Primary gastric SCC is rare, accounting for only 0.04–0.07% of all gastric cancers [
2]. And primary gastric SCC is known to have a bad prognosis [
3]. In this case, although curative resection was obtained and adjuvant chemotherapy was performed, the tumor recurred during adjuvant chemotherapy, and the prognosis was poor. After relapse, the tumor was treated by chemotherapy in accordance with the Japanese Gastric Cancer Treatment Guidelines, but it was ineffective. DCF, according to the chemotherapy regimen for esophageal cancer, that followed a previous chemotherapy showed a certain effect. This fact suggests that gastric SCC belongs to a different population that of gastric adenocarcinoma. And, for its rarity and bad prognosis, we have to accumulate more gastric SCC cases and analyze each case in detail as the first step of improving prognosis. Then, it is indispensable to pursue the mechanism of the disease in order to find out the fundamental treatment. And, as a starting point to consider the mechanism of the occurrence of this tumor, a histopathological approach may be helpful.
If the SCC is located in a place where squamous epithelium does not exist and SCC cannot arise naturally, we must determine whether the tumor is primary or metastatic. SCC occurring in the cardia or in conjunction with the esophagus is not considered to be primary gastric SCC; therefore, to be considered a primary lesion, the tumor must arise elsewhere in the stomach, and there should be no evidence of SCC in any other organs [
4]. As mentioned in the introduction, the Japanese classification of gastric carcinoma defines primary gastric SCC as tumors consisting of SCC components definitively generated from the stomach [
1]. In this case, preoperative CT revealed only the primary tumor and one enlarged perigastric lymph node, and we did not find any abnormality or metastatic lesion in the abdominal cavity during the operation. Therefore, we diagnosed this tumor as a primary SCC of the stomach.
Once the possibility of metastasis is excluded and it is confirmed that the tumor is primary gastric cancer, we should consider whether it is really SCC. As the histopathological examination of this biopsy specimen indicated adenocarcinoma, histopathology alone was insufficient to make a diagnosis of SCC. In such cases, an immunohistochemical approach is useful as supporting evidence. For the detection of SCC, immunostaining of CK5/6, p63, and p40 (an isoform of p63) should be performed [
5,
6]. In contrast, for the detection of the origin of adenocarcinoma (including gastric cancer), we usually use a combination of CK7 and CK20 [
5]. We also use CDX2 as a marker of intestinal adenocarcinoma [
7]. The reactivity for these markers in gastric SCC is of course unknown. Hence, it will be inferred from the reactivity in esophageal carcinomas. In the previous report, esophageal SCC consistently expressed CK5/6 (in 98% of all cases) and p63 (100%) with strong reactivity, partially expressed CK7 (34%) and CDX2 (27%), but most of CK7 and CDX2 expression demonstrated weak reactivity [
8]. In this case, the tumor was positive for CK5/6 and CDX2, slightly positive for CK20, and negative for CK7, p63, and p40. CK, expressed in the epithelium and epithelial-derived cells, is one of the proteins that form intermediate filaments. CKs function by forming heterodimers, and the combinations vary depending on the type of epithelial cell, so they are useful for differentiating epithelial cells and epithelial carcinoma cells. In this case, the results of immunostaining for CK5/6, CK7, and CK20 supported that this tumor was SCC.
A question remains: why, in this case, was the tumor negative for p63/p40 and positive for CDX2? Considering what cell types these markers are staining and where they are expressed helps us to infer the pathways of tumor development. There are several hypotheses for the development of primary gastric SCC [
9], including (1) arising from the ectopic squamous epithelium [
2,
10], (2) arising from the squamous metaplasia of the gastric mucosa [
2,
11], (3) squamous differentiation from a pre-existing adenocarcinoma [
12], and (4) arising from multipotent stem cells [
9]. In the histopathological examination of this case, intestinal metaplasia was observed in the mucous membrane around the tumor, but there was no squamous metaplasia. Moreover, the entirety of the tumor consisted of SCC component, and adenocarcinoma component was not detected. Pathological report of the biopsy specimen revealed “adenocarcinoma,” but, with limited specimen, it was difficult to distinguish poorly differentiated carcinoma whether it is adenocarcinoma or SCC. In gastric cancer cases, Japanese pathologists usually diagnose these cases as poorly differentiated “adenocarcinoma,” because almost all of the gastric cancers are adenocarcinoma. We re-examined HE staining of the biopsy specimen, it showed poorly differentiated “carcinoma” and was akin to the part of the tumor specimen. From these findings, hypotheses (2) and (3) are not appropriate for this case. There is a well-known theory that, under the exposure of chronic inflammation, the transformation of tissue stem cells causes metaplasia and dysplasia of the epithelium, then epithelial cancer arises [
13]. In this case, pathogens that could cause chronic inflammation might be HP, EBV, and HPV. As mentioned in the “
Case presentation” section, the patient had a history of former HP infection and background gastric mucosa showed atrophic change with intestinal metaplasia. And chronic gastritis which is developed by HP is known to cause repetitive injury and repair of gastric mucosa, resulting in hyperproliferation and increase of mitotic error, then progress to gastric cancer [
14]. From this point of view, hypothesis (4) seems not to be appropriate, but under the environment that is exposed to chronic inflammation or tissue damage due to infection; there may be a possibility that stem cells will directly become cancer skipping the step of metaplasia or dysplasia (hypothesis (4)). At this time, hypotheses (1) and (4) remain, and it is useful to next consider whether and where p63/p40 and CDX2 are expressed.
p63 is located on chromosome 3q27–29, and p63 acts as a switch for initiation of epithelial stratification in the embryonic epidermis [
6]. p63 is consistently expressed in the nuclei of basal cells of normal epithelia, such as the skin, esophagus, tonsils, ureter, uterine cervix, vagina, prostate, breast, and bronchi [
6]. The 3q27–29 region containing
p63 is amplified in many SCCs, and therefore p63 overexpression is observed in those SCCs [
6]. The majority of the p63 amplified isoforms in SCC are the dominant-negative ΔNp63 forms called p40, and p40 staining is highly specific in those SCCs [
6].
CDX2 is a caudal-type homeobox gene, encoding a transcription factor that promotes proliferation and differentiation of intestinal epithelial cells [
7]. CDX2 is present in the nuclei of epithelial cells of the intestine from the duodenum to the rectum, and the expression of
CDX2 mRNA is highly restricted to the intestinal epithelium [
7]. CDX2 expression is also observed in intestinal metaplasia and gastric adenocarcinoma [
15]. In this case, the tumor was negative for p63/p40 and positive for CDX2. As this suggests that the tumor cells were not derived from ectopic squamous epithelium but from intestinal metaplasia, hypothesis (1) can be excluded. In addition to these facts, the fact that tumor cells looked like homogeneous SCC suggested that these tumor cells might directly arise from stem cells in the intestinal metaplasia of the stomach.