Selective Use of Endoscopic Submucosal Dissection Appropriate for Large Nonpedunculated Colorectal Neoplasms

Abstract

The report from Ohata et al1Ohata K. et al.Gastroenterology. 2022; Google Scholar on colorectal endoscopic submucosal dissection (ESD), with 5-year follow-up of colorectal lesions ≥ 20 mm in size, is impressive for its planning, size, execution, and description of technical expertise. However, the conclusion that “these results suggest that ESD can be conceived as a first line therapy for large superficial colorectal neoplasms” is unwarranted if it means all large (≥20 mm) nonpedunculated colorectal neoplasms (LNPCNs) require ESD. A number of reasons to reject this conclusion were presented in a letter by O’Sullivan et al2O'Sullivan T. et al.Gastroenterology. 2022; Google Scholar in response to Ohata et al. In addition to the concerns of O’Sullivan et al, several other points warrant mention:1.Serrated lesions accounted for 6% of the Ohata et al cohort. To the extent these were sessile serrated lesions (SSLs), these lesions have a low cancer prevalence compared with adenomas.3Parsa N. et al.Endoscopy. 2019; 51: 452-457Crossref PubMed Scopus (10) Google Scholar Because of their relatively flat morphology, cancer is nearly always endoscopically evident in SSLs when present. Further, SSLs can be reliably and very safely snare-resected without electrocautery.4Tutticci N.J. et al.Gastrointest Endosc. 2018; 87: 837-842Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar Together, these features indicate ESD is unnecessary for benign-appearing SSLs.2.Ohata et al reported that 52% of lesions in their cohort were granular laterally spreading tumors and discuss that these lesions warrant ESD because 10.8% had unexpected submucosal invasion in a previously reported study. To the extent the granular laterally spreading tumors in the Ohata et al cohort were of the homogenous type, this high rate of invasion is unexpected, which is typically associated with a risk of 1%–2% of submucosal invasion,5Moss A. et al.Gastroenterology. 2011; 140: 1909-1918Abstract Full Text Full Text PDF PubMed Scopus (444) Google Scholar thus requiring 50–100 ESDs to prevent 1 surgical resection in lesions with this morphology.3.Ohata et al reported 7.4% had T1a cancer (superficial submucosal invasion). Because only T1a cancers avoid surgical resection with ESD in their management scheme, 13 ESDs were required to avoid 1 surgical resection in patients with T1a cancer. Preventing unnecessary surgical resections is a worthy goal. However, the risks and costs associated with unnecessary surgical resection of benign colorectal lesions poses a substantially larger public health burden in some countries compared with the costs and risks associated with unnecessary surgical resection of T1a cancers. A trend toward reduced surgical resection for benign lesions in the United States has only been recently described.6Kruger A.J. et al.Am J Gastroenterol. 2021; 116: 1938-1945Crossref Scopus (0) Google Scholar Many US centers do not have the personnel or expertise to convert to ESD for all LNPCNs, because ESD is inefficient to learn and perform compared with endoscopic mucosal resection (EMR). A selective approach to ESD, focused on high-risk rectal lesions, allows centers to efficiently provide curative endoscopic resection of the large numbers of referred lesions, of which the overwhelming majority are benign. This approach likely produces the greatest overall reduction in unnecessary surgical resections.4.Practical impediments to a policy of universal ESD for all LNPCNs in some countries often have their foundation in reimbursement and cost issues. For example, whereas Japan has a reimbursement policy that pays substantial increments for EMR and particularly ESD relative to routine colonoscopy and polypectomy, US reimbursement for EMR is often considered inadequate, and there is currently no consistent or coherent policy for ESD reimbursement. The relative risks of EMR and ESD also affect costs and thereby choices. Modern EMR has almost no risk of delayed perforation, and immediate perforations and muscle injuries are easily treated during EMR, so that EMR is almost entirely outpatient. The higher perforation risk of ESD, which included 7 delayed perforations in the study by Ohata et al, more often leads to hospitalization for observation, which is expensive in the United States. These factors are additional points to consider beyond those of O’Sullivan et al and further support a selective approach to ESD for LNPCNs. In some settings, selective use of ESD, based in part on lesion morphology and partly on rectal location where the morbidity of surgical resection is highest, remains appropriate, practical, and in the best interest of many patients. ReplyGastroenterologyPreviewWe express our appreciation for the comments from O’Sullivan et al and Rex et al on our recent study. We agree with performing piecemeal endoscopic mucosal resection (p-EMR) in cases with low risk of cancer, such as sessile serrated lesion (SSL) and laterally spreading tumor (LST) homogenous type. Although SSL is not included in the JGES guidelines1 as an indication for endoscopic submucosal dissection (ESD), SSLs > 2 cm are sometimes associated with cancer or dysplasia. Therefore, we finally decided to include SSLs in our study. Full-Text PDF