Dysplastic lesions may be overlooked because those in UC patients are often flat. To find such flat dysplasia, step biopsy is applied to surveillance colonoscopy. A general procedure for step biopsy is that four biopsy specimens are taken in every 10 cm from the cecum to the rectum, requiring at least 32 biopsy specimens in one surveillance colonoscopy. Despite this time-consuming procedure, the detection rate of neoplastic lesions is very low [12]. A recent randomized controlled study demonstrated the detection rates of neoplasia are comparable between the step biopsy and targeted biopsy methods [13]. Although step biopsy is still recommended in most of the guidelines for surveillance in patients with UC, as discussed below, target biopsy will replace step biopsy, along with the advancement of new endoscopic technology such as high-definition endoscopy and chromoendoscopy.

Chromoendoscopy with dyes such as methylene blue or indigo carmine can increase the detection rate of dysplasia in surveillance compared with conventional white-light colonoscopy. Kiesslich et al. demonstrated that chromoendoscopy with methylene blue detected three times more dysplastic lesions than conventional white-light colonoscopy [14]. A meta-analysis of seven studies comparing chromoendoscopy with white-light colonoscopy reported that the relative risk of detection of dysplasia with chromoendoscopy is 1.8 [95% CI, 1.2–2.6] compared with white-light colonoscopy, and chromoendoscopy increases the absolute risk by 6% [95% CI, 3–9%] [11]. Surveillance with chromoendoscopy has an advantage in terms of cost, because it needs fewer biopsy specimens than white-light colonoscopy. Thus, chromoendoscopy with target biopsy becomes the recommended procedure for surveillance in recent guidelines [11]. Step biopsy is recommended only if chromoendoscopy is not available.

The usefulness of narrow-band imaging in the detection of UC-associated CRC/dysplasia is controversial [15]. Other new imaging techniques such as endomicroscopy and confocal endoscopy were also examined but have not yet been implemented in clinical practice.

Pit pattern diagnosis with magnifying endoscopy is implemented in the diagnosis of sporadic CRC/adenoma. This diagnostic method is useful to differentiate neoplastic lesions from non-neoplastic lesions; Types I and II are non-neoplastic lesions and Types III, IV, and V are neoplastic lesions. Data on the usefulness of pit pattern diagnosis in the diagnosis of UC-associated dysplasia is limited. Kiesslich et al. reported that, in UC patients, 30 out of 32 dysplastic lesions (93.8%) showed Types III or IV and 82 (95.3%) out of 86 inflammatory hyperplastic lesions showed Type I or II [14]. In contrast, a Japanese group reported that specificity of the pit pattern diagnosis in UC-associated dysplasia was 100%; however, neoplastic patterns (Type III-V) were also observed in non-dysplastic areas, because inflamed mucosa could show various surface patterns [16], resulting in a low sensitivity of 57%. The diagnostic value of pit pattern diagnosis in UC-associated neoplastic lesions remains undetermined.

The management of dysplasia differs based on whether the lesion is endoscopically visible or not. If the dysplastic lesion is visible with a distinct border, and no dysplasia is found in other areas in colitic mucosa, it can be removed endoscopically. A recent meta-analysis of ten studies involving 376 patients in whom polypoid dysplasia was resected reported that the incidence of CRC was 5.3 cases/1,000 patient-years [95% CI, 2.7–10.1] during follow-up [17]. This risk of CRC can be acceptable for both patients and physicians, considering the risks of colectomy. After confirming complete resection of the lesion, the patient should be on close surveillance.

Detecting HGD in flat mucosa in surveillance colonoscopy, total colectomy must be considered, because the rate of concurrent CRC ranges from 47 to 67% in patients with flat HGD [11]. Colectomy may be considered in patients with flat LGD. It has been reported that advanced neoplasia was found in 18 patients (16%) out of 113 patients with flat LGD after a median follow-up of 48 months [18]. A meta-analysis involving 477 patients with flat LGD showed the rate concurrent CRC was 22% and that of progression to any advanced lesion was 14.6% [19]. When colectomy is avoided, frequent surveillance at the interval of 3–6 months is recommended for patients with flat LGD.

Sporadic adenoma may coincidentally develop in patients with UC. The diagnosis is challenging if the lesion is inside colitic areas, because it is difficult to distinguish UC-associated dysplastic polypoid lesions from sporadic adenoma. UC-associated dysplastic polypoid lesions that resemble endoscopically and histologically sporadic adenoma are referred as adenoma-like DALM. Engelsgjerd et al. followed 24 UC patients with endoscopically resected adenoma-like DALM. Fourteen patients (58%) developed further adenoma-like DALM, but only one patient (4%) developed LGD during the mean observational period of 42.4 months, suggesting a benign clinical course of adenoma-like DALM [20].