Mixed adenoneuroendocrine carcinomas of stomach and ampulla of vater after curative-intent resection: a single center cases series

Background Mixed adenoneuroendocrine carcinoma is a rare clinical manifestation, especially in the gastric and ampullary. The purpose of this study was to investigate the clinicopathological features and prognosis of mixed adenoneuroendocrine carcinoma in the gastric and ampullary and summarize related treatment suggestions. Methods In all, 32 cases of mixed adenoneuroendocrine carcinoma in the gastric and ampullary that were diagnosed from resected specimens were analyzed from 2009 to 2015. The corresponding demographic, clinicopathological and survival data were retrospectively reviewed. Results The 1-year, 3-year and 5-year survival rates were 78.1%, 28.1 and 9.4%, respectively, and the median overall survival was 28.0 months. In all, 75.0% (24/32) had lymph node metastasis at the time of initial diagnosis. A multivariate analysis revealed that TNM stage (HR 6.444 95%CI 1.477–28.121 P = 0.013), lymph nodes metastasis (HR10.617 95%CI 1.409–79.997 P = 0.022), vascular invasion (HR 5.855 95%CI 1.719–19.940 P = 0.005), grade of the adenocarcinoma component (HR 3.876 95%CI 1.451–10.357 P = 0.007) and CD56 positivity (HR 0.265 95%CI 0.100–0.705 P = 0.008) were independent predictors of overall survival. Conclusions Mixed adenoneuroendocrine carcinoma is an aggressive clinical entity with a poor prognosis. Taking both the neuroendocrine component and the adenocarcinoma component into consideration of optimal treatment is strongly recommended.


Background
According to the recent WHO classification from 2010, mixed adenoneuroendocrine carcinomas (MANEC) are composed of both malignant neuroendocrine and exocrine components, and each of them must exceed 30% of the entire tumor cell population [1]. Synaptophysin (Syn), chromogranin (CgA) and CD56 are commonly used neuroendocrine makers, two of which must be positive for a diagnosis of MANEC [2,3]. MANEC is clinically rare, and its morphology is recognizable as both gland-forming epithelial and neuroendocrine neoplasms [4]. MANECs have been reported in the colon, rectum, esophagus, stomach, and pancreas, among other sites. Although most of them are case reports [5][6][7][8], a few studies of MANEC cases in the gastric and ampullary on a much larger scale have been published, the clinicopathological Open Access features and prognosis of patients with MANEC in the gastric and ampullary are still unclear, however, there are several researches reported that biological and histological characteristics of MANEC in the stomach and ampulla of vater were similar at some aspect [9][10][11], most of patients were diagnosed in advanced stages with lymph nodes metastasis, both MANEC in stomach and ampulla of vater are aggressive and lack of effective treatments [12,13], what'more, p53 nuclear accumulation was detected in most patients of MANEC in stomach and ampulla of vater, and there is a special kind of MANEC with a component "composite glandular and endocrine tumor with pancreatic acinar differentiation" founded in only stomach and ampulla of Vater, which may indicate that there colud be a modicum of similar mechanisms of cancerogenesis in those two sites [4,10,14]. In our study, the data of 32 MANEC patients who were treated at West China hospital between 2009 and 2015 were retrospectively analyzed. In this study, we analyzed the clinicopathological features and prognosis of patients with MANEC in the gastric and ampullary and summarized related treatment suggestions according to other recently published literatures.

Study subjects
This is a retrospective study. First of all, we choosed "adenocarcinoma" and "neuroendocrine carcinoma" as key words to do a search on our pathological database of resected specimens from 2009 to 2015, then we went through the qualified pathological reports, only the sites of tumor were in stomach or in the ampulla of Vater did the patients are involved in the next step. Finally, tumor that were not composed of adenocarcinoma or a poorly differentiated neuroendocrine carcinoma (G3) were excluded. If either the exocrine or neuroendocrine components did not exceed 30% of the entire neoplasm, cases were also excluded (as showed in Fig. 1). The pathological diagnosis was supported by multiple histologic examinations including hematoxylin and eosin staining and immunohistochemical staining (Fig. 2).

Clinicopathological data
All the data obtained from the patients was retrospectively reviewed and included age, gender, symptoms, blood type, tumor location, levels of serum alpha fetoprotein (AFP), carbohydrate antigen 19-9 (CA19-9) and carcino embryonic antigen (CEA), ultrasonography, computed tomography, magnetic resonance image and treatments. All the resected specimens were reviewed pathologically and were assessed for tumor size, depth of invasion, pathological classification, histological differentiation, lymph node status, presence of lymphatic and perinerual invasion, cell types of neuroendocrine carcinoma, vascular invasion, Ki-67 index and immunohistochemical staining. The margins were examined for the presence of residual tumor, which was described by the residual tumor classification, as follows: R0, no residual tumor and resection margin was > 0 mm; R1, microscopic residual tumor; R2, macroscopic residual tumor. Tumors were staged according to the American Joint Committee on Cancer (AJCC) 7th TNM staging system [15].

Follow-up
All the patients received clinical and radiographic followup every 2-3 months the first year after the surgery and then 3-6 months annually thereafter. The patients who were at high risk for recurrence received chemotherapy. Follow-up ended December 31, 2017.

Statistical analysis
Overall survival (OS) was calculated from the time of surgical resection to the time of death. Disease-free survival (DFS) was estimated from the time of surgical resection until the time of relapse confirmed by biopsy or radiologic imaging. Survival curves were generated by the Kaplan-Meier method, and the log-rank test was performed to evaluate the survival difference. All predictors that were statistically correlated with survival in the univariable analysis were included in a multivariable analysis using the Cox proportional regression model. When the P value was less than 0.05, the difference was considered statistically significant. SPSS 24.0 statistics software was used for data analysis.

Discussion
MANEC is clinically rare since only when the exocrine and neuroendocrine components each exceed 30% can the tumor be considered as MANEC. Gastric and ampullary MANEC are highly exceptional entities [13,16,17]. To the best of our knowledge, this research is the largest case series of MANEC in the gastric and ampullary.
Three commonly used neurendocrine makers, Syn (96.9%), CgA (81.3%) and CD56 (56.3%), were all observed in our study. This result is similar to that of the study by Brathwaite et al. in that Syn and CgA were expressed in 97 and 82% of cases, respectively 18]. A previous report indicated that several markers, including CK7 and CK20, detected by immunohistochemical staining did not reveal any correlation with prognosis in MANECs [19]. However, regarding hematopoietic stem cell markers, immunoreactivity for CD117 has already been considered an indicator of malignancy in pancreatic neuroendocrine carcinomas [20][21][22][23]. In our study, we found that although CD56 was not as sensitive as CgA (81.3%) and Syn (96.9%), CD56 demonstrated a strong association with overall survival. CD56 is an important marker of neuroendocrine tumors and is also a glycoprotein of the immunoglobulin (Ig) superfamily expressed on natural kill (NK) cells, NK-T cells, and in vitroexpanded cytokineinduced cells [24]. It had been shown that CD56 + cells exhibited cytotoxic effect against tumor targets [25]. In our study, the OS of the CDK56-positive group was much longer than that of the CD56-negative group (30.0 months vs. 12.0 months) (P = 0.006). In the multivariate analysis, CD56 positivity (HR 0.265 95%CI 0.100, 0.705 P = 008) was an independent prognostic predictor of OS. This may be related to the cytotoxic effect of CD56 + cells. However, CD56 was also expressed in a subset of biliary epithelial cells, especially in intrahepatic  small bile ducts, and thus it might not be the best marker of neuroendocrine tumors in the biliary duct [26]. In this study, more than half of the patients presented with perinerual invasion (13/32, 61.9%) or vascular invasion (13/32, 61.9%), which is a little lower than Wantanabe's and La Rosa's results [19,27]. In our study, vascular invasion (HR 5.855 95%CI 1.719-19.940 P = 0.005) was one of the independent predict factors on overall survival, which was consistent with La Rosa's conclusion that vascular invasion had a negative effect on the prognosis [19]. There were 84.4% of patients with the presence of the lymphatic invasion and it was a predict factor of bad outcome in univariate analysis (P = 0.018). which was disagreement with the La Rosa's result that the patients with the presence of lymphoid infiltrate had a better prognosis [19]. But both of two analysis had a limited case samples of MANECs and were lost in multivariate analysis. Hence, further studies are needed on impact of lymphatic invasion.
The optimal treatment of MANEC is still controversial. Most published studies have indicated that the treatment should be based on the most aggressive histologic component [4,16,28], but which histologic component is the most aggressive? Some research has focused on this question. Chen et al. reported that the pure neuroendocrine component accounts for a large proportion (6/9) of lymph node metastasis in MANEC patients, and patients who have died had a predominantly neuroendocrine component in the primary tumor or in metastatic lymph nodes [29]. This finding was in agreement with that of the study by Harada et al. [30]. They found that at the surface of the tumor, the adenocarcinoma components were in the highest flight. Neuroendocrine components are involved in most stromal and vascular invasion and LN metastasis. Moreover, neuroendocrine components showed higher proliferative activity than adenocarcinoma components [30,31], which suggested that the neuroendocrine component was more aggressive and defined the prognosis of MANEC [32]. However, in our study, we excluded all the patients with neuroendocrine tumors (G1/G2) according to the definition of MANEC by WHO [1], and we found that patients with well/moderately differentiated adenocarcinoma had a better prognosis. In the multivariate analysis, the grade of the adenocarcinoma component (HR 3.876 95%CI 1.451-10.357 P = 0.007) was an independent predictor of overall survival, which was agreed with the conclusion of Nie's research [17]. Hence, we suggested that the optimal treatment of MANEC should take both the neuroendocrine component and the adenocarcinoma component into consideration.
Several limitations of our study should be noted. Our data were retrospectively collected from a single medical center, which may carry an inherent risk of bias. The small sample size in this study also limited our ability to conduct further analyses. Despite these limitations, this study reflected the actual clinical features and prognostic factors of MANEC of the gastric and ampullary.