating the base in a “loop-and-let-go” fashion to induce necrosis and ablative sloughing is not a reliable method to ensure tumor-free margins.

Lacking documentation of R0 resection, clinical follow-up is mandatory. The follow-up period of 3 – 6 months reported in this study is far too short to be confident that a GIST was curatively resected and will not recur. As a rule, in the absence of documented R0 resection, patients undergoing ELE should not be considered “cured”, nor should they be discharged from follow-up. Of note, this limitation applies equally to any method that does not achieve full-thickness resection, including ESD.

If we accept the premise that R0 resection can only be accomplished by full-thickness resection, then perforation is an inherent and necessary trade-off. The question then boils down to the endoscopist’s comfort level managing perforation. The recent availability of novel clips and suturing devices has certainly made closure of perforations easier, but no less intimidating. Once perforation occurs, the peritoneum has been contaminated and the clock is ticking. Pneumoperitoneum may need to be relieved by percutaneous puncture. Antibiotics need to be administered and the patient should be admitted for overnight observation.

Is the benefit of R0 resection worth the perforation? For larger GISTs, perhaps “yes” – as an alternative to surgery. But for smaller GISTs that are nearly always benign and where the management alternatives include observation, perhaps “no.” I would argue that the risks of perforation reported with any of the resection methods, including ELE, outweigh the benefit when dealing with small SETs that appear benign on EUS. 

Is there a safer alternative to full-thickness resection that can accomplish the same outcome? A technique of “full-thickness ligation” without complete resection may qualify. We developed a technique of “ligate unroof biopsy” (LUB) with the specific intent of full-thickness ligation for the diagnosis and treatment of SETs [8] [9]. The bowel wall is involuted using either aggressive cap suction or instrumental retraction to enable full-thickness ligation with serosa-to-serosa apposition. After it has been ligated, the SET is “unroofed” by incising the overlying mucosa to expose the underlying SET for generous tissue sampling. All treated GISTs were found on surveillance EUS to have no evidence of residual or recurrent disease. Further study is certainly needed to determine whether we can truly rely on ischemic ablation and spontaneous enucleation after LUB to provide a cure for GISTs.

In the management of SETs, each endoscopist must carefully weigh the risk and benefit on an individual basis and opt for a course of action that reflects the available resources and expertise. A community practice will have different constraints from a tertiary, academic one. ESD, STER, and full-thickness resection are time-, labor-, and cost-intensive procedures. ELE may sound easier and quicker, but its high perforation rate still virtually disqualifies it from use in a community setting. While perforation may be endoscopically treatable and tolerated in most patients, an ideal technique of GIST removal would render this complication unnecessary and avoidable. Future device innovation should enable the integration of full-thickness ligation with full-thickness resection to prevent any violation of the extra-gastrointestinal space. Small GISTs could then be removed in the community setting no differently from small polyps.

• References

• 1 Krinsky M, Binmoeller KF. Endoscopic ultrasound for the characterization of subepithelial lesions of the upper gastrointestinal. tract. Available from: https://www.uptodate.com/contents/endoscopic-ultrasound-for-the-characterization-of-subepithelial-lesions-of-the-upper-gastrointestinal-tract Accessed: 8 October 2018
• 2 Zhang D, Lin Q, Shi R. et al. Ligation-assisted endoscopic submucosal resection with apical mucosal incision to treat gastric subepithelial tumors originating from the muscularis propria. Endoscopy 2018; 50: 1180-1185
• 3 Chang KJ, Yoshinaka R, Nguyen P. Endoscopic ultrasound-assisted band ligation: a new technique for resection of submucosal tumors. Gastrointest Endosc 1996; 44: 720-722
• 4 Huang WH, Feng CL, Lai HC. et al. Endoscopic ligation and resection for the treatment of small EUS-suspected gastric GI stromal tumors. Gastrointest Endosc 2010; 71: 1076-1081
• 5 Zheng Z, Jiao G, Wang T. et al. Ligation-assisted endoscopic enucleation for the resection of gastrointestinal tumors originating from the muscularis propria: analysis of efficacy and facility. Dig Surg 2016; 33: 488-494
• 6 Guo J, Liu Z, Sun S. et al. Ligation-assisted endoscopic enucleation for the diagnosis and resection of small gastrointestinal tumors originating from the muscularis propria: a preliminary study. BMC Gastroenterol 2013; 13: 88
• 7 Zhang JS, Ye LP, Wang CY. et al. Endoscopic submucosal enucleation of small gastric gastrointestinal stromal tumors with cross-shaped incision: report of sixty-nine cases. Hepatogastroenterology 2012; 59: 440-443
• 8 Binmoeller KF, Shah JN, Bhat YM. et al. Retract-ligate-unroof-biopsy: a novel approach to the diagnosis and therapy of large nonpedunculated stromal tumors (with video). Gastrointest Endosc 2013; 77: 803-808
• 9 Binmoeller KF, Shah JN, Bhat YM. et al. Suck-ligate-unroof-biopsy by using a detachable 20-mm loop for the diagnosis and therapy of small subepithelial tumors (with video). Gastrointest Endosc 2014; 79: 750-755