Comparison of clinical and pathological features between early-stage gastric-type and intestinal-type differentiated adenocarcinoma: a retrospective study

Gastric carcinoma (GC), one of the most common malignancies in the world, is known as a kind of heterogeneous disease displaying a variety of phenotypes with different prognostic characteristics. Histologically, gastric cancer has been divided into two major categories, intestinal and diffuse types according to Lauren classification system [1], or differentiated and undifferentiated types according to Nakamura classification system based on glandular structure [2]. According to the latest classification released by Japanese Gastric Cancer Association (JGCA) in 2017, gastric cancer is generally classified into five major types: papillary, tubular, poorly differentiated, mucinous, and signet-ring cell [3]. Among different types, papillary adenocarcinoma (pap), well differentiated tubular adenocarcinoma (tub1), and moderately differentiated tubular adenocarcinoma (tub2) belong to differentiated GC. Poorly differentiated, mucinous, and signet-ring cell belong to undifferentiated GC. The histological type is determined according to its predominant component. Therefore, if the proportion of differentiated components in the lesion is greater than or equal to 50%, it is regarded as differentiated gastric cancer. Otherwise, it is regarded as undifferentiated gastric cancer.

Mucin is a heavily glycosylated glycoprotein produced by cancers [4]. With increasing research on mucin and the progress of immunohistochemical studies, a mucin phenotype subclassification come into sight, independent of histologic type. Differentiated gastric adenocarcinomas can be classified as having either gastric, intestinal, mixed or unclassified mucin phenotype based on the expression of MUC5AC (a marker of foveolar cells), MUC6 (a marker of gastric mucous neck cells and pyloric glands), MUC2 (a marker of intestinal goblet cells), and CD10 (a marker of the brush border of the small intestinal absorptive cells) (Fig. 1).

Accumulating evidence has indicated that mucin phenotype correlated with the development pattern and malignant degree of gastric adenocarcinoma. Several studies indicated that the gastric phenotype had a higher potential for lymph node metastasis, deeper invasion depth, and correlated with a worse prognosis than the intestinal type [5‐7]. On the contrary, Kim et al. indicated that the gastric phenotype was linked to a better prognosis [8]. Most of the above studies investigated advanced GCs and did not focus on early GCs. Whether early gastric-type differentiated adenocarcinoma (GDA) have more malignant potential than early intestinal-type differentiated adenocarcinoma (IDA) remains controversial due to insufficient data.

Magnifying endoscopy with narrow band imaging (ME-NBI) is an effective method for the detailed visualization of microsurface and microvascular structures within the superficial layer of the gastric mucosa. Recent studies showed that the ME-NBI is able to predict histological characteristics of early gastric cancers [9‐13]. Two studies have also reported that mucin phenotype was significantly correlated with ME-NBI results in early gastric cancers. However, the results in these studies were not consistent. In addition, the sample size of these studies was small.

It is important to characterize the clinicopathological and endoscopic differences between early GDAs and IDAs in order to clarify the clinical significance of mucin phenotype. In this study, we investigated the clinicopathological and endoscopic features of early GDAs in comparison with early IDAs. We also compared the endoscopic curability between GDAs and IDAs based on the Japanese gastric cancer treatment guideline. To the best of our knowledge, the present study is the largest study, conducted to date, evaluating the significance of mucin phenotype of early differentiated gastric adenocarcinoma [5, 8, 14‐16].

In the present study, 657 early differentiated gastric carcinomas were classified as gastric phenotype tumors (307; 46.7%), intestinal phenotype tumors (109; 27.5%), mixed phenotype tumors (181; 27.5%), and unclassified phenotype tumors (60; 9.1%). Previous reports have shown the incidences of gastric, intestinal, mixed, and unclassified phenotype GCs to be 16.3–53.1%, 10.3–35%, 19.4–56% and 0–19.6%, respectively. The percentage of mucin type varied considerably among different studies. One possible reason is that different study uses different definitions of mucin positivity, antibodies, and case groups. We showed that gastric type is the most common mucin phenotype among early differentiated gastric adenocarcinomas [15, 19, 26].

To clarify the significance of the mucin patterns in early differentiated gastric cancer, we analyzed the relation between gastric/intestinal patterns and the clinicopathologic parameters using ESD cases from our hospital. No significant difference was observed in terms of gender, age, tumor size, gross type, tumor location, background mucosa, lymphatic invasion and vascular invasion between patients with GDA and IDA. However, GDA cases were associated with deeper invasion than IDA cases. This may indicate a more biological aggressiveness of GDAs than IDAs.

Differentiated-type-predominant early GC including undifferentiated-type components has been shown to have higher malignancy than pure differentiated-type early GC [27, 28]. Song et al. reported that gastric-type mucin was expressed in mixed-type early GC with a higher frequency, compared with intestinal-type mucin. Gastric-type tumors were associated with a higher proportion of undifferentiated-type tumors [29]. Our study also shows GDA was associated with a higher proportion of undifferentiated components. However, the difference did not reach statistically significance (p = 0.107).

Although clinicopathological factors of different mucin phenotype have been reported in many studies, no study has investigated the association between mucin phenotype and endoscopic curability. According to the most recent Japanese gastric cancer ESD guideline, endoscopic curability of early gastric cancer can be divided into eCura A, eCura B, eCura C1, and eCuraC-2 [24]. eCura A and eCuraB have been regarded as curative resection, which are considered as having the same therapeutic effect as surgery. eCura C1 and eCura C2 have been regarded as no-curative resection, needing additional treatment after ESD. In this study, the proportion of no-curative resection in GDAs was significantly higher than that in IDAs, indicating that GDA is less endoscopically resectable than IDA. Valente et al. reported that the depressed type of intramucosal differentiated-type gastric cancer has higher malignant potential in comparison to the elevated type [30]. The GDA showed a higher proportion of depressed lesions and undifferentiated-type components than the IDA, though the difference did not reach statistically significance. The above result may in part illustrate the higher malignant potential of the GDA compared with the IDA. Therefore, endoscopic submucosal dissection for early gastric-type differentiated adenocarcinoma must be carried out with more caution and followed-up more intensively. This is the first study to investigate an association between mucin phenotype and curability of endoscopic resection based on the classification of endoscopic curability in the Japanese gastric cancer guideline.

Previous studies have posited that the GDA contains a higher potential for lymphatic and vascular invasion and lymph node metastasis than the IDA, consequently having worse prognosis [8‐10]. In our study, the GDA cases revealed deeper invasion than the IDA cases, but there were little lymphatic and vascular invasion either in the GDA cases or in the IDA cases. Owing to the endoscopic resection, we cannot compare the rate of lymph node metastasis in different groups. Although GDA invaded deeper than IDA, these lesions were still confined to the submucosa, which may explain why the prognosis had no significant difference between gastric type and intestinal type in early gastric cancer.

Since GDA shows more malignant potential, early recognition and diagnosis under endoscopy is quite important. ME-NBI system can predict the histological characteristics of gastric cancer lesions [10]. Yokoyama et al. have shown that ILL and FNP were the main microvessel patterns for most differentiated GCs. This study showed that GDAs mainly displayed an ILL pattern, whereas IDAs demonstrated a higher ratio of FNP pattern. Kobayashi et al. also reported that GC with gastric mucin showed a high ratio of ILL pattern (61.5%) and GC with intestinal mucin showed a high ratio of FNP pattern (66.7%). Our results confirmed these findings, which indicated ILL and FNP were the main microvessel patterns for GDAs and IDAs, respectively. In contrast, OK et al. described that the most common microvessel pattern in gastric type adenocarcinoma was CSP pattern [12]. However, they also included 34.4% undifferentiated adenocarcinomas and 90.9% undifferentiated adenocarcinomas displayed CSP pattern in their study. This may account for the predominance of the CSP pattern in gastric type adenocarcinomas. Study by OK et al. also classified microsurface patterns as tubular/oval, papillary, destructive, and absent, and showed that microsurface patterns did not vary significantly with mucin phenotypes. Our study did not compare the microsurface pattern between different groups due to the following reasons. First, it was subjective to classify the microsurface into destructive and absent pattern. Second, surface glandular structure in some differentiated adenocarcinomas could not be observed clearly without crystal violet or acetic acid staining. In this study, there was a significant difference in the microvascular pattern between GDAs and IDAs. These findings suggest that expression of gastrointestinal mucins may in part help to explain the differences in the vascularization patterns of early differentiated gastric adenocarcinomas. However, it is difficult to use microvascular patterns alone as a meaningful diagnostic tool to predict mucin phenotypes due to the low specificity. Microvessel patterns combined with immunohistochemistry of biopsy specimens prior to ESD may be helpful in differentiating mucin types. Further prospective studies will be needed to verify the clinical significance of microvascular patterns in early differentiated gastric adenocarcinomas.

Our study has some limitations. First, only endoscopically resected early gastric cases were included and early or advanced gastric cancer cases resected by surgery were excluded. The methods of sampling of ESD specimens and surgical specimens may be different and the pathological evaluation of the depth of invasion of surgical specimens is less accurate than that of ESD. Moreover, a great number of surgically resected GDA cases did not undergo preoperative magnifying endoscopy. Second, mixed phenotype and unclassified phenotype were not included because we aimed to compare the clinicopathological features between early gastric-type and intestinal-type differentiated adenocarcinoma. In addition, there is little or no understanding of mixed phenotype and unclassified phenotype, especially in the mechanism of carcinogenesis and development. The underlying processes resulting in these two phenotypes require further research.

The clinicopathological characteristics of gastric cancer with gastric phenotype differ from those of gastric cancer with intestinal phenotype in terms of invasion depth, microvessel pattens in ME-NBI, and endoscopic resectability.