Low-grade intraepithelial neoplasia is an unequivocal neoplastic condition confined to the glandular epithelium, not to be mistaken for inflammatory or regenerative changes. The earliest morphological precursor of intraepithelial neoplasia is the aberrant crypt focus (ACF), i.e. a “single-crypt adenoma” [16]. Microscopic examination of mucosal sheets dissected from the bowel wall in patients with familial adenomatous polyposis or mucosal examination with a magnifying endoscope reveals ACFs to have crypts of enlarged calibre and thickened epithelium with reduced mucin content [16]. Progression from ACF through adenoma to adenocarcinoma characterises carcinogenesis in the large intestine. The situation is more complex in individuals with chronic inflammatory bowel disease in which the diagnosis of a sporadic adenoma has clinical implications different from those of colitis-associated neoplasia.

Histologically, lesions characterised by low-grade intraepithelial neoplasia show reduced mucin content (reduced number of goblet cells) with increased cytoplasmic basophilia and a variety of nuclear changes, such as hyperchromasia, enlargement (spindle-like elongation) and pseudostratification (palisading). Although the number of mitotic figures is increased, atypical mitoses are generally not observed, and nuclear polarity is retained (Fig. 21.2a, b).

Morphological changes suggestive of high-grade intraepithelial neoplasia should involve more than just one or two crypts. In addition to the histological characteristics of low-grade intraepithelial neoplasia described above, high-grade intraepithelial neoplasia shows distinctive structural features: complex glandular crowding and irregularity (note that the word “complex” is important and excludes simple crowding of regular tubules that might result from crushing), prominent glandular budding, cribriform gland formation, back-to-back pattern as well as prominent intraluminal papillary tufting. Some of these criteria may be found in low-grade intraepithelial neoplasia/dysplasia as well. This makes it necessary to use further cytological features: loss of cell and/or nuclear polarity and distinct nuclear changes such as vesicular and/or markedly enlarged irregular nuclei, often with a dispersed chromatin pattern and prominent nucleoli. Atypical mitotic figures are common, and prominent apoptosis, giving rise to intraluminal debris, may be found in some cases (Fig. 21.2c, d). Often several of the above-mentioned features are detected within the same lesion. Caution should be exercised if only a single criterion is observed in order to avoid overinterpretation. Since ordinary adenomas do not erode, erosions should always prompt thorough evaluation of an adenomatous lesion not to miss presence of high-grade intraepithelial neoplasia. However, overinterpretation of isolated surface changes that may be caused by trauma or prolapse needs to be excluded.

Misplacement (entrapment) of adenomatous epithelium into the submucosa of a polyp (pseudoinvasion) represents a well-known histopathological pitfall which needs to be distinguished from invasive carcinoma [17]. If in doubt, the relevant findings should be stated in the pathology report, and a second opinion and/or additional biopsies from the polypectomy site should be considered. Sigmoid polyps are particularly prone to inflammation, a feature which tends to exaggerate the histological features of neoplasia. When associated with epithelial misplacement, the potential for misdiagnosing these lesions as early carcinoma is evident. Histological changes suggestive of epithelial misplacement are acellular submucosal mucin lakes (not to be misinterpreted for mucinous adenocarcinoma) and hemosiderin-laden macrophages within the stroma.

The danger of surgical overtreatment due to misinterpretation of a lesion as invasive cancer must always be considered in biopsy diagnoses. Postoperative mortality (within 30 days) varies between 5 and 10 % in colonic cancers, depending on the population, age of the patient and quality of services available [18–20].

Achieving the optimum balance between removing all disease by resection and minimising harm is crucial. Since criteria of intramucosal criteria are currently not well defined in columnar epithelium, we strongly recommend to adhere to the definition given by the WHO classification of tumours of the digestive system [3], which regards invasion of neoplastic cells through the muscularis mucosae into the submucosa as prerequisite for diagnosing invasive adenocarcinoma at this site (Fig. 21.3a). It needs, however, to be recognised that diagnoses made on the basis of this definition will hamper comparison with series from Japan for which the diagnosis of intramucosal adenocarcinoma is possible. Of note, the AJCC/UICC TNM classification scheme [5, 6] allows to use the term carcinoma in situ (pTis) in columnar epithelium, but as pointed out above, this category is rather vague and ill defined, and, therefore, its use is discouraged by the European guidelines on colorectal cancer screening [7, 8].

Histological criteria of invasion, such as single tumour cells, are more likely to be seen in advanced carcinomas, but not in early carcinomas. Likewise, a desmoplastic stromal response is only rarely observed, particularly when carcinoma cells have just started to invade the submucosal layer (Fig. 21.3b). In addition, basal membranes are frequently discernible around invading glands in well-differentiated early carcinomas [21–23], and definitions using the term “invasion through the basement membrane” are for this reason misleading. Nevertheless, a subclassification of early carcinomas into low-risk and high-risk categories, based on the estimated risk of regional lymph node involvement, should always be performed.

Histological typing of colorectal cancers is generally performed according to WHO guidelines and does not differ between early and advanced tumours [3]. Thus, most adenocarcinomas are gland forming, with variability in size and configuration of glandular structures. In well- and moderately differentiated cases, which make up the vast majority of early cancers, the epithelial cells are usually large and tall, and the gland lumina may contain varying amounts of debris. The designation of mucinous adenocarcinoma is used when more than 50 % of the lesion is composed by pools of extracellular mucin. Signet ring cell carcinoma and adenosquamous carcinoma represent rare variants that are almost never observed as early cancer.

For submucosal carcinomas, substaging of pT1 is essential. The most widely used approach implies the pragmatic subdivision of the submucosal layer into three parts: a superficial, a middle and a deep third, also referred to as Kikuchi levels sm1, sm2 and sm3 [32]. The risk of lymph node involvement has been calculated to account for 1.4–5 % for sm1, 8 % for sm2 and 23 % for sm3 tumours, respectively [36, 37]. In polypoid lesions, the stalk always represents the upper third of the submucosal layer. Haggitt et al. [33] identified the level of invasion into the stalk of a pedunculated polyp as being important in predicting outcome. The authors classified polypoid tumours as follows: “level 1” lesions show invasion of the submucosa limited to the head of the polyp, “level 2” lesions extend into the neck of the polyp, “level 3” lesions show invasion of any part of the stalk and “level 4” lesions invade beyond the stalk but still confined to the submucosal layer. An example illustrating the use of the Kikuchi and the Haggitt system is given in Fig. 21.4.

It has to be noted that both the Kikuchi (for sessile tumours) and the Haggitt (for polypoid or pedunculated tumours) classification systems may be difficult to apply, especially if there is fragmentation or suboptimal orientation of the tissue. Separation of Haggitt level 2 from levels 1 and 3 may be difficult even if the lesion is well preserved. Moreover, although level 4 invasion is generally accepted as a major risk factor for adverse outcome, regional lymph node metastasis may be observed also in level 2 or 3 lesions.

Thus, recent classification systems stressed the advantage of exact measurement of tumour invasion beyond the muscularis mucosae. Ueno et al. [38] suggested to assess both the depth (cut-off value 2,000 μm) and the width (cut-off value 4,000 μm) of invasion. The authors demonstrated that by doing so, a more objective and accurate assessment of the risk of lymph node metastasis is possible: 3.9 % vs. 17.1 %, when depth of submucosal invasion is lower or higher than 2,000 μm, as well as 2.5 % vs. 18.2 %, when width of submucosal invasion is lower or higher than 4,000 μm. According to other studies, tumours with submucosal invasion not deeper than 1,000 μm generally lack lymph node metastasis, thus rending the 1,000-μm cut-off value as major criterion for stratification of patients for endoscopic or surgical therapy [39, 40].

Each classification system has its advantages and disadvantages: (1) the Kikuchi system cannot reliably be used if the muscularis propria is not present (i.e. in most endoscopic resections), (2) the Haggitt levels are not applicable in non-polypoid (sessile) lesions and (3) exact measurement depends on the feasibility to identify the muscularis mucosae which may be difficult, if not impossible, when cancer tissue has destroyed this histological landmark. Lacking a worldwide consensus, a definitive recommendation for one method or another can currently not be given. The European guidelines of colorectal cancer screening, however, recommend the Kikuchi system for non-polypoid (sessile) and the Haggitt system for polypoid (pedunculated) lesions [5, 6].

Grading of colorectal cancer is generally performed according to the WHO guidelines assessing the extent of glandular appearance and should be divided into well-, moderately and poorly differentiated or into low-grade (encompassing well- and moderately differentiated cancers) and high-grade (including poorly differentiated and undifferentiated cancers) tumours, respectively [3]. Poorly differentiated adenocarcinomas should at least show some gland formation or mucus production. When a carcinoma is heterogeneous, grading should be based on the least differentiated component, not including the leading front of invasion. The percentage of the tumour showing gland formation can be used to define the grade. According to the current WHO classification [3], well-differentiated tumours (grade 1) exhibit glandular structures in >95 % of the tumour area, moderately differentiated tumours (grade 2) in 50–95 % and poorly differentiated tumours (grade 3) in >0–49 %, respectively. By convention, morphological grading of tumours applies only to adenocarcinoma “NOS” (not otherwise specified). Other morphological variants (e.g. mucinous adenocarcinoma, signet ring cell carcinoma) carry their own prognostic significance and grading does not apply [3].

In general, no substantial differences exist between the grading of early and advanced cancers. In the absence of good evidence, the European guidelines for colorectal cancer screening recommend that a grade of poor differentiation should be applied in a polyp cancer when any area of the lesion is considered to show poor tumour differentiation [7]. As for advanced tumours, tumour cell dissociation at the leading edge of invasion (“budding”) should not influence the grading of early cancers.