Surveillance and Endoscopic Resection of Ulcerative Colitis-Associated Neoplasia: A Japanese Perspective

Ulcerative colitis (UC) is a chronic inflammatory disease that affects the colon and has a relapsing and remitting course, and its incidence is rising worldwide [1]. Patients with a long history of UC are at risk of developing a serious complication known as UC-associated neoplasia (UCAN). Since the first report by Crohn and Rosenberg in 1925, a large number of cases of UCAN have accumulated [2]. The characteristics of UCAN are different from those of usual sporadic tumors because UCAN occurs primarily via the p53-mediated oncogenic pathway [3]. As the treatment strategy for UCAN greatly differs from that for sporadic tumors, UCAN needs to be distinguished from sporadic tumors.

The cumulative incidence of UCAN associated with long-term UC has been reported to reach approximately 20% after 30 years [4], although the reported rates vary. To reduce the risk of UCAN and associated mortality, the current guidelines recommend initiating surveillance colonoscopy (SC) 8–10 years after the confirmation of the UC diagnosis [5, 6].

The standard treatment for UCAN is proctocolectomy [7]. However, since the publication of the SCENIC international consensus statement [8], advanced endoscopic resection techniques have been more widely applied to manage dysplasia and early neoplasia in patients with UC. Several reports have highlighted the usefulness of endoscopic submucosal dissection (ESD), which was developed in Japan and can even remove neoplastic lesions with submucosal fibrosis arising within the affected colonic mucosa of UC [9]. This article provides an overview of the current status and future challenges regarding the SC and ESD of neoplastic lesions in patients with UC from a Japanese perspective.

The progression from adenoma to carcinoma that occurs during the development of sporadic colorectal tumors appears to follow an inflammation-dysplasia-carcinoma sequence in UCAN [10]. The sequence for UCAN differs from the adenoma-carcinoma sequence in sporadic colorectal cancer (CRC) [11]. If dysplasia is found on biopsy, cancer may already be present in other parts of the body. And there is a risk of cancer progression in a short period of time when biopsy shows dysplasia. Thus, the early detection, accurate diagnosis, and appropriate treatment of dysplasia are important with CS. A meta-analysis reported that SC results in a 42% reduction in UCAN development and 64% reduction in CRC-associated deaths compared with patients who do not undergo surveillance monitoring [12].

The current guidelines for the timing of SC are primarily based on the disease extent and duration, which are demonstrated risk factors for UCAN. Many guidelines suggest that all patients with inflammatory bowel disease (IBD) should undergo screening CS 8 years after the onset of the first symptoms to assess the disease extent and exclude dysplasia [13].

The European Crohn’s and Colitis Organisation (ECCO) recommends stratifying the timing of subsequent SC after the initial screening in accordance with the patient’s risk level. For patients with colonic disease extending beyond the rectum, the interval of SC is annual, every 2–3 years, or every 5 years depending on the presence of high-risk, intermediate-risk, or low-risk features, respectively [13]. High-risk groups include patients with active total colitis, a family history of CRC in those under 50 years of age, primary sclerosing cholangitis, or a history of dysplasia detection within 5 years. Intermediate-risk groups include patients with a family history of CRC in those over 50 years of age, persistence of mild endoscopic and histologic inflammation. Low-risk groups include patients with total colitis or left-sided colitis without endoscopic or histologic activity, no dysplasia in the past two consecutive SCs. The next SC should be scheduled for 1 year in patients with high-risk features, 2–3 years in patients with intermediate-risk factors, and 5 years in patients with neither intermediate- nor high-risk features.

Japanese guidelines also recommend SC 8 years after the diagnosis in patients with total colitis and left-sided colitis [14]. In contrast to the ECCO guidelines that recommend CS intervals based on the patient’s risk level, the Japanese guidelines recommend SC annually or biennially after screening CS for patients with total colitis or left-sided colitis due to a lack of clear evidence for the optimal timing of SC. It is challenging to determine the appropriate SC interval considering the risk of disease onset and the speed of cancer progression.

The early detection of UCAN is difficult due to the complex background, variable morphologic types, and unclear lesion boundaries with an inflammatory mucosa [15]. As the presence of inflammation makes it difficult to detect lesions, SC should preferentially be performed during the remission state. Furthermore, the lesion boundaries are often unclear, which can make it difficult to detect UCAN lesions with white-light imaging (WLI). Therefore, chromoendoscopy (CE) and image-enhanced endoscopy are required to improve the detection rate of UCAN.

It has traditionally been difficult to use standard WLI to detect dysplasia in UC, which leads to high rates of cancers occurring between scheduled surveillance visits. The SCENIC international consensus statement recommends CE using indigo carmine dye over WLI for surveillance. A meta-analysis found that the dysplasia detection rate is 1.6 times higher with CE compared with high-definition WLI [16]. Furthermore, a recent long-term study revealed an increased incidence rate of dysplasia [17], likely attributable to improved endoscopic imaging and the recent use of CE. CE-based targeted biopsy using high-definition CS requires fewer biopsies than WLI-based random biopsy using high-definition CS and has no disadvantage regarding the UCAN detection rate [16, 18, 19]. However, CE requires the spraying of indigo carmine or methylene blue throughout the entire colon, requires expertise in its use, and can be time-consuming compared with WLI.

The Paris classification system included in the SCENIC international consensus statement is used to describe the morphology of superficial gastrointestinal tract neoplasms. Detected tumors are endoscopically classified into the following five categories: pedunculated, sessile, superficial elevated, flat, and depressed lesions. However, there are many UCAN lesions that do not fit this classification [15]. UCAN differs from sporadic CRC in that UCAN rarely presents as pedunculated lesions and often presents as flat and serrated lesions. In addition, the tumor size and border clarity must be examined to enable complete resection using targeted CE.

Virtual CE (VCE) refers to electronic endoscopic imaging technologies that provide detailed contrast enhancement of the mucosal surface and blood vessels in the colon and rectum. VCE with targeted biopsies is considered an alternative modality to increase the detection of dysplasia in patients with UC. Recent studies have reported no difference in the detection rate of dysplasia between CE and VCE [20‒23]. Furthermore, a recent meta-analysis of the current literature comparing narrow-band imaging (NBI), blue-laser imaging, iSCAN, CE, and high-definition WLI reported that VCE is equivalent to CE in detecting dysplasia and is more cost-effective than CE [17]. The updated SCENIC international consensus statement recommends surveillance using VCE in addition to CE [19].

In selecting the therapeutic strategy, UCAN should critically be distinguished from sporadic tumors in patients with UC because sporadic adenoma can be managed by endoscopic resection [24]. However, it is difficult to distinguish between the two lesion types, even with the use of CS and histological examination of the resected specimens. This is because there is a small amount of non-neoplastic inflammatory and regenerative epithelium with atypia in UC. It has been proposed that the detection of dysplasia in the flat mucosa around a dysplastic lesion is suggestive of UCAN. Additionally, a history of strong inflammatory findings in the background mucosa supports the diagnosis of UCAN.

The endoscopic diagnosis of UCAN is occasionally challenging and requires a stepwise approach using multiple endoscopic modalities, including CE, VCE, and the magnification function. The lesion is observed using indigo carmine to determine the characteristics, non-magnification to determine the extent and irregularities, and a combination of VCE and magnification with crystal violet dye to narrow down the diagnosis.

The endoscopic features of UCAN are shown in Table 1. UCAN differs from sporadic neoplasia in that it often has a flattened and chorioallantoic surface structure with indistinct margins. Furthermore, flat and depressed high-grade dysplasia (HGD) is often erythematous [13]. The Japan NBI Expert Team (JNET) published an NBI-magnifying endoscopy-based classification system for predicting the histology of colorectal tumors in evaluating UCAN [17]. However, the diagnostic performance and intra-observer reliability of the JNET classification and the pit pattern classification for UCAN are low, even among experts [25]. The applicability of the current differentiation systems may be limited by the variety of UCAN lesions. In addition to the endoscopic findings, it is currently recommended to consider the clinical information and histopathologic biopsy findings (including immunostaining for Ki-67 and p53) to obtain a comprehensive diagnosis [21]. However, the performance of preoperative biopsy-based diagnosis is not high in distinguishing UCAN from sporadic tumors and in diagnosing the grade of dysplasia of UCAN, even using immunostaining [26, 27]. Therefore, endoscopic resection is being used for complete biopsy to differentiate between UCAN and sporadic tumors, or between HGD and low-grade dysplasia (LGD).

A convenient paradigm proposes that colon carcinogenesis in IBD follows a progression from no dysplasia to indefinite dysplasia, LGD, HGD, and finally invasive cancer. However, although this model is conceptually useful, it is by no means absolute.

There is a worldwide consensus that a diagnosis of HGD or higher is an indication for proctocolectomy. In contrast, the management of LGD remains controversial. The SCENIC consensus statement recommends complete removal of “endoscopically resectable” LGD, followed by monitoring. The term “endoscopically resectable” is defined as a lesion that (1) has distinct margins, (2) is completely removable by endoscopic resection, (3) is completely removable histologically, and (4) has no dysplasia detected in biopsies of the mucosa adjacent to the resection site [5].

The latest ECCO Guideline/Consensus Paper statement supports the endoscopic resection of visible LGD when the dysplastic lesion has clearly demarcated borders, especially if the lesion is ≤2 cm in diameter, and there is no invasive cancer or submucosal fibrosis; however, large, flat, non-polypoid lesions may require ESD by an expert endoscopist or surgery [13]. Although endoscopic mucosal resection (EMR) is one of the options for endoscopic resection, UCAN is often associated with fibrosis, and complete resection is often difficult with EMR.

This guideline also recommends complete removal of “endoscopically visible dysplasia,” followed by monitoring. From the above, this review focuses on ESD as a method of resection for UCAN. The removal of “endoscopically resectable” lesions is a simple strategy for lesions in patients with UC and has the possibility of expanding the indications for endoscopic resection. However, the definition of “endoscopically resectable” is ambiguous, no strong evidence supports this strategy, and the role of ESD generally remains unclear. In patients with UCAN (dysplasia), the best treatment option (i.e., advanced endoscopic resection or surgery) should be considered based on the disease-related risk factors for cancer and the patient’s age, comorbidities, and perspectives.

Western guidelines propose that the indications for proctocolectomy are (1) if the lesion is not resectable, (2) if there is evidence of dysplasia at the base of the lesion after resection, or (3) if there is endoscopically invisible HGD or multifocal LGD. The criterion “the lesion is not resectable” is based on endoscopic resection, but we believe that this criterion should be revised with the introduction of ESD to facilitate en bloc resection of HGD and LGD because it is difficult to preoperatively differentiate between HGD and LGD, even when using VCE with magnification.

A recent Japanese multicenter cohort study of patients with UC who initially underwent endoscopic resection (n = 199) or surgery (n = 137) for UCAN and sporadic lesions reported that the age of onset was significantly lower in the surgical cases [9]. Furthermore, endoscopically treated patients were more likely to have relapsing-remitting-type disease, while surgically treated patients were more likely to have pancolitis [9]. Endoscopic activity was significantly lower in the endoscopically treated group than the surgically treated group [9]. Although it may not be appropriate to make comparisons due to the different background stages, there was a significant difference between the endoscopically and surgically treated groups in the survival curve at a median follow-up of 34 months, and there were no deaths in the endoscopically treated group [9]. The endoscopic treatment comprised EMR for 142 lesions and ESD for 96 lesions [9]. Sporadic adenoma was the most common lesion type in the EMR group, while UCAN was the most common in the ESD group [9]. In addition, most of the lesions had distinct borders in the ESD group because they were treated endoscopically [9]. The lesion size was significantly larger in the ESD group than in the EMR group, and the left-sided colon and flat type were the most common location and morphology, respectively [9]. The en bloc resection rate was 89% for EMR and 97.9% for ESD [9]. The perforation rate was significantly higher in the ESD group (6.3%) than in the EMR group [9]. Although the perforation rate seems to be higher than that reported for conventional ESD, most patients could be managed using conservative therapy [9]. Consequently, this perforation rate seems acceptable.

The latest systematic review and meta-analysis of the short-term efficacy and safety of ESD for dysplasia in UC reported that the en bloc resection rate and complete resection rate were 94% and 84%, respectively [28]. The short-term outcomes were better in the systematic review and meta-analysis [28] than in the cohort study [9]; however, the heterogeneity was strong regarding the complete resection rate in the meta-analysis. It is possible that the preoperative diagnostic accuracy in detecting the lesion borders differed between endoscopists or institutions.

Metachronous lesions must be monitored to clarify the usefulness of ESD. A Japanese multicenter study that performed surveillance endoscopy in 146 patients for a median follow-up period of 15 months reported a local recurrence rate of 2.7% and a metachronous recurrence rate of 6.1%, with no difference between EMR and ESD and no deaths after endoscopic treatment [9]. Therefore, this study concluded that endoscopic resection is a feasible and valid approach for neoplastic lesions in selected patients with UC [9]; however, these procedures were performed by experts and the endoscopic treatment of UCAN is not easy.

A meta-analysis reported that the incidence of metachronous tumors is 6% [28]; therefore, although ESD has a high complete resection rate, it is not sufficient for dysplasia in UC due to metachronous recurrence. Medications that target inflammation are crucial to prevent the occurrence of dysplasia. This meta-analysis also revealed that the incidence of additional surgery after ESD was 10% [28], which is much higher compared with that reported for sporadic tumors (0.4–1.1%). Furthermore, additional surgery after ESD was performed for patients with metachronous tumors (57%), noncurative resection (39%), and failed ESD (4%) [28]. Among the metachronous tumors, 13 were treated by colectomy, and 3 were treated by ESD. Therefore, ESD of dysplasia in patients with UC is not a definitive cure, and colectomy or another ESD may be needed.

The systematic review results suggest that ESD is a safe and effective treatment for dysplasia in patients with UC [28]. The local recurrence rate was 5%, and the rates of metachronous tumors and additional surgery after ESD were 6% and 10%, respectively [28]. However, all the abovementioned evidence comes from observational studies; therefore, future, randomized controlled multicenter studies with less heterogeneity and longer follow-up are needed to better assess the clinical outcomes of ESD in patients with UC.

The process of ESD for UC is shown in Figure 1. The roles of ESD for UCAN are to treat LGD and to enable the histopathological examination of complete excisional biopsy specimens to differentiate between UCAN and sporadic tumors and grade the dysplasia of UCAN.

Close endoscopic surveillance should be performed after complete endoscopic resection; however, the ideal timing of subsequent procedures is unclear. The local recurrence rate after achieving en bloc resection is very close to zero, and thus, there is no need to fixate on local recurrence. High-quality SC is necessary because of the risk of metachronous lesions. The Global Interventional IBD Group recommends colonoscopy with CE and biopsies at the resection site 3 months after endoscopic resection [29], while a Japanese expert consensus meeting recommends follow-up SC 6 to 12 months after ESD [30]. A longer SC interval is considered acceptable after ESD because of its higher curative rate. Prospective cohort studies are needed to better assess the clinical outcomes of ESD and determine a suitable surveillance interval for UCAN.

We provided an overview of the current status and future challenges regarding the SC and ESD of neoplastic lesions in patients with UC from a Japanese perspective. In future, prospective cohort studies are needed to better assess the clinical outcomes of ESD in patients with UC.

We thank Dr. Kelly Zammit, BSc, BVSc, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Yu Hashimoto, Syota Tomaru, Yuki Itoi, Keigo Satou, Hiroko Hosaka, Hirohito Tanaka, Shiko Kuribayashi, Youji Takeuchi, and Toshio Uraoka declared no conflicts of interest.

Any research was not funded related to the study.

Y.H. is responsible for drafting of the manuscript. S.T., Y.I., K.S., H.H., H.T., S.K., and Y.T. are responsible for revision of the manuscript. T.U. is responsible for critical revision of the manuscript.