Clinicopathological heterogeneity between primary and metastatic sites of gastroenteropancreatic neuroendocrine neoplasm

Background Chromogranin A (CgA), synaptophysin (Syn) and the Ki-67 index play significant roles in diagnosis or the evaluation of the proliferative activity of gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs). However, little is known about whether these biological markers change during tumor metastasis and whether such changes have effect on prognosis. Methods We analyzed 35 specimens of both primary and metastatic tumor from 779 patients who had been diagnosed as GEP-NENs at Wuhan Union Hospital from August 2011 to October 2019. The heterogeneity of CgA, Syn and Ki-67 index was evaluated by immunohistochemical analysis. Results Among these 779 patients, the three most common sites of NENs in the digestive tract were the pancreas, rectum and stomach. Metastases were found in 311 (39.9%) patients. Among the 35 patients with both primary and metastatic pathological specimens, differences in the Ki-67 level were detected in 54.3% of the patients, while 37.1% showed a difference in CgA and only 11.4% showed a difference in Syn. Importantly, due to the difference in the Ki-67 index between primary and metastatic lesions, the WHO grade was changed in 8.6% of the patients. In addition, a Kaplan–Meier survival analysis showed that patients with Ki-67 index variation had a shorter overall survival (p = 0.0346), while neither Syn variation nor CgA variation was related to patient survival (p = 0.7194, p = 0.4829). Conclusions Our data indicate that primary and metastatic sites of GEP-NENs may exhibit pathological heterogeneity. Ki-67 index variation is closely related to the poor prognosis of patients with tumor metastasis, but neither Syn variation nor CgA variation is related to patient prognosis. Therefore, clinicopathologic evaluation of the primary tumor and metastatic sites could be helpful for predicting the prognosis.


Introduction
Neuroendocrine neoplasms (NENs) are a group of rare and highly heterogeneous neoplasms originating from peptidergic neurons and neuroendocrine cells, which can exist in all parts of the body [1]. Recently, the incidence of NENs has been on the rise. The actual incidence rate increased 6.4-fold from 1973 (1.09 per 100, 000) to 2012 (6.98 per 100,000) according to data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) Program [2]. Gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) are reportedly the most common of these, and account for about 66% of all NENs [3]. In addition, 21 to 73.4% of GEP-NENs were found to have metastases at the time of diagnosis [4][5][6].
Chromogranin A (CgA) and synaptophysin (Syn) widely exist in neuroendocrine cells, and are currently necessary markers for the diagnosis of NENs [7,8]. CgA was initially found in chromaffin particles of adrenal medulla containing catecholamines [9]. It is a kind of acidic hydrophilic secreted protein found in the secretory vesicles of neuroendocrine cells [10]. Syn is an integral membrane glycoprotein present in presynaptic neuron vesicles and the vesicles of normal and neuroendocrine tumor cells [7,11]. In addition, the Ki-67 index plays a crucial role in the World Health Organization (WHO) grading system for GEP-NENs: (grade 1) G1 tumors have a Ki-67 index < 3%; (grade 2) G2 tumors have an index of 3-20%; and (grade 3) G3 tumors have an index > 20% [12]. Syn, CgA and the Ki-67 index play important roles in the diagnosis of NENs. However, little is known about whether these compounds change during tumor metastasis and whether such changes play a role in the overall process of metastasis.
Therefore, the purposes of this study were to investigate the clinicopathological heterogeneity of CgA, Syn, and the Ki-67 index between the primary and metastatic sites, and to evaluate the influence of these clinicopathological features on the prognosis in patients with metastatic GEP-NENs.

Patients and data collection
This study retrospectively collected and analyzed data from 779 patients diagnosed with GEP-NENs at Wuhan Union Hospital from 2011 to 2019; specimens of both primary and metastatic sites were available for 35 patients. The diagnosis of NENs mainly depended on the histological morphology as well as the immunohistochemical staining results of various neuroendocrine markers like CgA and Syn. NENs could also be diagnosed by CD56 or neuron-specific enolase when the immunostaining result of CgA or Syn was absent [13]. All specimens were obtained by surgical excision, fine needle aspiration, and/or core biopsy. Patients for whom complete clinical data were not available for extraction were excluded. The following demographic and clinical characteristics of the patients were collected: age, sex, primary and metastatic tumor sites, immunohistochemical results (CgA, Syn and Ki-67 index), and survival time. The grading system used in this study was based on the WHO 5th edition classification (2019) of digestive system tumors, in which NENs are classified as welldifferentiated neuroendocrine tumors (NETs), poorly differentiated neuroendocrine carcinomas (NECs), and mixed neuroendocrine-non-neuroendocrine neoplasm (MiNEN). NETs were further divided into three grades based on the Ki-67-positive index: G1: < 3%, G2: 3-20%, G3: > 20% [12]. According to the AJCC cancer staging manual, if the metastasis occurred simultaneously with the primary tumor or within 4 months after the initial resection of the primary tumor, the metastasis was considered to be synchronous; if the metastasis occurred more than 4 months after the initial resection of the primary tumor, it was considered as metachronous [14].
The study was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (IORG number: IORG0003571) and was conducted in accordance with the Declaration of Helsinki.

Immunohistochemistry
We performed immunohistochemical staining for CgA, Syn and Ki-67. The primary and metastatic tumor tissues were fixed in 4% paraformaldehyde and embedded in paraffin blocks. Each block was cut into 4 μm-thick section, and deparaffinized with xylene and rehydrated with ethanol. Then the sections were incubated with hydrogen peroxide at the room temperature, then antigen retrieval and serum blocking were performed. After these steps, they were incubated with the CgA (M0869, Dako, Glostrup, Denmark), Syn (M7315, Dako) and Ki-67 antibodies (MIB-1, DAKO, Agilent Technologies, Santa Clara, CA, USA) at 37°Cfor 1 h. Then they were incubated with the secondary antibody containing the biotin. Last, slides were visualized by using the DAB (AR1022, BOSTER Biological Technology), stained with hematoxylin, dehydrated with ethanol, transparentized with xylene, and counted.
The staining results for CgA and Syn were graded according to the extent of positive cells as follows: grade I (negative, less than 5% positive tumor cells), grade II (focally positive, 5 to 50% positive tumor cells), grade III (positive, more than 50% positive tumor cells) [15]. For Ki-67 staining, the areas with abundant and more positive tumor cells were selected. The Ki-67 labeling index was determined by calculating the percentage of positive nuclei of 2000 tumor cells in the densest area of each slide [16]. Each slide was reexamined by the pathologist (Q.Z.).

Statistical analysis
SPSS software v24.0 (IBM, Armonk, NY, USA) or GraphPad Prism v6.0c (GraphPad Software, San Diego, CA, USA) was used for statistical calculation and data processing. The clinicopathological characteristics of patients were expressed as median and range, absolute value or fractions. Overall survival (OS) was defined as the date of initial diagnosis to the last follow-up or date of death. Survival curves were drawn according to the Kaplan-Meier analysis, and differences between groups were assessed using the log-rank test. P < 0.05 was regarded as statistically significant.
Heterogeneity of CgA, Syn and the Ki-67 index between primary and metastatic sites For 35 patients with both primary and metastatic specimens, the clinicopathological differences between the primary and metastatic sites were analyzed. There was no significant difference in the overall positive rates of CgA and Syn, or in the grading or classification of tumors between the primary and metastatic sites (Supplementary Table 1). However, the variation of CgA, Syn and the Ki-67 index existed between the primary and metastatic sites. Examples of CgA, Syn and the Ki-67 index variation between the primary tumor sites and metastatic sites were shown in Fig. 1.

Survival analysis
The  in survival time between NET and NEC groups, between synchronous and metachronous metastatic tumor groups, between Ki-67 index up-regulated and down-regulated groups (Fig. 5g-i). To explore the different treatment methods on patient survival, patients were grouped into four groups including surgery, adjuvant therapy, surgery and adjuvant therapy, and no treatment. The results showed that the prognosis of patients receiving surgery or both surgery and adjuvant therapy was better than that of patients who received adjuvant therapy only (p = 0.0350, p = 0.0109) (Fig. 6).  I  I  III  III  70  70  8  -23  stomach  iliac fossa  III  II  III  III  35  30  16  NEC → NEC   24  stomach  lymph node  I  I  III  III  /  80  15  -25  stomach  lung  I  II  III  III  70  70  15  -26  stomach  liver  I  I  II  III  80  50  1  NEC → NEC   27  stomach  liver  III  II  III  III  80  50  36  NEC → NEC   28  duodenum  liver  III  III  III  III  5  2  5  NET G2 → NET G1   29  duodenum  liver  II  II  III  III  5  5  11  -30  duodenum  adrenal gland  II  II  III  III  20  35  29  NEC → NEC   31  colon  mesentery  II  II  III  III  60  /  55  -32  unknown  liver  /  /  III  III  80  60  5  NEC → NEC   33  unknown  liver  II  III  III  III  80  90  1  NEC → NEC   34  unknown  peritoneal  II  II  II  II  30  30  21  -35  unknown  rectum  III  III  III  III  1  2  61 NET G1 → NET G1 Abbreviations: CgA chromogranin, Syn synaptophysin, NEC neuroendocrine carcinoma; a CgA and Syn graded according to the percent of positive cells as follows: I: < 5%, II: 5-50%, and III: > 50%

Discussion
GEP-NENs represent a heterogeneous family of neoplasms with variable biological and clinical characteristics. Few studies have attempted to assess the heterogeneity of the expression of biological markers in primary and metastatic GEP-NENs. Therefore, in this study, we investigated the pathological heterogeneity of primary and metastatic sites of GEP-NENs and the effect of such heterogeneity of pathological indicators on the prognosis of patients.
In the SEER database, the rectum and small intestine are the most common sites of GEP-NENs [2,3]. Our results showed that the pancreas, rectum and stomach were the most common sites, which were consistent with the results of a large retrospective single-center study in China [17]. We found that the metastasis rate of patients with GEP-NENs was 39.9%, and the most common site of distant metastasis was the liver. It has been reported that distant metastasis of NENs is a major factor affecting the prognosis [18].
NENs synthesize, store and secrete various peptides and amines [19], which exist in the blood or tissues of patients. These have become important markers for the diagnosis and follow-up of patients with NENs [20]. Among them, CgA and Syn are considered to be the most important biomarkers for the diagnosis of NENs [21,22]. Our results showed that the positive rates of CgA and Syn in the primary site were 57.5 and 85.4%, respectively. Many studies have shown that the CgA level and Ki-67 index are closely related to NENs metastasis [23][24][25][26], and are useful for determining the prognosis [27][28][29][30].
Although most studies have explored the relationship between these biological indicators and tumor metastasis and prognosis, little attention has been paid to the heterogeneity of marker expression between primary and metastatic tumor sites, and the impact of such heterogeneity on disease progression has not been fully evaluated. Lindholm et al. [31] found that in small-intestinal NENs, 38.5% showed the grade variation in CgA and 54% showed the grade variation in Syn between the primary and metastatic sites. In this study, 37.1% of cases showed the variation in CgA and 11.4% showed the variation in Syn between the primary and metastatic tumor sites. The variation of CgA was in line with Lindholm's study, and we found that the variation of CgA in different primary tumor sites was similar, ranging from 42.9 to 50%. We consider the difference may be due to the heterogeneity of NENs at different sites in Syn variation. In our study, the variation rates of Syn were different in different sites of NENs, with the highest variation rate of Syn in rectal NENs (28.6%) while the lowest variation rate of Syn (7.1%) in pancreatic NENs. Due to the high incidence of smallintestinal NENs in Europe and America, Lindholm et al. only focused on small-intestinal NENs, while the incidence of small-intestinal NENs in China was extremely low, so our study focused on NENs in the whole digestive tract. We consider this heterogeneity of Syn variation might be site-related, and further Singh et al. [32] reported that, in 37.0% of NENs patients, the Ki-67 index changed during the course of the disease, and the WHO classification was upregulated in 27.8% of these patients. Grillo et al. found 39% discrepancy in grade between primary and metastatic tumors [33]. Keck KJ et al. found grade variation occurred in 34% of patients between primary and metastatic tumors [34]. In our study, we found that, in 54.3% of GEP-NENs patients, the Ki-67 index differed between primary and metastatic lesions; as a result, the WHO classification changed in 8.6% of the patients.
In addition, we further analyzed the heterogeneity of CgA, Syn, and the Ki-67 index according to the primary tumor site, and found that these markers differed between the primary and metastatic sites. While there was no significant variation in CgA or Syn in the pancreas, rectum or stomach, the Ki-67 index showed obvious variation, especially in the rectum, where variation was seen in 85.7% of the patients. Interestingly, we also found that CgA and Ki-67 variability were more common in metachronous metastatic tumors than in synchronous metastatic tumors. The Ki-67 variation, in particular, was as high as 80% in metachronous metastases, which was in line with the study, that showed 83% of patients had an increase in Ki-67 rate in the metachronous metastatic site and a change in grade [33]. A Kaplan-Meier survival analysis showed that the overall survival time of patients with the Ki-67 index variation was significantly shortened, which was consistent with the previous report [35], while CgA and Syn variation had no significant correlation with patient survival. The mechanism of this variation in CgA, Syn and the Ki-67 index between primary and metastatic NENs is still unclear. A previous study deduced that inconsistencies in the expression of biomarkers in primary and metastatic tumors may be caused by the heterogeneity of the tumor itself [33,36], which may affect the clinical treatment [37].
In summary, this study showed that the Ki-67 index showed heterogeneity between the primary and metastatic foci, and that patients with such Ki-67 index variation had a poor prognosis. In contrast, there was little variation in CgA or Syn expression during disease progression, and CgA and Syn variation had no effect on the prognosis. The mechanism that underlies this heterogeneity has yet to be determined. Further studies with a larger sample size will be needed to specifically explore the mechanism of this tumor-marker variation.

Conclusions
Our results show that there is pathological heterogeneity between the primary and metastatic sites of GEP-NENs.
The Ki-67 index shows obvious heterogeneity and patients with such Ki-67 index variation have a poor prognosis. In contrast, there is little heterogeneity in CgA and Syn expression throughout disease progression, and their variations have no effect on the prognosis. Close attention should be given to the clinicopathological heterogeneity between the primary and metastatic lesions to better monitor the progress of NENs and potentially contribute to a more effective treatment.