This article is mentioned in:

Introduction

Due to their generally good prognosis, small (≤1 cm) and well-differentiated neuroendocrine neoplasms of the stomach, duodenum, appendix, or rectum may be considered as ‘early’ tumors. The new 2010 World Health Organization (WHO) classification refers to these neoplasms as neuroendocrine tumors (NETs)/carcinoids, grade (G)1 or 2, whereas poorly differentiated neuroendocrine carcinomas are classified as G3 (1,2). Over the past 35 years, the age-adjusted incidence of NETs in the USA has risen by ~700%. Colorectal carcinoids display different biological behaviors compared with other tumors (3,4). Colorectal carcinoids are defined by the WHO classification as benign when confined within the submucosa, are ≤20 mm in size, and are not associated with vascular invasion. However, several reports are critical of this definition (5–7). Histological lymph node involvement in G1-G2 differentiated rectal NETs/carcinoids 1–2 cm in size has not been extensively investigated and, thus, its clinical significance is not well known worldwide.

Patients and methods

Patients

Between January 2005 and December 2013, 40 consecutive patients undergoing endoscopic resection of rectal NETs and 12 patients undergoing surgical resection of rectal NETs were enrolled in this retrospective study. The criterion for the performance of endoscopic resection [including endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD)] at the Gifu University Hospital is a tumor size of <20 mm. In addition, the criteria for surgical resection are a tumor size of ≥20 mm, a positive resection margin, and lymphatic or venous invasion in patients who receive local treatment.

All the patients were examined prior to surgery by pelvic computed tomography (CT) at a slice thickness of 5 mm. A diagnosis of lymph node metastasis was made if lymph nodes were detected on this scan. The resected tissue specimens underwent conventional processing and staining with hematoxylin and eosin to evaluate invasion depth, lymphatic and venous invasion, and lymph node metastasis. Clinical and surgical data were correlated with pathological findings. Patient follow-up included clinical assessment, assessment of laboratory data, chest radiography, abdominal ultrasonography, and CT or magnetic resonance imaging for early detection of tumor recurrence, at least every 6 months until 5 years. Survival time was calculated from the endoscopic resection to the time of the last follow-up examination or the patient's death.

Written informed consent was obtained from all the patients enrolled in the present study. The study protocol conformed to the the ethical guidelines of the 1975 Declaration of Helsinki after the approval of the Institutional Review Board of the Gifu University Graduate School of Medicine.

Immunohistochemical staining for Ki-67

An LSAB2 kit (LSAB2 System-HRP; Dako, Carpinteria, CA, USA) was used for Ki-67 immunohistochemical staining. The 4-µm tissue sections were placed on slides and then deparaffinized and dehydrated. The sections were then placed in 0.01 M citrate buffer (pH 6.0) and heated in a microwave (MI-77; Azumaya, Tokyo, Japan) for 40 min at 400 W and at 95°C for antigen retrieval. Following pretreatment at room temperature with 0.3% H2O2 in methanol to quench endogenous peroxidase activity, the sections were blocked with Protein Block Serum-Free (Dako, USA) for 30 min and incubated for 1 h with anti-Ki-67 antibody (dilution, 1:50; clone MIB-1; cat. no. M7240; Dako, Carpinteria, CA, USA). The sections were then incubated for 15 min with biotinylated secondary antibody (pooled goat anti-mouse/anti-rabbit antibody, 1:1, LSAB2 System-HRP, Dako, USA) and washed with phosphate-buffered saline prior to a 20-min treatment with peroxidase-conjugated streptavidin. The sections were visualized by incubation for 3 min in 3,3′-diaminobenzidine tetrahydrochloride with 0.05% H2O2 (Liquid DAB+Substrate Chromogen System; Dako, USA) and counterstaining with Carazzi's hematoxylin.

Statistical analysis

The data are presented as means ± standard deviation and were evaluated with the Student's t-test, Wilcoxon's signed-rank test, Kaplan-Meier method, a log-rank test, and the Pearson product-moment correlation coefficient, as appropriate. A P-value of <0.05 was considered to indicate statistically significant differences.

Results

Patient characteristics

The study comprised 26 men and 14 women, with a median age of 59.3 years (range, 34–81 years). The median tumor size was 7.4 mm (range, 3–15 mm). The locations of the NETs were the rectosigmoid colon (n=3), the upper rectum (n=13), and the lower rectum (n=25) (Table I).

Table I. TNM classification for NETs of the colon and rectum.

Table I.

TNM classification for NETs of the colon and rectum.

TNM	Characteristics
Primary tumor
Tx	Primary tumor cannot be assessed
T0	No evidence of primary tumor
T1	Tumor invades mucosa or submucosa
	T1a size <1 cm
	T1b size 1–2 cm
T2	Tumor invades muscularis propria or size >2 cm
T3	Tumor invades subserosa/pericolic/perirectal fat
T4	Tumor directly invades other organs/structures and/or perforates the visceral peritoneum
Regional LNs
Nx	Regional LN status cannot be assessed
N0	No regional LN metastasis
N1	Regional LN metastasis
Distant metastasis
Mx	Distant metastasis cannot be assessed
M0	No distant metastasis
M1	Distant metastasis

[i] NET, neuroendocrine tumor; TNM, tumor-node-metastasis; LN, lymph node.

NET classification and outcome

The NETs were classified by size as 0–5 (n=7), 6–10 (n=29) and 11–15 mm (n=4). The surgical procedures performed included low anterior resection (plus esophagectomy) in 1 patient, laparoscopic low anterior resection in 7 patients, and laparoscopic intersphincteric resection in 4 patients. Lymph node metastases were present in only 1 patient (tumor size, 6–10 mm, with associated lymphovascular invasion). NET recurrence was not detected in any of the patients. According to the WHO guidelines, the tumors were classified as G1 (n=8), G2 (n=3), and as G1/G2 (n=1) (1). The tumor in the patient with lymph node metastasis was classified as G1 (Table II). No patient received adjuvant chemotherapy. At the time of writing of this manuscript, with the exception of 1 patient who succumbed to heart disease, all the patients remained alive and recurrence-free.

Table II. Grading of gastrointestinal neuroendocrine neoplasms according to proliferative activity (WHO 2010 classfication).

Table II.

Grading of gastrointestinal neuroendocrine neoplasms according to proliferative activity (WHO 2010 classfication).

Grade	Mitotic count (/10 HPFs)	Ki-67 (%)
G1	<2	≤2
G2	2–20	3–20
G3 (NEC)	<20	<20
Mixed adenoneuroendocrine carcinoma

[i] Modified according to reference, MIB1 and antibody, % of 100 tumor cells in areas of highest nuclear labeling. WHO, World Health Organization; HPF, high-power field; NEC, neuroendocrine carcinoma.

Discussion

Historically, the term ‘carcinoid’ has been used to label NETs originating in the gut. There is an international consensus towards the use of the term ‘NET’ rather than ‘carcinoid’, although the latter term is still widely used clinically and in the literature (8,9). A NET is a well-differentiated, neuroendocrine neoplasm comprised of cells with characteristics similar to those of normal gut endocrine cells. According to its location, NETs express general markers of neuroendocrine differentiation (generally diffuse and intense chromogranin A and synaptophysin expression) and hormones (generally intense but not always diffuse) and exhibit mild-to-moderate nuclear atypia with few mitoses [<20 per 10 high-power fields (HPF)]. G1 and G2 are defined on the basis of proliferation fraction and histology (Table III). This definition also covers neoplasms termed ‘carcinoid tumor’ in the WHO 2000 classification (8). In the WHO 2010 classification, these tumors are generically referred to as NETs of the well-differentiated type and rough division and neuroendocrine carcinoma in poorly differentiated neuroendocrine cancer (3,4) that have the characteristics and expression pattern of the endocrine system and gut tumors. NETs are identified as neuroendocrine neoplasms based on the number of mitotic figures showing the proliferative capacity, and by the Ki-67 index as G1 and G2 tumors (3,4) (Table IV). Ito et al reported that it is important to understand the background of the patients, particularly their epidemiological background, and to be aware of the differences in epidemiology between Japanese and Western patients (10).

Table III. Patient characteristics (n=40).

Table III.

Patient characteristics (n=40).

Characteristics	No.
Age at treatment, years
Median	59.8
Range	34–81
Sex
Male	26
Female	14
Tumor size, mm
Median	7.4
Range	3–15
Tumor location
Rs	3
Ra	13
Rb	25
Categories by size, mm
0–5	7
6–10	29
11–15	4

[i] Rs, rectosigmoid; Ra, upper rectum; Rb, lower rectum.

Table IV. Surgical parameters (n=12).

Table IV.

Surgical parameters (n=12).

Variables	Mean	Range
Age, years	55.9	(40–69)
Submucosal, µm	1,816.7	(100–4,500)
Max tumor size, mm	8.5	(5–14)
Operative time, min	458	(275–643)
Blood loss, ml	423	(5–1,770)
Lymph node dissection, no. site	8.6	(2–19)

Soga et al calculated the rate of 1,914 gastrointestinal (GI) submucosal (SM) carcinoids (among a total of 6,799 GI carcinoids) in 10 growth sites and ranked them in descending order from the highest (44.4% in the rectum; n=849) to the lowest (0% in the gallbladder). Despite an overall ratio of GI carcinoids to all GI carcinoids of 28.2% (1,914/6,799), the incidence rate of SM carcinoids at each of the 10 sites varied widely, from 0% (gallbladder) to 51.0% (rectum), with 37.8% in the duodenum, 36.0% in the stomach, 27.1% in the esophagus, and 11.3% in the jejunoileum (11–13).

Identification of the site distribution of colorectal carcinoids among 90,057 cases of tumors registered between 1984 and 1998 in Japan showed 345 cases of carcinoids, all from the Asian population. After excluding 2 cases of unknown sites, the sites were distributed as follows: 3 (0.9%) in the ileum, 8 (2.3%) in the appendix, 28 (8.2%) in the colon, and 304 (88.6%) in the rectum, indicating a high prevalence for carcinoids in the rectum in the Japanese population. Colon carcinoids were most often present in the cecum (9/28, 32%), sigmoid colon (8/28, 28%), transverse colon (5/28, 18%), ascending colon (4/28, 14%), and descending colon (2/28, 7%). The majority of the rectal carcinoids occurred in the lower rectum (267/304, 88%) (14,15).

Fahy et al also reported on the effect of lymphovascular invasion in rectal NETs by investigating 70 rectal NETs surgically resected at a single institution to assess the association between various clinicopathological variables and poor oncological outcomes. The authors observed a strong association of the presence of lymphovascular invasion with metastasis and poor relapse-free and disease-specific survival (16,17). When selecting the appropriate treatment, it is often particularly difficult to evaluate NETs sized <20 mm in the clinical setting. Other clinicopathological variables are not addressed in the current guidelines and staging systems.

At present, the surgical techniques of EMR and ESD are spreading in Japan, and several endoscopic therapies are being performed for rectal NETs. The malignant potential of NETs sized ≥2 cm is high, similar to that of colorectal cancer. The possibility of lymph node metastasis in NETs of this size is also high. However, the management of G1 NETs sized 1–2 cm continues to be a subject of debate (18–20). Unfortunately, controlled prospective studies comparing the endoscopic and surgical approach for NETs of this size have not been performed. Due to the particular biology of G1 NETs of this size, the endoscopic approach is preferable to open surgery in patients with significant comorbidities and in elderly patients at high surgical risk.

Evaluation of the grade of malignancy and the choice of an appropriate treatment for colorectal NETs may be difficult in the clinical setting. Even tumor size and invasion depth are apparently not adequate for stratifying the risk of this rare tumor, and the current guidelines and staging systems do not address other clinicopathological variables. However, several recent studies have reported the effect of lymphovascular invasion on outcome in patients with colorectal NET. The presence of lymphovascular invasion is one of the strongest risk factors for metastasis, as are tumor size and depth of invasion. Furthermore, tumors sized <1 cm with submucosal invasion and no lymphovascular invasion are at minimal risk for metastasis, as reported in studies from Japan and the USA showing 100% 5-year survival (5,6).

This suggests that such tumors may be treated curatively with endoscopic resection or transanal local excision. By contrast, the risk for metastasis of colorectal carcinoids with lymphovascular invasion or a tumor size of >1 cm is equivalent to that of adenocarcinomas. Thus, it is important to emphasize the mandatory histological examination of lymphovascular invasion in specimens obtained by endoscopic resection or transanal local excision, in order to determine the need for additional radical surgery with regional lymph node dissection. Although the effects of lymphovascular invasion are not covered by the present guidelines and the TNM staging system, this may be promising for improving the guidelines and staging systems, particularly for early-stage colorectal NETs (7).

The present study evaluated rectal NETs sized <20 mm in comparison with the WHO 2010 guidelines. An open surgical procedure was performed in 12 patients to evaluate the presence of lymphatic or venous invasion. According to the WHO guidelines, 8 of our patients had G1 NETs, with 1 patient having lymph node metastasis; 4 patients had G2 NETs, and 1 patient had G1/G2 NET. ESD and EMR were performed in 28 patients, all of whom had G1 NETs (data not shown).

According to their mitotic index or Ki-67 index, NETs may be subdivided into either G1 or G2. Although this revised classification is a simple and useful grading system based on proliferative activity, the assessment of tumors with a Ki-67 index of >2 and ≤3% remains unclear. Despite this, inter-observer differences in mitotic counts remain larger compared with those of the Ki-67 index, and it is difficult to routinely scan at least 50 HPFs (1 HPF=2 mm2), as required by the WHO 2010 classification for evaluation of the mitotic index. Thus, the validity and reproducibility of the Ki-67 index are superior to those of the mitotic index (21). A tumor with a a Ki-67 index of <2% is classified as G1, whereas a tumor with an index of 3–20% is classified as G2. As only one cut-off value was used to divide continuous values into two groups, the Ki67 index criteria of G1 NETs of the WHO 2010 classification were validated to elucidate the assessment of tumors with a Ki-67 index of 2–3%.

Yamaguchi et al reported that analysis of a Ki-67 index between 2 and 3% confirmed 2.8% to be the best Ki-67 index cut-off value for predicting metastasis or recurrence. However, none of our patients with G1/G2 and G2 NETs had lymph node metastasis (22). Only 1 of 12 (8.3%) patients had lymphatic or venous invasion, which is a low rate compared with that of other reports.

It remains debatable whether rectal NETs sized 10–19 mm require radical lymph node dissection or can be treated with local resection. As reported by Mani et al, the lymphatic spread of NETs sized 10–19 mm was clear in 10–15% of their patients. Guidelines from UKNET also recommend that lesions sized <10 mm may be adequately treated with complete endoscopic removal (23).

However, Konishi et al reported a 7% incidence of lymph node metastasis in tumors sized ≤10 mm, and Soga et al reported a 9.7% rate of metastasis of early-stage rectal carcinoids sized ≤10 mm. Fahy et al indicated that invasion depth, lymphovascular invasion and mitotic rate, all correlate with prognosis. Konishi et al found the risk factors for lymph node metastasis of colorectal carcinoids to be tumor size ≥11 mm and lymphatic invasion. Our patient with lymph node metastasis developed lymphatic invasion of a 10-mm tumor. In their evaluation of specimens from endoscopic resection, Fijimoto et al reported that 70% of the patients had lymph node metastasis with tumor sizes of 0–30 mm. The incidence of lymph node metastasis in the present study is significantly higher compared with that in previous reports, suggesting that the Fujimoto's criteria for radical resection are appropriate (24). Of note, two of three lesions sized <10 mm with lymph node metastasis were accompanied by lymphovascular invasion, confirming that lymphovascular invasion may be an important predictor of lymph node metastasis. The absence of lymphovascular invasion is key to confirming a favorable outcome in patients with a colorectal NET (Table V). To the best of our knowledge, this is the first study to report a comparison between lymphovascular invasion and Ki-67 in regard to rectal NETs. Further evaluation should be undertaken to determine whether mandatory histological examination of lymphovascular invasion is necessary in specimens obtained by endoscopic resection or transanal local excision, as histological examination can provide useful information to aid in determining the necessity of additional radical surgery with regional lymph node dissection.

Table V. Surgical parameters (n=12) mitotic count and Ki-67.

Table V.

Surgical parameters (n=12) mitotic count and Ki-67.

No.	Depth	SM invasion, µm	Size, mm	Lymphatic invasion	Vascular invasion	Mitotic count, na	Ki-67, %a	WHO 2010
1	SM	2,000	7	0	1	1	0.4	G1
2	SM	100	8	1	0	0	1.0	G1
3	SM	1,000	8	0	0	1	2.4	G2
4	SM	1,000	8	0	1	1	3.0	G2
5	SM	2,200	10	1	0	0	1.0	G1
6	SM	500	14	0	0	1	3.0	G2
7	SM	4,000	11	1	0	1	3.2	G2
8	SM	4,500	10	0	1	1	2.0	G1
9	SM	2,000	7	0	0	0	1.4	G1
10	SM	2,000	7	0	1	1	2.0	G1
11	SM	1,000	5	0	1	0	2.0	G1
12	SM	1,500	7	0	0	0	1.0	G1

a per 10 high-power fields. SM, submucosal; WHO, World Health Organization; G, grade.

The findings of the present study suggest that NETs sized <10 mm may be curatively treated by endoscopic resection. However, NETs with either lymphovascular invasion or sized >1 cm carry a risk for metastasis equivalent to that of adenocarcinomas. Therefore, it is mandatory to histologically examine lymphovascular invasion in specimens retrieved via endoscopic resection to determine the necessity for further radical surgery with regional lymph node dissection. The treatment of NETs sized <20 mm, as presently defined in the WHO 2010 guidelines, requires further evaluation.

References