Colorectal LLM Wiki

Browse
Home / Procedures / Endoscopic Polypectomy, EMR & ESD
Procedure created 2026-08-24 · updated 2026-08-25
colorectalendoscopypolypectomyEMRESDcolonic-stentingtextbookreference

Endoscopic Polypectomy, EMR & ESD

Endolumenal management of colorectal polyps — from basic forceps/snare polypectomy through endoscopic mucosal resection (EMR), endoscopic submucosal dissection (ESD), and colonic stenting for obstruction. Synthesized from ASCRS Ch. 5. Textbook reference content, not primary literature. Complements Colonoscopy & Lower GI Endoscopy (general colonoscopy technique) and Local Excision of Rectal Neoplasms (the surgical/transanal excision techniques — TAMIS, TEM — for lesions not amenable to purely endoscopic removal); that page's brief EMR/ESD mention has been trimmed to point here for full mechanistic detail.

Forceps and snare polypectomy

Polypectomy is required in 30–40% of all colonoscopies. Forceps: cold biopsy (opens to 6mm), jumbo cold biopsy (opens to 8.6mm, superior complete-resection rate vs. standard), and hot biopsy forceps — the latter's theoretical advantage (fulgurating residual dysplastic tissue) has been refuted (17% persistent-tissue rate on repeat endoscopy in one series), and hot biopsy carries more delayed-hemorrhage risk plus histologic distortion; ESGE recommends against routine hot biopsy forceps use. Cold forceps can also methodically avulse large spreading polyps too big for snare resection, sometimes combined with hot-snare-tip soft coagulation of the margins ("CAST" technique) for non-lifting laterally spreading lesions.

Snares allow larger resections than forceps. For polyps <6mm, cold snare has a higher complete- resection rate than cold forceps (93% vs 76%). Cold snares have gained favor over hot (electrocautery) snares due to lower delayed-bleeding/thermal-injury risk, equivalent complete-resection rates, and shorter procedure time; thinner-wire dedicated cold snares improve complete resection further, especially for 8–10mm or sessile polyps. Current ESGE guidance: cold snare preferred for polyps <5mm (strongly favored for 6–9mm); hot snare reserved for sessile polyps 10–19mm (after submucosal injection, to decrease thermal-injury risk) and recommended for pedunculated polyps (decreases bleeding). Overall polypectomy bleeding rate ~1–2/1000 patients (10× colonoscopy-without-polypectomy risk) — cold-snare bleeding tends to be immediate/apparent at the time of colonoscopy, while hot-snare bleeding (0.1–0.7% of polypectomies) is often delayed up to 30 days. Prophylactic clip placement after routine polypectomy does not reduce delayed-bleeding risk and should be avoided (cost without benefit) — selective clip use for larger lesions is a separate question (see EMR below).

Optimal snaring: position the polyp at the bottom half of the screen (instrument channel exits at the 5 o'clock position); working the scope tip further from the lesion can improve the angle of attack for polyps in difficult positions. A 10% incomplete-resection rate for 5–20mm polyps was found in one 1,400-patient study; risk factors were larger size and sessile serrated morphology vs. adenoma.

Lifting (submucosal injection)

Flat lesions often need a submucosal lift to separate the resection target from the muscularis propria and reduce full-thickness injury risk. Normal saline is the standard agent (lasts ~3 minutes, equivalent to other agents for complete-resection rate/bleeding/perforation); more viscous/durable agents (hyaluronic acid, glycerol, dilute albumin, hetastarch, proprietary gels lasting >40 minutes) suit longer, more complex resections. Adding colored dye (methylene blue, indigo carmine) helps delineate the submucosal plane and can double as the technique for endoscopic tattoo placement. Injection should be tangential to the bowel wall (not perpendicular) for a taller, more localized lift; lesions that fail to lift may reflect wrong-plane injection, prior-attempt scarring, or submucosal invasion (see NICE classification/non-lifting sign below).

Endoscopic mucosal resection (EMR)

Used for sessile/flat lesions 20–50mm, confined to the superficial colon wall. Standard technique: submucosal lift → snare resection in as few pieces as possible, with a 2–3mm normal-tissue margin. Lesions >2cm are usually removed piecemeal (limiting pathologic margin assessment) — early EMR series showed 30% recurrent polyp formation on follow-up, though a more recent 1,000-procedure multicenter series found 17% overall recurrence (5% for polyps <20mm). Recurrence risk factors: size >40mm (OR 8.2 vs. <20mm), argon plasma coagulation (APC) use (OR 2.4 — APC likely only superficially ablates rather than eradicating polyp tissue), and bleeding (OR 1.6). Avulsion with hot biopsy forceps for residual tissue after EMR had a significantly lower adenoma recurrence rate than APC ablation (10% vs. 59% in one 278-patient study).

Clips: routine prophylactic clipping after EMR is discouraged on cost grounds, but selective clip placement to close mucosal defects — particularly for non-pedunculated polyps >20mm in the proximal colon — reduced post-polypectomy bleeding in a multicenter RCT (absolute risk reduction 6.3%: 9.6% bleed without clips vs. 3.3% with, p<0.001); no benefit was seen for distal-colon large polyps.

Underwater EMR: lumen filled with 500mL–1L sterile water instead of air, avoiding submucosal injection altogether (which can go to the wrong plane or make snare application harder as the snare slips over a distended mucosa). The polyp edges are marked with APC, then removed piecemeal with a cutting-current snare. Compared to traditional EMR, underwater technique has shown increased complete macroscopic resection and decreased recurrence in selected series, and has also been applied to ESD for the same theoretical reasons (mucosa floats away from submucosa, improving the dissection plane; submersion also improves heat dissipation, theoretically reducing thermal injury).

Endoscopic submucosal dissection (ESD)

Developed in 1990s Japan for early gastric cancer; colonic ESD first described in the early 2000s and is more technically challenging (thinner colon wall, higher perforation risk) but achieves a more complete resection with a substantially lower recurrence rate than EMR — a >350-patient retrospective comparison found a sevenfold lower recurrence rate with ESD, at the cost of a nearly fivefold higher perforation rate (6.2% ESD vs. 1.3% EMR). Only ~20% of polyps endoscopically deemed unresectable and referred to surgery actually harbor invasive malignancy on final pathology — most large "benign-appearing" polyps can be managed endoscopically, sparing the morbidity of colectomy. Comparing ESD to laparoscopic formal resection: ESD patients had significantly shorter hospital stay and lower cost, with similar complication rates but lower complication severity.

Complications: abdominal pain, bleeding, perforation, and recurrence are the most common (as with most endoscopic polypectomy techniques). Post-ESD electrocoagulation syndrome (similar to post-polypectomy syndrome, see Colonoscopy & Lower GI Endoscopy) occurs in up to 40% of patients. Post-ESD bleeding occurs in ~2–7%. Perforation rate 5–20%, associated with larger tumor size and fibrosis; perforation during ESD of a malignant lesion carries a risk of peritoneal tumor seeding (9% in one 22-perforation gastric-ESD review — a gastric-literature figure, extrapolated cautiously to colon). Local recurrence after ESD is low (~1%), and even recurrences have generally been manageable with repeat endoscopic resection (none containing invasive cancer in one series).

Patient selection: EMR generally favored for lesions <2cm; ESD reserved for lesions >2cm without malignancy features, or as an excisional-biopsy technique when the diagnosis is unclear (used cautiously given increased full-thickness-injury risk from malignant-invasion-related fibrosis). Endoscopic classification systems used to predict underlying malignancy before deciding EMR vs. ESD vs. surgical referral:

  • Paris classification — polypoid vs. non-polypoid gross morphology; non-polypoid lesions subdivided by protrusion (slightly elevated, flat, slightly depressed, excavated) — depressed lesions carry higher malignancy risk, though inter-observer reliability of the full system is limited (motivating interest in a simpler pedunculated/elevated/depressed three-category scheme).
  • Kudo pit pattern / NICE (NBI International Colorectal Endoscopic) classification — combines narrow-band-imaging color, vessel pattern, and surface pattern into three types: Type 1 (same/ lighter color, no/isolated vessels, dark-or-white-spot or homogeneous pattern) → most likely hyperplastic; Type 2 (browner, brown vessels around white structures, regular oval/tubular/ branched white structure) → most likely adenoma; Type 3 (brown-to-dark-brown, disrupted/missing vessels, amorphous/absent surface pattern) → most likely deep submucosal invasive cancer. Reported accuracy 98%, PPV 80%, NPV 95% in one series.
  • Non-lifting sign: failure of a lesion to lift with submucosal injection is a strong (though not perfectly sensitive) invasive-malignancy indicator — overall accuracy 95%, sensitivity only 62%, specificity 98% in one 270-lesion study — meaning endoscopic visual assessment is actually more sensitive than the non-lifting sign for detecting invasion in flat/depressed lesions; the non-lifting sign's main value is that a well-lifting lesion with benign morphology is still a reasonable ESD candidate even after failing to lift for a technical (not malignant) reason (fibrosis, prior injection scarring).

ESD technique: mark a 2–3mm margin outside the lesion edge → submucosal injection outside the marks → circumferential incision (partial incision preferred over complete, to avoid submucosal fluid leak reducing subsequent lift) → tangential submucosal dissection with repeated injection, using a clear cap attachment and gravity-favorable positioning for exposure/traction → hemostasis of visible vessels during dissection → closure of the resection bed with clips, endoscopic suturing, or an over-the-scope clip for larger defects. A hybrid ESD/EMR approach (ESD to define borders and start dissection, large-snare EMR to finish the central resection) trades a higher recurrence rate for shorter procedure time. Learning curve is substantial — up to 30 supervised cases for technical proficiency even in high-volume centers, with perforation rate dropping from 12% (first 100 cases) to 2% (second 100 cases) in one single-center series; en bloc resection rate similarly improved (80%→92%).

Postoperative care: EMR patients typically go home same-day; ESD patients may benefit from overnight observation. No prophylactic antibiotics needed for either. Abdominal X-rays are frequently used after difficult ESD to screen for free air.

Controversies with ESD vs. EMR: critics argue en bloc resection of benign-appearing lesions is unnecessary overtreatment (malignancy rate <10%) and that EMR's higher recurrence rate is readily managed with surveillance endoscopy. Proponents note EMR's piecemeal excision is detrimental for lesions actually harboring carcinoma (fragments the specimen, compromising margin assessment), and that ESD can be curative for superficial (Sm1, <1000μm submucosal invasion) cancers given their low lymph node metastasis rate — but this is a narrow population: in one 1,100-patient ESD series, 19% had invasive cancer, half of which were Sm2 (requiring surgical resection anyway) — so only ~10% of ESD patients actually captured the "complete resection of a superficial malignancy" benefit that justifies the procedure's added complexity/perforation risk over EMR.

Endoscopic suturing and stabilization platforms

Endoscopic suturing (e.g., the OverStitch® platform, 2006) closes large post-EMR/ESD defects more durably than clips — a dual-channel endoscope with a grasper passes suture through the mucosa in interrupted or running fashion, partial- or full-thickness; has been used to close full-thickness defects without needing transabdominal surgery.

Stabilization platforms address the challenge of operating through a mobile flexible scope in a folded lumen — double-balloon (DiLumen®) and stabilization-cage-plus-working-channel (ORISE Tissue Retractor System®) devices create a flattened, stable therapeutic zone and counter-traction to ease dissection during advanced endoscopic surgery.

Approach to referral for an "unresectable" polyp

Historically, colectomy for endoscopically-challenging large polyps carried only a 20% malignancy rate — implying 80% of patients were over-treated with a formal resection. Evaluation of referral photos/pathology for ulceration, mucosal contour, and vascular pattern (or repeat colonoscopy with attempted ESD/EMR if photos are inadequate) can often identify polyps still amenable to endoscopic removal. A step-up approach in the OR — endolumenal resection first, escalating to combined endoscopic-laparoscopic surgery (CELS) or formal laparoscopic colectomy only if needed, with the patient consented for all three — lets the least invasive option be attempted first while avoiding a second procedure if it fails; the main downside is scheduling complexity (needing all resources/ consent ready for the most invasive contingency).

Colonic stenting

Self-expanding metallic stents (SEMS, uncovered — to allow tissue ingrowth and prevent migration) offer a minimally invasive alternative to emergent surgery/ostomy for large bowel obstruction. Two indications, no role for prophylactic stenting:

  • Bridge to surgery (left-sided colonic obstruction): relieves obstruction, allows colonic edema to resolve, and enables elective (often primary, non-diverted) anastomosis instead of emergency surgery. Fivefold decreased likelihood of permanent stoma formation, with a significant increase in primary anastomosis rate and decreased wound infection rate vs. straight-to-surgery — this directly answers Large Bowel Obstruction's previously flagged gap on stenting outcomes. Complications: perforation 5%, migration 4–10%, repeat obstruction 30%. Two RCTs of colonic stenting as bridge-to-surgery were closed early due to stent-related complications, and the only current bridge-to-surgery guideline (ESGE 2014, meta-analysis-based) recommends against stenting as the preferred method in an otherwise-healthy patient with a left-sided obstruction, but says it could be considered in higher-anesthetic-risk patients (ASA ≥3 or age >70) — refining Obstructing Colon Cancer's previously flagged "no comparative data" gap: there is guidance, it's just risk-stratified rather than universal. Since that 2014 guideline, at least two systematic reviews/meta-analyses have found bridge-to-surgery stenting oncologically safe (similar 5-year overall survival, disease-free survival, and local recurrence vs. emergent surgery) — so current practice is at the individual surgeon's discretion, weighing patient risk profile. Extrinsic causes of colonic narrowing (e.g., metastatic disease compressing the colon from outside) have a higher stenting migration/perforation rate than intrinsic obstructing lesions, per the chapter authors' stated personal experience (not a cited study) — because the colon wall isn't thickened/diseased enough to allow stent ingrowth and anchoring.

The named early-closed RCTs and subsequent trial evidence, per ASCRS Ch39 - Large Bowel Obstruction (cited at the textbook/secondary level only — none of the below independently ingested from primary literature): the Dutch Stent-In 2 trial was stopped when interim analysis revealed a higher-than-anticipated morbidity rate in the SEMS arm, driven mainly by a high perforation rate (13%) — a relatively high 17% failure-to-pass-a-guidewire rate raised the question of whether adequately experienced endoscopic personnel were staffing the trial sites. Pirlet's RCT was closed early when an interim analysis found the primary outcome (decreased ostomy rate) not achieved in the SEMS group vs. straight-to-surgery — attributed to a low rate of successful stent deployment. Alcantara's RCT was closed early due to a high morbidity rate in the emergent-surgery arm (the inverse direction from Stent-In 2/Pirlet). A meta-analysis of these and five other RCTs (Arezzo, 8 trials/497 patients) found no significant difference in 60-day mortality between SEMS and emergent surgery (9.6% vs. 9.9%, RR 0.99, p=.97), but a significantly lower temporary-ostomy rate favoring SEMS (33.9% vs. 51.4%, RR 0.67, p<.001). The UK's CREST trial (246 patients, published in abstract form as of the source chapter) found no significant difference in 30-day mortality (5.3% SEMS vs. 4.4% surgery), but markedly lower ostomy-formation rates with SEMS (45% vs. 69%, p<.001) — the largest RCT on this question to date, run with stenting workshops beforehand to standardize technique.

Covered vs. uncovered stents: a meta-analysis (Mashar, one RCT + nine observational studies) found no difference between covered and uncovered stents in successful deployment, perforation, or bleeding rates, but uncovered stents had a lower risk of tumor overgrowth (RR 0.29), lower risk of stent migration (RR 0.29), and lower need for stent reinsertion (RR 0.38) — consistent with this section's existing uncovered-stent rationale above (tissue ingrowth prevents migration, at the cost of potential tumor overgrowth long-term). - Palliation (unresectable/metastatic malignant obstruction): ESGE-preferred method for palliation. Similar mortality to emergent surgery with shorter hospital stay and decreased stoma rate per a recent meta-analysis; long-term stent placement allowed 95% of patients to avoid stoma formation in one study.

Technique: water-soluble contrast enema (fluoroscopy or CT) maps tumor location/stenosis length/ lumen caliber and rules out an obstructing cause needing a different approach. A guidewire crosses the lesion under fluoroscopy; location is confirmed by catheter contrast/air injection (haustration visible on double-contrast). The largest-diameter, longest available stent is generally favored; shorter stents are chosen for rectal lesions to keep the stent ≥5cm from the anus (avoids significant tenesmus). A radio-opaque clip 5cm distal to the lesion marks the landing zone. Balloon dilation of the stent is not recommended; the scope is not passed back through the stent after placement (risk of dislodgement). Abdominal X-ray confirms position/rules out free air; expansion occurs over the following 48 hours; stool softeners help prevent fecal impaction of the stent. If stenting fails or complicates, emergent surgery with diverting colostomy follows.

Stenting anastomotic leaks: covered stents (originally an esophageal device) have been used off-label, with case-report-level evidence, to treat contained colorectal anastomotic leaks by blocking further stool extravasation and allowing the sinus to heal — left in place 20–50 days (small cohort range), with 80–100% successful closure reported after removal, though migration rates are high and may require stent replacement.

  • Colonoscopy & Lower GI Endoscopy — general colonoscopy technique and complications this page's procedures are performed within.
  • Local Excision of Rectal Neoplasms — the surgical/transanal (TAMIS/TEM) alternative for lesions not suited to purely endoscopic removal.
  • Large Bowel Obstruction — its previously flagged "endoscopic stent outcomes not covered" gap is filled by this page's Colonic Stenting section.
  • Obstructing Colon Cancer — its previously flagged "no comparative bridge-vs-surgery data" gap is refined (risk-stratified guidance exists) by this page.
  • Rectal Cancer — colonic stenting for rectal lesions and the watch-and-wait/local-excision interplay both touch this page's content.
  • Colorectal Polyps — that page's Malignant polyps section (substantially expanded via ASCRS Ch23 - Management of Malignant Polyps) builds directly on this page's Paris/Kudo/NICE classification and EMR/ESD technique detail without repeating it; see that page for malignant-polyp-specific depth-of-invasion, lymph node metastasis risk, and curative-resection criteria.

Open items / gaps

  • No primary literature yet on any specific cited trial (CAST technique, clip-closure RCTs, the three closed-early stenting RCTs now named — Stent-In 2, Pirlet, Alcantara — plus CREST/Arezzo/ Mashar, and the ESD learning-curve series) — presented as the textbook's own synthesis, per ASCRS Ch05 - Endoscopic Management of Polyps and Endolumenal Surgery and ASCRS Ch39 - Large Bowel Obstruction.
  • The chapter's claim about extrinsic-cause stenting having worse outcomes is explicitly sourced to the authors' personal experience, not a cited study — flagged as lower-strength evidence within this page itself, per the source's own framing.
  • ~~Interaction between this page's ESD/EMR malignancy-risk figures and Colorectal Polyps's existing malignant-polyp-management content not yet cross-checked for consistency~~ — resolved: ASCRS Ch23 - Management of Malignant Polyps substantially expanded that page's malignant-polyp content with quantitative depth-of-invasion/lymph-node-metastasis data and a formal curative-resection criteria set, cross-linked back to this page's technique detail.