"Salvage" endoscopic mucosal resection in the colon using a retroflexion gastroscope dissection technique: a prospective analysis

BACKGROUND AND STUDY AIMS: Endoscopic mucosal resection and submucosal dissection can provide curative endoscopic therapy for Paris type I/II adenomas and node-negative early cancer. No studies have addressed the technical feasibility of retroflexion endoscopic dissection methods for luminal "salvage" therapy in patients considered unresectable using conventional forward-viewing resection. PATIENTS AND METHODS: Colonoscopy using an Olympus GIF-XQ240 gastroscope was carried out in 76 patients with Paris type I/II adenomas, early colorectal cancer (CRC), or laterally spreading tumors (LSTs) when the index endoscopist considered the lesion to be unresectable due to retrograde fold involvement. Endoscopic mucosal resection (EMR) and submucosal dissection were carried out using a complete retroflexion technique. Endoscopic and miniprobe 20-MHz or 12.5-MHz ultrasound follow-up data were collected prospectively up to 24 months after the index resection. RESULTS: Cecal intubation or cannulation to the neoterminal ileum was achieved in 76 (100 %) cases. Forty lesions (53 %) were classified in accordance with the Paris criteria as Is; 16 (21 %) as type II; 10 (13.5 %) as LST-G; and 10 (13.5 %) as LST-NG. Eight lesions (10 %) were excluded from EMR on the basis of endoscopic ultrasound criteria, with 68 of the 76 lesions (89 %) meeting the criteria for endoluminal resection. The median intubation time was 16 min (range 3-32 min). The median resection times were 98 min (range 30 - 242 min), 36 min (range 10-60 min), 172 min (range 20 - 240 min), and 60 min (range 10-116 min) for Paris Is, II, LST-G, and LST-NG lesions, respectively. LST-G morphology was associated with a high median submucosal injection volume in comparison with all other Paris types ( P < 0.05) and with a prolonged resection time ( P < 0.01). Sixty-one patients (94 %) completed the surveillance protocol. Higaki recurrence criteria were met in seven patients (11 %), with six undergoing successful adjunctive endoluminal resection. After 24 months of follow-up, the "cure" rate with endoscopic resection was 60 out of 61 (98 %). CONCLUSIONS: This is the first prospective study to address the safety and medium-term efficacy of retroflexion endoscopic resection in the colon. When appropriate exclusion criteria are applied, selected patients can receive curative resection using the retroflexion technique. "Salvage" endoluminal therapy may therefore be possible in such cases when surgical resection would otherwise have been required.