Endoscopic submucosal dissection of lower gastrointestinal lesions

Colorectal lesions may be benign, premalignant or malignant. Patients may be asymptomatic or may present with blood in the stool, change in bowel habit, abdominal pain or unexplained weight loss.

Lower gastrointestinal lesions may be investigated radiologically and/or endoscopically. Treatment normally involves resection of the lesions, which may be performed endoscopically or surgically. Current management of small colorectal lesions usually involves snare polypectomy or endoscopic mucosal resection (EMR). EMR usually removes lesions in small pieces, while endoscopic submucosal dissection (ESD) aims to resect lesions intact.

In ESD, an electrocautery knife is used to resect the lesion in one piece (en bloc), aiming to reduce the risk of recurrence and to allow a more accurate histopathological assessment.

The procedure is performed with the patient under sedation (usually) or general anaesthesia. Lesions are visualised at colonoscopy and the submucosa injected with fluid to raise the lesion.

A circumferential mucosal incision is made initially around the lesion with the electrocautery knife. Submucosal dissection is then performed under endoscopic visualisation, parallel to the muscle layer. Diathermy coagulation is used to achieve haemostasis, but endoscopic clips may also be required to control bleeding and/or treat small perforations.

A systematic review of 14 studies including 1,314 patients reported rates of en bloc lesion resection of 85% and complete cure (en bloc resection with histologically clear margins) of 75% (follow-up not stated).

A non-randomised comparative study of 536 lesions (number of patients not stated) reported higher rates of en bloc resection in patients treated by ESD (99%, 463 out of 468) compared with those treated by 'simplified' ESD using a snare (91%, 40 out of 44) and those treated by small incision EMR (83%, 20 out of 24; p<0.004 for both comparisons with ESD).

Case series of 400 (405 lesions) and 278 (292 lesions) patients, reported en bloc resection rates of 87% (352 out of 405) and 90% (263 out of 292) of lesions respectively. Case series of 278, 42 and 35 patients reported en bloc resection with completely free margins in 15% (44 out of 292) of lesions, and 74% (31 out of 42) and 63% (22 out of 35) of patients respectively.

The case series of 278 patients reported recurrent rectal intramucosal cancer in 3% (1 out of 38) of lesions with incomplete resection that were followed-up for a median of 36 months (cancer was successfully removed).

The Specialist Advisers listed key efficacy outcomes as one-piece resection rate, complete resection rate with clear margins, endoscopic cure rate, clinical cure rate and avoidance of bowel resection.

The non-randomised comparative study of 536 lesions reported perforation in 1% (7 out of 468), 5% (2 out of 44) and 0% (0 out of 24) of patients treated by ESD, 'simplified' ESD using a snare, and small incision EMR respectively (no further details provided). The case series of 400 patients reported colonic wall perforation in 3% (14 out of 405) of patients: all were detected intraprocedurally; all were managed successfully with endoscopic clips insertion; and 1 required surgical repair. In a case series of 186 patients (200 lesions) a perforation rate of 6% (12 out of 200) was reported: 11 of the perforations were detected intraprocedurally and 1 was detected 2 days later (requiring surgical repair).

The case series of 186 patients reported rectal bleeding, prompting emergency colonoscopy for application of endoscopic clips in 1% (2 out of 200) of lesions. One bleed occurred on the day of the procedure and the other occurred 10 days later.

Acute intestinal obstruction 18 hours after the procedure was reported in a case report of 1 patient (treated by aggressive fluid resuscitation and colonoscopic decompression).

The Specialist Advisers considered theoretical adverse events to include conversion of a curable cancer to an incurable cancer because of perforation.