A nationwide, multi‐institutional collaborative retrospective study of colorectal neuroendocrine tumors in Japan

Abstract Aim Neuroendocrine tumors (NETs) are one of the subtypes of neuroendocrine neoplasms and are defined as epithelial neoplasms with predominant neuroendocrine differentiation. The aim of this study was to clarify the clinicopathological characteristics of colorectal NETs through a nationwide retrospective study in Japan. Methods This multicenter retrospective cohort study of NETs in Japan was conducted by the study group of the Japanese Society for Cancer of the Colon and Rectum. In this study, we aimed to clarify the characteristics of Japanese patients with colorectal NETs. This cohort study included patients with colorectal NETs who were treated from January 2011 to December 2015. Results Most NETs developed in the lower rectum. Predictive factors of lymph node metastasis included size (>10 mm), depth of invasion (muscular propria or greater), NET grade (NET G2), depressed lesion of the tumor, and lymphovascular infiltration. In particular, depressed lesion of the tumor and lymphovascular infiltration were independent predictive factors of lymph node metastasis. The presence of an increased number of these predictive factors increased the lymph node metastasis rate. Conclusion Surgical resection with lymph node dissection is considered in the colorectal NETs patients with predictive factors of lymph node metastasis, the number of which is correlated with incidence of lymph node metastasis.


| INTRODUC TI ON
Neuroendocrine tumors (NETs) are one of the subtypes of neuroendocrine neoplasms (NENs) and are defined as epithelial neoplasms with predominant neuroendocrine differentiation. Since neuroendocrine cells are distributed widely throughout the body, and NENs can arise at various locations, including the respiratory system and the digestive system. 1 The World Health Organization (WHO) previously proposed a classification scheme for digestive NENs that divides them into three categories based on the mitosis count and Ki-67 labeling index value: NET G1, NET G2, and neuroendocrine carcinoma (NEC). 2 In particular, a mitotic count of <2 per 10 highpower fields (HPFs) and/or a Ki-67 index <3% corresponds to NET G1, a mitotic count of 2-20 per 10 HPFs and/or a Ki-67 index of 3%-20% corresponds to NET G2, and a mitotic count of >20 per 10 HPFs and/or a Ki-67 index >20% corresponds to NEC. In 2019, the WHO revised its former classification scheme and instead established a well-differentiated subtype as NET G3 from among those cases previously classified as NEC. 3 A total of five to seven new digestive NET cases per 100 000 people per year are diagnosed, 4,5 and colorectal NETs, including appendiceal NETs, account for approximately 50% of all digestive NETs. 4,6 Difference in the location of development of colorectal NETs in white and non-white patients due to ethnic differences has been reported. It has been reported that NETs derived from the hindgut were predominant in non-white patients, whereas those derived from the midgut were predominant in white patients. 4,[6][7][8][9] To date, only few reports have analyzed a large number of colorectal NET cases. Thus, the aim of this study was to clarify the clinicopathological characteristics of colorectal NETs through a nationwide retrospective study in Japan.

| ME THODS
This multicenter retrospective cohort study of NETs in Japan was conducted by the study group of the Japanese Society for Cancer of the Colon and Rectum (JSCCR). In this study, we aimed to clarify the characteristics of Japanese patients with colorectal NETs. This cohort study included patients with colorectal NETs who were treated from January 2011 to December 2015. The NET classification was performed according to the 2010 edition of the WHO classification scheme. The cohort study protocol was approved by the JSCCR Ethics Committee and the institutional review board of each involved center. Clinical information was collected either from medical records or directly from the patients.
The data in this study are presented as totals, medians (ranges or standard deviations), or percentages (95% confidence intervals).
Statistical analysis was performed using Fisher's exact test and the Mann-Whitney U test. Multivariable logistic regression models were performed to identify clinical and pathologic differences between lymph node-negative and lymph node-positive patients.
We included the following variables in the multivariate logistic regression analysis: tumor size (≥10 mm or <10 mm), depressed lesion of the tumor, lymphovascular infiltration, invasion to muscular propria, and NET grade (G1 or G2), as these factors are well associated with lymph node metastasis in past reports. [10][11][12][13] The depressed lesions of the tumor were diagnosed endoscopically and/or histologically.
Statistical significance was defined at the level of P < .05. All statistical analyses were performed with EZR (http://www.jichi. ac.jp/saita ma-sct/Saita maHP.files/ statm edEN.html; Kanda, 2014; Saitama Medical Center, Jichi Medical University), which is a graphical user interface for R (version 3.6.0; The R Foundation for Statistical Computing). More specifically, the interface is a modified version of R Commander (version 2.5-3) that was designed to add statistical functions frequently used in biostatistics.

| RE SULTS
In total, 416 patients from 25 institutions were enrolled in the present multicenter retrospective cohort study of colorectal NETs in Japan. Twenty-six patients were excluded from the primary analysis because they could not be confirmed as colorectal NET cases.
Thus, a total of 390 patients were included for the primary analysis The patients with colorectal NETs without any metastasis underwent endoscopic resection (n = 315) and surgical resection (n = 51).
Of the 254 patients with colorectal NETs who received no additional treatment after endoscopic resection, one experienced lymph node recurrence. Additionally, after surgical resection, five patients experienced liver metastasis and three experienced local recurrence.
The univariate analysis demonstrated that size (≥10 mm), depth of invasion (muscular propria or greater), NET grade (NET G2), depressed lesion of the tumor, and lymphovascular infiltration were predictive factors of lymph node metastasis (Table 2). Overall, the greater the number of predictive factors for lymph node metastasis, the greater the rate of lymph node metastasis (P < .0001; Figure 2).
In addition, a multivariate analysis demonstrated that depressed lesion of the tumor and lymphovascular infiltration were independent predictive factors of lymph node metastasis.
The clinicopathological characteristics of colorectal NETs according to liver metastasis were shown in Table 3. However, we performed neither a univariate nor a multivariate analysis because of the small number of the patients with liver metastasis.

| D ISCUSS I ON
The present study demonstrated the following: (a) most NETs devel- Predicting lymph node metastasis by preoperative computed tomography is difficult. 14 In general, it is recommended that the NET patients with some risk factors for lymph node metastasis undergo surgical treatment. It has previously been reported that tumor size (≥10 mm or <10 mm), depressed lesion of the tumor, lymphovascular infiltration, invasion into the muscular propria, and NET grade (G1 or G2) are predictive factors for lymph node metastasis. [10][11][12][13] These factors were associated with lymph node metastasis in this study as well. Additionally, we demonstrated that an increase in the number of these predictive factors increased the lymph node metastasis F I G U R E 1 The treatment outline of colorectal neuroendocrine neoplasms patients. Clinical diagnosis, treatment course, recurrence site and tumor grade were shown. After a total of 390 patients had undergone surgical resection, liver metastasis in five patients and local recurrence in three patients occurred. CTx, chemotherapy; ETx, endoscopic treatment; H−, negative of liver metastasis; H+, positive of liver metastasis; N−, negative of lymph node metastasis; N+, positive of lymph node metastasis; NET, neuroendocrine tumor; Sur., surgical resection  18 However, they also added that long-term follow-up of 10-20 years is recommended to assess for any delayed recurrence in rectal NETs patients without surgical resection.

TA B L E 1 Clinicopathological characteristics of colorectal neuroendocrine tumors
The liver is the second most common distant organ of metastatic NET following lymph node. 19 Tumor size was reported to Nonetheless, considering that we elucidated features of colorectal NETs in Japanese patients in our investigation, we believe that our findings will help researchers and physicians alike to clarify colorectal NETs.

ACK N OWLED G EM ENTS
The authors would like to acknowledge all the patients and their families. In addition to the investigators in the author list, we acknowledge the following investigators who also participated in this