Usually, endoscopic findings are enough to distinguish colorectal benign lesions from malignant lesions. However, in some cases, endoscopic findings are so ambiguous that it is difficult to discriminate between benign and malignant. Although biopsy and pathologic examination are the only way to confirm the malignancy, the initial endoscopic evaluation may be of a great significance for planning the treatment strategy. Thus, endoscopist should be reminded on the endoscopic features strongly suggesting a malignant component.

According to data from US National Polyp Study , the size of polyp is strongly associated with malignancy potential. If the size of polyp is less than 5 mm in diameter, the prevalence of malignancy is less than 0.1%, that is, rarely malignant [5]. On the other hand, if a polyp is larger than 50 mm in diameter, the probability of malignancy is over 30% [6].

Malignant transformation is more frequent in villous type than in tubular type. And the other gross configurations suggesting malignant transformation are as follows; nodule on polyp (“Buddha-like” polyp), color change or easy-touch bleeding, depression or ulceration, firm consistency or induration, broadening of the stalk or fold irregularity, irregular surface contour and lack of air-induced deformation.

The orifices of colonic mucosal glands are called by pits, and the specific shape and arrangement of pits in each lesion is called by a pit pattern (PP) . The pit pattern can be clearly observed with chromoendoscopy and it reflects the lesion’s histological nature. Kudo et al. has firstly characterized pit patterns [7]. The pit patterns are classified into five types (I–V); PP type I (round pattern), PP type II (asteroid or papillary pattern), PP type III (tubular pattern), PP type IV (branch or gyrus-like pattern), PP type V (unstructured pattern). PP type III can be classified into two sub-types; PP type IIIs (small tubular pattern) and PP type IIIL (large tubular pattern). PP type V also can be classified into PP type Vi (irregular pattern) and PP type Vn (non-structured pattern). Among these PP types, PP type V carries very-high risks of malignant transformation (60–95%) [8] (Fig. 4.1).

Sano, et al. investigated the alterations of the surface capillary patterns, using a enhanced endoscopy called narrow band imaging (NBI) endoscopy [9]. It emphasize the superficial capillary structure without any dye, using the light of specific blue and green wavelengths. The colonic capillary patterns (CP) are classified into four patterns: CP type I (no meshed capillary vessels, faint pattern), CP type II (regularly meshed capillary vessel) and CP type III (irregularly meshed capillary vessels). CP type III lesions are divided into two sub-types, CP type IIIA (lack of uniformity, high density of capillary vessels) and CP type IIIB (loose capillary vessels or nearly avascular). Among these CP types, CP type III show very high incidence of malignant transformation (70–90%) [8].

The concept of the “superficial neoplastic lesion” is very similar to that of early cancer. The superficial neoplastic lesion is the lesion with endoscopic features suggesting the limited invasion depth (no more than into the submucosa). The Japanese group classified the superficial neoplastic lesions as “type 0”, distinguished from types of advanced lesions, i.e., the Borrmann type 1–4 [10, 11].

On 2002, an international group of endoscopist, surgeons, and pathologists participated in the Paris workshop, and explored the utility and clinical relevance of the Japanese endoscopic classification of superficial neoplastic lesions [12]. In the Paris workshop, the participants proposed a general framework for the endoscopic classification of the superficial neoplastic lesion (type 0). According to the Paris classification , type 0 lesion of large bowel is classified as three types basically, based on the absence or the presence of protrusion/ulceration. The basical three types include polypoid (type 0-I), non-polypoid, non-excavated (type 0-II), and non-polypoid, excavated (type 0-III) lesions (Fig. 4.2).

Type 0-I lesions are polypoid lesions, i.e., protruding type. With respect to the presence or the absence of neck, Type 0-I lesions are classified into two sub-types, pedunculated type (type 0-Ip) and sessile type (type 0-Is).

Type 0-II lesions are relatively flat lesions, neither protruded nor excavated. Type 0-II lesions are classified into three sub-types, slightly elevated type (type 0-IIa), flat type (type 0-IIb), and slightly depressed type (type 0-IIc). Among type 0-II lesions, some have both findings of elevation and depression. Type 0-IIc + IIa lesion is a depressed lesion with partially elevated area and type 0-IIa + IIc is an elevated lesion with partially depressed area.

Because both type 0-Is and 0-IIa lesions are all elevated type, sometimes it is difficult to make a distinction between type 0-Is and 0-IIa. The distinction between type 0-Is and type 0-IIa is based on the height of the lesion. When compared with the height of closed biopsy forceps (2.5 mm), type 0-Is lesions are >2.5 mm and type 0-IIa lesions are <2.5 mm. Type 0-III lesions are excavated lesions, i.e., ulcerated type.

Laterally spreading tumor (LST) is defined by lateral growth of lesions larger than 10 mm in diameter with a low vertical axis. Based on the surface morphology, LSTs are usually classified into two types, granular and non-granular type. The granular type LSTs are classified into two sub-types, the hemogenous type and the nodular mixed type. The non-granular type LSTs are classified into two sub-types, the flat elevated type and the pseudodepressed type [See Chap. 3]. Malignant transformation is more frequent in nodular mixed type and pseudodepressed type.

For planning the treatment strategy of the superficial neoplastic lesions (surgery or endoscopic resection), proper evaluation of endoscopic respectability is critical.

Although there are several unfavorable histopathological factors for lymph node metastasis in ECC, endoscopic respectability is absolutely depend on the invasion depth, because only the invasion depth is a predictable histopathologic factor before resection, and also it is technically difficult to completely resect the lesions with deep invasion.

Methods for evaluating the invasion depth of superficial neoplastic lesions include EUS, magnifying endoscopy, and the nonlifting sign.

EUS has a high level of accuracy for T staging, especially differentiation between Tis and T1 colorectal cancer (CRC), showing 90% accuracy in differentiation between Tis and T1 CRC [13, 14].

Magnifying chromoscopy is a reliable method for differentiating between non-neoplastic and neoplastic lesions, or between early cancer and adenoma [14, 15].

Clinically, in evaluation of invasion depth, making a differentiation between limited submucosal invasion and deep submucosal invasion is important. However, neither EUS nor magnifying endoscopy is suitable for making a differentiation between limited and deep submucosal invasion.

Checking for the nonlifting sign (NLS) is a simple and reliable method for making a differentiation between limited and deep submucosal invasion. NLS was first described by Uno et al. in 1994, defined as when the lesion is not lifted by saline solution injection into the submucosal layer of the tumor base [16] (Fig. 4.3). Lesions with deep submucosal invasion are not lifted by a submucosal saline solution injection because of the dense fibrosis associated with invasive carcinoma, which prevents fluid infiltration through the submucosal connective tissue. Some previous studies showed the correlations between deep submucosal invasion and the nonlifting sign. According to previous researches, ECC lesions with deep submucosal invasion show 72.7–100% of nonlifting sign, while lesions with limited submucosal invasions show 0–20% of nonlifting sign [17–19]. Although nonlifting sign is every effective for evaluating invasion depth, mechanical stimulation such as forceps biopsy may cause the false-positive event. Han, et al. reported that forceps biopsies may lead to submucosal fibrosis, causing the nonlifting sign in colorectal tumors although they have not invaded the deep submucosal layer. They also reported that an increase in the number of postbiopsy days (more than 21 days) may influence the nonlifting sign in endoscopically resectable colorectal tumors. They recommended that mechanical stimulation such as forceps biopsies should be minimized before endoscopic resection, and endoscopic resection should be tried as soon as possible if biopsy was performed [17].

Endoscopic resection methods for superficial neoplastic lesions include snaring polypectomy, endoscopic mucosal resection (EMR), and endoscopic submucosal dissection (ESD). In choosing the endoscopic resection method, the size, endoscopic type, and predicted invasion depth of the lesion should be taken into consideration.

Snaring polypectomy should be only used for the pedunculated type. In snaring of pedunculated lesion, the snare should be placed near the base for acquiring enough resection margin.

Since endoscopic mucosal resection (EMR) was first described in 1973 as a strip biopsy technique, EMR is widely used for large polyp resection [20]. Clinically, EMR is the most commonly used method for endoscopic resection of superficial neoplastic lesions. However, in case of the larger sized lesion, it is difficult to acquiring en bloc resection by EMR. Practically, 2 cm may be the largest size that can be easily resected en bloc by EMR.

Endoscopic piecemeal mucosal resection (EPMR) could be used for larger sized lesions. If the superficial neoplastic lesions is suspected to be noninvasive, EPMR could be performed. But, it should be noted that EPMR is associated with a high incomplete resection rate and a high local recurrence rate [21].

As is well known, en bloc resection is desirable for accurate pathologic examinations, especially for evaluating the resection margin. Endoscopic submucosal dissection (ESD) is a new technique for en bloc resection of larger sized tumor. Since ESD was first contrived for gastric tumor resection, colorectal ESD was adopted later and is increasing now. Although colorectal ESD has not become a common method yet because of the difficult technique and high risk of complications, colorectal ESD can be safely performed by experienced experts [21, 22].

After endoscopic resection of superficial neoplastic lesions, the localization of the original lesion site may be necessary for colonoscopic surveillance or additional surgery. If a submucosal invasive cancer is suspected clinically, marking at the site of excision is recommended, using submucosal injection of India ink [23–25].

The endoscopically resected specimens should be pinned to a board by pin immediately after resection, and soaked to 10% neutral buffered formalin to avoid shrinkage, autolysis or other tissue artifacts related to poor fixation (Fig. 4.4).

There has been controversy for the definition of a positive resection margin , and following three definitions are used commonly; (1) tumor cells present <1 mm from the transected margin, (2) tumor cells present <2 mm from the transected margin, and (3) tumor cells present within the diathermy of the transected margin [26–29].

Resection margin is more important in invasive lesions, and the status of vertical (deep) margin is more valuable than of lateral margin. If endoscopically resected T1 CRC shows a positie resection margin, additiona surgery should be considered.

Tis lesion has no risk of lymph node metastasis, so complete endoscopic resection of Tis is accepted as a curative therapy [30, 31].

However endoscopic resection of T1 CRC should be selectively applied because lymph node metastasis occurs in 7–15% [32–36]. Many previous researches reported the risk factors for LNM in T1 CRC, and commonly accepted pathologic risk factors for LNM in T1 CRC are deep submucosal invasion, vascular invasion, histopathologic high grade, and budding.