In general, dysplasia itself is known to be associated with an increased risk of CRC [24, 31], and often leads to a recommendation for prophylactic colectomy. To understand the endoscopic findings and surveillance strategies one must understand the classifications of dysplasia. Dysplasia is histologically defined as neoplastic alteration of the epithelium without invasion into the lamina propria, also termed intraepithelial neoplasia [31]. Traditionally the microscopic classification of dysplasia is divided into negative for dysplasia, indefinite for dysplasia, and positive for dysplasia [31]. Indefinite for dysplasia refers to changes in the epithelium that cannot be classified definitively as positive or negative for dysplasia either due to the confounding effects of inflammation and regeneration or to technical factors that impede their pathologic interpretation [32]. Dysplasia is further classified into LGD (also referred to as low-grade intraepithelial or noninvasive neoplasia), HGD (also referred to as high-grade intraepithelial or noninvasive neoplasia) or invasive cancer [31, 33]. It is important that histologic diagnoses of dysplasia be confirmed by a second expert gastrointestinal pathologist.

Grossly, most dysplastic lesions are endoscopically visible and their appearance in chronic colitis is heterogeneous, but is usually described as flat or elevated [34]. Flat dysplasia is the same as invisible dysplasia found on random nontargeted biopsies, and elevated lesions can be described as polypoid, plaque-like, slightly raised (flat), unifocal, or multifocal [35]. The growing use of the term “flat colon polyps” in non-colitic colons, which are now better visible through improved endoscopic imaging, can lead to confusion with flat LGD. Endoscopists must be clear whether the dysplasia was macroscopically visible (elevated) or invisible (flat). Raised dysplastic lesions discovered in areas of macroscopic or microscopic inflammation or prior inflammation are also referred to as dysplasia-associated lesion or mass (DALM) [36]. DALMs can then be further characterized into lesions that endoscopically appear to be consistent with sporadic adenomas, called “adenoma-like DALMs” or those that do not, called “nonadenoma-like DALMs.” If a well-circumscribed polyp with dysplasia is found outside the area of inflammation it is commonly referred to as a sporadic adenoma (Fig. 16.1a). If a similar appearing lesion is found within inflammation it is referred to as an adenoma-like DALM (Fig. 16.1b). Nonadenoma-like DALMs found in areas of inflammation typically have irregular and indistinct margins [36] (Fig. 16.1c). This terminology can be confusing since there might not be a difference between a sporadic polyp and an adenoma-like polyp when discovered in colitic mucosa; and when present in inflamed irregular mucosa it can be challenging to distinguish from a nonadenoma-like DALM [37]. These distinctions carry important prognostic implications with different management outcomes summarized in Table 16.2.

Unfortunately, carcinogenesis in IBD may not follow a progressive pattern from low-grade dysplasia (LGD), through high-grade dysplasia (HGD), to cancer [30]. Ullman et al. showed that cancer can arise in patients without prior dysplasia, without progression from LGD to HGD [23]. In fact even the diagnosis of indefinite for dysplasia carries risk of advanced neoplasia in 9 % of patients [5, 38].

The degree of risk for associated or synchronous CRC or HGD in patients with flat LGD is varied with studies reporting rates of 0–28 % [23, 24, 39, 40]. A recent meta-analysis examining 20 studies with 447 cases of flat LGD reported that the positive predictive value of flat LGD was 22 % for synchronous CRC and 36 % for synchronous HGD and CRC [41]. Further raising controversy, two large cohort studies from Europe found only a 2–10 % frequency of advanced pathology over a 10-year period in patients with flat LGD [25, 26]. The debate will continue regarding the degree of risk for synchronous or future advanced neoplasia, but there clearly seems to be a risk. There is a better consensus with the risk of CRC from flat HGD with historical cohort studies and reviews indicating a risk of synchronous or metachronous CRC in up to 42–67 % of patients [3, 22, 24, 42].

There is also agreement on the risk of advanced neoplasia after finding a raised lesion with dysplasia, although it depends on whether it is found in an area of IBD or not (summarized in Table 16.2). Nonadenoma-like DALMs carry a risk for synchronous occurrence of CRC as high as 43–58 % and thus are referred for total colectomy [22, 36]. Adenoma-like DALMs carry a negligible risk of associated CRC similar to that of sporadic polyps [43, 44]. Additionally the risk of CRC does not change if the adenoma-like DALM is inside or outside the area of inflammation [45]. The key in this situation is to make sure the polyp in the area of inflammation is actually an adenoma-like DALM.

Surveillance strategies are aimed at early detection and mortality reduction from CRC in IBD. Evidence for the effectiveness of surveillance programs reducing morbidity and mortality from CRC is not strong due to the limitations of currently available studies, reviewed in Ahmadi et al. [46]. A recent Cochrane systematic review reported that there was no strong evidence that surveillance colonoscopy prolonged survival in people with chronic colitis, but there was evidence that suggested CRC found at an earlier stage led to a better prognosis [47]. However, a recent retrospective study of 149 subjects from the Netherlands identified a significant survival benefit for patients diagnosed with CRC during surveillance colonoscopy, with a 5-year overall mortality of 0 % for people undergoing surveillance versus 36 % for those not undergoing surveillance [48].

Although there is consensus among international gastroenterology societies that persons with chronic ulcerative colitis and chronic Crohn’s colitis undergo surveillance colonoscopy, differences exist on when to start screening, the interval of surveillance, and even the method of surveillance. These differences among American, British, European and Asian societies are summarized in Table 16.3 [49–57].

Current American guidelines do not stratify patients according to risk exclusive of a diagnosis of PSC and disease extent. However, evidence supports that risk stratification to determine surveillance intervals may be a cost-effective approach, and is currently followed by British and European guidelines. European risk stratification is defined as a scoring system, with one point for each risk factor (pancolitis, endoscopic and/or histological inflammation, pseudopolyps, and family history of CRC) with low-risk patient scores of 0–2 and high-risk scores of 3–4 [55]. British risk stratification is defined as: low risk (no endoscopic/histological inflammation or left-sided colitis or Crohn’s disease colitis affecting <50 % surface area of the colon); intermediate risk (mild endoscopic/histological active inflammation or presence of post-inflammatory polyps or family history of CRC in a first-degree relative aged 50 years or over); and high risk (moderate or severe endoscopic/histological active inflammation or stricture within past 5 years or confirmed dysplasia within past 5 years in a patient who declines surgery or PSC/post-orthotopic liver transplant for PSC or family history of CRC in a first-degree relative aged <50 years [53].

Colonoscopic screening and surveillance for CRC in chronic colitis is widely accepted, and even though guidelines evolve, several limitations remain. First, a high sampling error exists, with <1 % of the entire mucosal colonic surface area sampled if 32 biopsies are obtained [30]. Despite growing evidence that most dysplasia in chronic colitis is visible, approximately one-third is flat invisible dysplasia requiring optimal sampling or improved endoscopic techniques [34]. Current guidelines are based off the estimate that 33 biopsies provide a 90 % probability of dysplasia detection [58].

Next, adherence to guidelines by patients and providers are paramount to screening success. A large cohort study from California reported that less than one-third of at-risk patients underwent surveillance colonoscopy in a two-year period. These low rates were also reported by a nested study from the French CESAME cohort that showed a reduced rate (54 %) of surveillance in at-risk subjects with the lowest rates of surveillance in subjects with Crohn’s colitis and those not cared for at specialized centers [59]. Studies have also demonstrated that providers often stray from guidelines and take less than 33 biopsies [60–62].

Making an accurate histologic diagnosis is also vital for proper surveillance outcomes. In addition to the reactive atypia that may confound pathologic interpretation, there is a large degree of interobserver disagreement among pathologists when distinguishing between LGD and HGD [63, 64]. As part of the guidelines it is recommended that all specimens concerning for dysplasia be confirmed by a second gastrointestinal pathologist [65].

Finally, as the guidelines evolve to recommend resection of raised dysplastic lesions as an alternative to colectomy, the adequacy of resection is increasingly important [66]. However, anatomical interference can limit the effectiveness of proper identification and complete resection of lesions. Three circumstances encountered during surveillance colonoscopy need to be addressed further: pseudopolyps, UC-associated strictures, and Crohn’s-associated colonic strictures. In the setting of multiple or diffuse pseudopolyps, identification of suspicious lesions can be a daunting task. A case–control study by Velayos et al. demonstrated that presence of pseudopolyps was associated with a 2.5-fold increased risk in the development of CRC [20]. Strictures in UC should arouse a high suspicion of CRC as studies have shown that a quarter of them are malignant [67]. In the setting of colonic Crohn’s strictures, the risk of underlying malignancy is not as high as in UC, but is still present. A review of 175 Crohn’s colon strictures revealed a 6.8 % frequency of colon cancer after 20 years of disease duration [68]. Therefore surgical resection should be considered if one is unable to fully evaluate the stricture or proximal colon [68, 69].

If flat or invisible LGD dysplasia is found on colonoscopy the management is not clear primarily because of conflicting data regarding the progression of flat LGD to CRC and the controversial existence of concurrent advanced neoplasia. Studies have revealed a LGD to HGD or CRC progression rate ranging anywhere from 2 to 50 % [22–26]. Therefore, it is important to discuss the risks and benefits of both aggressive surveillance (repeat surveillance colonoscopies every 3–6 months) and the option of a proctocolectomy with patients at the time LGD is identified. If the dysplasia is multifocal flat LGD (identified in different locations of the colon during the same colonoscopy) there may be an increased risk of progression to CRC and perhaps more a push toward colectomy. A prospective study from the University of Washington followed 42 patients with LGD and the rate of progression to advanced neoplasia over a 4-year follow-up period was low, but for those patients with four or more biopsies that contained LGD the risk of progression to advanced neoplasia was greatest, RR 5.8 (95 % CI, 1.29–26.04) [40]. The risk of progression to advanced neoplasia in the setting of recurrent flat LGD (also identified on the subsequent surveillance colonoscopy) is again less clear since sampling error might create a false negative. Thus for multifocal or recurrent flat LGD the Crohn’s and Colitis Foundation of America consensus strongly recommends a prophylactic total proctocolectomy [69].

Management recommendations for raised or visible lesions that contain dysplasia on histologic examination (DALMs) depend on their endoscopic appearance and degree of removal. Raised lesions that resemble sporadic adenomas (adenoma-like DALM) should be completely resected with snare polypectomy technique with biopsies obtained from the adjacent flat mucosa surrounding the polyp base, and placed in a separate container, demonstrated in Video 16.2. If the flat adjacent mucosa contains dysplasia, then these lesions should be treated as nonadenoma-like DALMs and colectomy should be recommended. Otherwise, if there is no dysplasia in the adjacent or distant tissue, a repeat colonoscopy should be performed within 6 months [49]. Then, if no further dysplasia is identified, surveillance can be resumed every 1–2 years. A long-term follow-up study of 34 patients revealed no significant difference in future polyp formation between UC patients with an adenoma-like DALM (62.5 %), UC patients with a sporadic adenoma (50 %), and non-UC sporadic adenomas (49 %) [85]. Two other retrospective cohort studies revealed a 0–2 % risk of developing advanced neoplasia in follow-up after polypectomy [37, 86]. The endoscopic management of adenoma-like DALMs is appropriate and can be safely monitored by surveillance colonoscopy. On the other hand, if the lesion is recognized as a nonadenoma-like DALM, there is a significant risk of concurrent or metachronous CRC (described previously), thus a colectomy should be recommended [37, 85].