Third-space endoscopy: Current techniques,clinical applications,and future directions

TSE

TSE is based on a unified set of procedural principles centered on safe access to and intervention within the submucosal space or “third space.” All TSE procedures begin with a controlled mucosal incision, followed by creation of a submucosal tunnel that provides a stable and relatively contained working space. Within this tunnel, a target-specific intervention is performed, such as myotomy (POEM), septotomy (Z-POEM and D-POEM), or resection (STER), depending on the underlying pathology. A defining feature of TSE is the preservation of the overlying mucosa at the treatment site (e.g. myotomy), which minimizes the risk of contamination. This separation of access from therapy allows a single technical platform to be adapted across diverse indications, contributing to the rapid expansion of TSE from POEM to a broad range of applications (Figure 2).

OPEN IN VIEWER

Figure 2.

Conceptual frame work and clinical maturity continuum of third-space endoscopy procedures.

POEM

POEM is an established treatment modality for patients with achalasia cardia and non-achalasia spastic esophageal motility disorders.

a. Technique

Mucosal lifting and incision

Several technical variations have been described, including the orientation of myotomy (anterior 1–2 o’clock vs. posterior 4–5 o’clock), length of esophageal myotomy (short vs. long esophageal myotomy), and depth of myotomy (selective circular vs. full-thickness). Notably, most of these variations have not demonstrated consistent or clinically meaningful differences in major outcomes including clinical success and gastroesophageal reflux disease (GERD). A shorter esophageal myotomy (<5 cm) is increasingly favored in patients with type I and II achalasia, primarily for procedural efficiency, with a possible, although not definitively proven, reduction in esophageal acid exposure.

b. Clinical outcomes

Technical success of POEM approaches 100% in experienced centers, while clinical success is reported in approximately 80%–90% of patients. Randomized trials indicate that POEM is superior to pneumatic dilatation and broadly comparable to laparoscopic Heller myotomy in terms of short- to medium-term efficacy.^6,7 Although long-term data remain limited, available evidence suggests durable symptom control in the majority of patients.^8–11

POEM has also demonstrated efficacy across diverse and traditionally challenging subgroups, including patients with spastic motility disorders, those with prior treatment failure, those with sigmoid achalasia, older patients, and pediatric populations.^12,13

Adverse events are generally infrequent and predominantly mild-to-moderate, including insufflation-related events (capnoperitoneum and capnothorax), bleeding, and mucosal injury. ¹⁴ Severe complications such as mediastinitis are rare, which has led to a shift toward shorter antibiotic prophylaxis strategies.^15,16

GERD remains the most important late adverse event. Both endoscopic and physiologic studies consistently demonstrate increased esophageal acid exposure following POEM, with randomized data suggesting higher reflux rates than those associated with both pneumatic dilatation and laparoscopic Heller myotomy with fundoplication.^6,7 However, clinically significant reflux is less common than objective reflux, and severe esophagitis (LA grade C/D) remains relatively infrequent. ¹⁷ Most patients respond to proton pump inhibitors, although a subset requires long-term therapy or adjunctive interventions such as endoscopic fundoplication.

c. Unresolved issues and future directions

Post-POEM reflux prevention remains the central unresolved issue. Technical modifications—including short gastric myotomy, sling fiber preservation, and POEM with fundoplication (POEM-F)—have been proposed, but the evidence remains inconsistent and largely nonrandomized. Other key areas for future research include long-term durability beyond 5–10 years, integration of EndoFLIP to tailor myotomy extent, identification of patient- and technique-related predictors of reflux and treatment failure.

D-POEM

D-POEM has emerged as a minimally invasive alternative to surgery for the management of epiphrenic diverticula. The procedure combines septotomy with esophageal myotomy to address the structural and functional components of the disease. ¹⁸ This approach also allows simultaneous treatment of the underlying motility disorder, which is present in up to 75% of patients with epiphrenic diverticula.

Technique

An important technical consideration is whether to perform POEM with septotomy (POEM + S) or POEM alone, with current data suggesting comparable short-term efficacy and safety.^19,20

a. Clinical outcomes

D-POEM demonstrates high technical success (92%–100%) and clinical success rates (84%–94%), with significant improvement in dysphagia scores and esophageal physiology.^20–23 These outcomes appear durable at medium-term follow-up (2–3 years), with sustained clinical success in approximately 85%–88% of patients and low recurrence rates (0%–2.6%). Safety outcomes are favorable, with overall adverse-event rates ranging from 2% to 11%, predominantly minor (e.g. subcutaneous emphysema and capnoperitoneum).

b. Unresolved issues and future directions

Despite these encouraging results, several limitations persist. First, the evidence base is dominated by retrospective series with heterogeneous inclusion criteria and outcome definitions. Second, the relative contribution of septotomy vs. myotomy remains unclear, particularly given comparable outcomes with POEM alone in selected cohorts. Third, long-term durability beyond 3 years and comparative effectiveness against surgical diverticulectomy with myotomy remain inadequately studied.

Gastric POEM (G-POEM) for refractory gastroparesis

G-POEM is an extension of TSE principles to the pylorus. It targets impaired pyloric relaxation which may act as a distal outflow obstruction in a subset of patients with gastroparesis.

a. Technique

The technical steps of G-POEM are now well standardized:

Mucosal incision in the distal antrum (typically 3–5 cm proximal to the pylorus)

Variations in technique include the approach to the pylorus (greater vs. lesser-curvature approach) and single vs. double myotomy.

b. Clinical outcomes

The technical success of G-POEM in published cohorts exceeds 95%, with overall adverse-event rates of approximately 5%–10%. Most adverse events are mild-to-moderate and include capnoperitoneum, intraprocedural bleeding, and mucosal injury.

Clinical success rates across studies range from 45% to 70%, with sustained improvement reported in 50%–80% of patients at 3–5 years of follow-up. ²⁴ The wide variation in clinical outcomes is likely attributable to variable definitions of clinical response (e.g. Δ gastroparesis cardinal symptom index (GCSI) ≥ 1 vs. ≥1.5), inconsistent use of objective endpoints (gastric emptying studies), and differences in follow-up duration across studies. A recent systematic review including 30 studies reported an overall clinical response rate of approximately 63%. ²⁵ Notably, prospective studies report more modest response rates (∼60%), suggesting that earlier retrospective data may have overestimated efficacy because of selection and reporting bias.^26–28

c. Patient selection and predictors of response

Identification of reliable predictors of response to G-POEM remains a major unmet need. Clinical variables associated with improved outcomes include older age, shorter disease duration, response to prior intrapyloric botulinum toxin injection, lower body mass index (BMI), and nondiabetic etiology. A high baseline GCSI score >2.6 (odds ratio (OR) 3.23) and baseline gastric retention >20% at 4 h (OR 3.65) have also been shown to be favorable predictors.^24,27,29 Negative predictors include longer disease duration, higher BMI, and psychiatric medication use (OR 1.33). ³⁰ Recent evidence incorporating wireless motility capsule data demonstrates that patients with preserved antral contractility exhibit significantly lower response rates, suggesting that pyloric disruption may be ineffective when upstream motor function is intact. ³¹ It is noteworthy that these associations have been inconsistently reproduced across studies, limiting their clinical applicability.

d. Unresolved issues and future directions

Despite rapid adoption, the evidence base for G-POEM has several critical limitations. First and foremost is the mechanistic heterogeneity underlying the disease. Gastroparesis is not a single disease, and pyloric dysfunction is not universal. Therefore, not every patient is likely to benefit from G-POEM, which may explain the heterogeneous responses reported across studies. Second, there is a lack of standardized definitions for clinical success, further contributing to the wide range of outcomes after G-POEM. Third, selection bias cannot be excluded because the bulk of the available data comes predominantly from single-center, high-volume cohorts. Finally, high-quality long-term studies are distinctly lacking in the evidence-based literature. There is only one sham-controlled trial, and durability beyond 1–2 years remains unclear. ²⁶

Submucosal tunneling endoscopic resection

STER enables en bloc resection of subepithelial tumors (SETs) arising from the submucosa or muscularis propria while preserving mucosal integrity. The main advantage of STER is that it minimizes the risk of perforation compared with exposed resection strategies.^32,33

a. Technique

Mucosal incision 2–3 cm proximal to the tumor

A variation of the technique is the double-opening technique, in which an additional mucosal opening is created along the distal end of the tumor to facilitate egress and improve exposure of the attachments.^32,34 While en bloc resection is the goal, piecemeal resection has also been reported as a safe and effective alternative in patients with large esophageal leiomyomas. ³⁵

b. Clinical outcomes

In large observational series and meta-analyses, STER has high technical success (>95%) and en bloc resection rates, with R0 resection rates typically ranging from 85% to 95%, although these vary according to lesion size and location.^36,37 Importantly, outcomes appear less favorable for larger (>3 cm) lesions or lesions located at the cardia, where technical complexity and the risk of incomplete resection increase.^38,39 Comparative evidence suggests that STER offers efficacy similar to that of endoscopic submucosal dissection (ESD) and endoscopic full-thickness resection (EFTR), with potential advantages of shorter procedure times and lower perforation rates. However, caution is advised when interpreting these findings, as they are derived largely from retrospective cohorts and indirect meta-analyses.^40,41 Adverse events occur in approximately 5%–15% of cases and most commonly include gas-related events, bleeding, or tunnel infection, the latter representing a unique complication spectrum specific to tunneling techniques. ⁴²

c. Unresolved issues and future directions

While STER is an established technique, several limitations merit consideration. First, the majority of data originate from high-volume Asian centers, limiting generalizability. Second, long-term oncologic outcomes (particularly for gastrointestinal stromal tumors) remain insufficiently characterized. Third, learning curve analyses confirm significant operator dependency, with procedural efficiency and safety improving only after substantial case volume. ⁴³ Future studies should focus on standardized selection algorithms, integration with hybrid techniques, and prospective comparative trials against EFTR and surgical approaches to better define the utility of STER. ⁴⁴

Zenker's POEM (Z-POEM)

Flexible endoscopic septotomy (FES) has traditionally been the cornerstone of endoscopic management for Zenker's diverticulum. However, symptom recurrence remains a recognized limitation, likely related in part to incomplete division of the cricopharyngeal septum. While a more extensive septotomy may reduce recurrence, it is often tempered by concerns regarding perforation. Z-POEM was developed to address this limitation by enabling complete septal division within the submucosal tunnel. Because the mucosa is preserved, the risk of perforation is negligible.

a. Technique

The conventional technique of Z-POEM is as follows

Mucosal incision on or proximal to the septum

Variations such as mucosal flap incision and “open Z-POEM” have been introduced to address limitations such as a remnant mucosal flap and difficult closure, respectively.^45,46

b. Clinical outcomes

Clinical outcomes across multicenter cohorts consistently report high technical success (97%–100%) and clinical success rates exceeding 90%, with rapid symptom resolution and a short hospital stay (≈1 day).^47,48 Importantly, durability appears acceptable, with recurrence rates of approximately 6%–8% at ≥2 years, which are lower than those reported in historical FES series. ⁴⁷ Comparative data remain underpowered but suggest equivalence in short-term outcomes between Z-POEM and FES, with a possible trend toward higher clinical success and lower recurrence with tunneling approaches.^49,50 However, these differences are modest and subject to study design, follow-up duration, and the definition of recurrence, raising concerns about overinterpretation of superiority claims.

Safety profiles are favorable, with adverse-event rates typically <10%, mostly mild-to-moderate, and rarely requiring intervention.^47,51 Notably, comparative datasets show no clear superiority of Z-POEM over other endoscopic approaches in terms of adverse events or hospital stay, suggesting that operator expertise may outweigh technique selection. ⁵¹

c. Unresolved issues and future directions

Future directions are constrained by three unresolved issues. First, there is no consensus gold standard, largely because of heterogeneity in patient selection, diverticulum size, and outcome definitions. ⁵¹ Second, the mechanistic advantage of tunneling, namely complete septal division, has not been conclusively linked to long-term clinical benefit, particularly reduced recurrence. ⁴⁹ Third, innovation (e.g. open Z-POEM) aims to address procedural inconveniences rather than clinically meaningful endpoints. High-quality randomized trials with standardized dysphagia scoring, objective functional endpoints, and long-term follow-up are essential before Z-POEM can be considered the preferred treatment modality.

Other applications

PREM is an extension of TSE principles to the colorectum. It targets the aganglionic or spastic distal bowel segment in patients with Hirschsprung's disease. The technique mirrors POEM, involving submucosal tunneling and selective myotomy of the affected rectosigmoid segment to relieve functional obstruction. ⁵² Evidence, however, remains extremely limited and heavily center-dependent, with early case reports and small series demonstrating symptomatic improvement, reduced laxative dependence, and acceptable safety profiles. ⁵³ However, short follow-up and the lack of comparative data against the current standard of care, namely surgical pull-through procedures, are noteworthy limitations. Consequently, PREM is still viewed as an experimental or highly selective alternative rather than a paradigm shift.

POETRE addresses a fundamentally different problem (complete esophageal obstruction) using bidirectional or antegrade submucosal tunneling to re-establish luminal continuity. ⁵⁴ While technical success has been reported in small prospective series, clinical durability is modest, with most patients requiring repeated dilations and adjunctive therapy. ⁵⁴ The procedure is technically demanding, relies on fluoroscopic and dual-endoscope guidance, and is associated with nontrivial adverse events. Overall, POETRE remains a rescue technique for highly selected cases.

Training, learning curve, and safe implementation of TSE

TSE has a long, experience-dependent learning curve, and training should not be reduced to a fixed case number alone. Available data suggest that for POEM, a first plateau in procedure time and adverse-event reduction is usually reached after approximately 20–40 cases, although this varies substantially according to prior experience. Endoscopists already proficient in ESD may achieve proficiency earlier, whereas less experienced operators may require a larger caseload. ⁵⁵ Training should therefore follow a stepwise pathway, beginning with theoretical knowledge of disease indications, anatomy, electrosurgery, devices, patient selection, and adverse-event management, followed by simulator/ex vivo or live animal model training before supervised human cases. The recent ESGE POEM curriculum further emphasizes procedural deconstruction, use of equipment checklists, standardized technique, recognition of anatomic landmarks, safe tunneling, appropriate myotomy length, mucosal closure, and proactive identification and management of mucosal injury, bleeding, and gas-related events.^56,57 Importantly, independent practice should be based on documented competence rather than case volume alone and should be supported by mentorship, high-volume training environments, direct observation, structured feedback, and prospective tracking of technical success, procedure time, adverse events, and clinical outcomes.