Metastatic colon cancer of the small intestine diagnosed using genetic analysis: a case report

Background Intestinal-type adenocarcinoma is widely detected in the gastrointestinal tract, head and neck, lower respiratory and urinary systems. Determining the nature (monoclonal or multicentric) of the intestinal adenocarcinoma is sometimes a diagnostic challenge owing to its occurrence at various locations of the body, especially in the lower gastrointestinal tract. Herein, we successfully diagnosed metastatic colon cancer in the small intestine using tumor protein 53 gene (TP53) mutation analysis. Case presentation An 83-year-old woman presented with severe abdominal pain and nausea at the emergency department of the hospital. Her history included surgery and adjuvant chemotherapy for colon and breast cancers. Abdominal computed tomography revealed small intestinal dilation, which was associated with the mural nodule detected on fluorodeoxyglucose positron emission tomography. Laparoscopy-assisted small bowel resection was performed based on the diagnosis of small bowel obstruction, probably due to recurrence of the colon or breast cancer. Macroscopically, an ulcerated tumor was present in the resected small intestine. Histologically, the cancer cells showed infiltrative growth of colonic dysplastic glands, whose non-specific finding made it difficult to determine the relationship with past colon cancers. Retrospective pathological examination confirmed that the previous breast and colon carcinomas were primary cancers. Immunohistochemical analysis revealed that the small intestinal and colon cancer cells showed diffuse positive tumor protein 53 (p53) expression. However, the breast cancer cells showed only weakly positive p53 expression. In addition, TP53 mutational analysis detected an identical missense mutation (p.T211I) between the two intestinal cancers. Moreover, further molecular genetic work-up revealed that both small intestinal and colon adenocarcinomas harbored an identical missense mutation (p.G12D) of KRAS gene. In conclusion, the small intestinal cancer in this case was identified as a metastatic adenocarcinoma arising from a past colon cancer. Conclusions Genetic analyses help in clarifying the identity of the cells in multiple cancer cases. In morphologically indeterminate cases, molecular analysis of common cancer-related genes can be useful for a precise and reproducible diagnosis.


Background
Histology of cancer cells shows cell differentiation and the neoplastic process. Accordingly, unique tumor morphology, which shows its histopathological type and expected tumorigenesis, is a diagnostic tool to identify its primary site. For example, colorectal cancer is generally classified as adenocarcinoma NOS (not otherwise specified) because it resembles normal intestinal crypts or conventional colonic adenoma [1]. However, the colonic or enteric subtype is also found in other tumor classifications, including head and neck, lung, and urinary tract cancers [2][3][4]. Consequently, the intestinal phenotype in cancer does not always originate from the lower gastrointestinal tract. In addition, distinguishing whether multiple colonic adenocarcinomas developed from single or multicentric tumor-initiating cells can be a diagnostic challenge. Although cancer predispositions such as genetic and inflammatory factors accelerate multicentric tumor formation [5][6][7][8], these clues are sometimes hidden in the practical diagnostic setting.
Herein, we report a case of adenocarcinoma in the small intestine diagnosed with immunohistochemical and genetic analyses, which also clarified the relationship of this adenocarcinoma with past breast and colon cancers.

Clinical history
An 83-year-old woman presented with severe abdominal pain and nausea at the emergency department of the hospital. She had undergone sigmoidectomy, followed by total mastectomy of the left breast 2 years ago. Pathological examination revealed that each lesion was a primary cancer; the colon cancer was a moderately differentiated adenocarcinoma (pT4aN0M0), whereas the breast cancer was an invasive ductal carcinoma with apocrine differentiation (pT2N1M0). After mastectomy, she received follow-up care, which included six cycles of adjuvant chemotherapy consisting of cyclophosphamide, methotrexate, and fluorouracil. In the emergency room, she was treated with scopolamine butylbromide because abdominal computed tomography (CT) showed mild dilation of the small intestine (Fig. 1a, b); she went home showing no symptoms. The next day, she returned to the hospital with relapse of the abdominal symptoms. The in-house radiological department noticed that the previous CT images showed an obstructed ileus arising from the nodule detected on a 18 Ffluorodeoxyglucose positron emission tomography scan 3 months ago (Fig. 1c, d). No postoperative adhesion or constriction seemed to be related to the bowel obstruction. Radiological findings and history led to the diagnosis of small bowel obstruction due to the mural nodule, which probably recurred from the colon or breast cancer. Subsequently, she was admitted to the digestive surgery department and received laparoscopy-assisted small bowel resection.

Pathological findings
Macroscopically, the resected small intestine was found to contain an ulcerated tumor (Fig. 2a), which was located 170 cm from the ligament of Treitz. Slices of the tumor suggested that the estimated tumor depth was up to the serosal surface of the intestinal wall (Fig. 2b). Histologically, infiltrative growth of colonic dysplastic glands was observed (Fig. 2c,  d). Immunohistochemically, tumor cells were diffusely positive for tumor protein 53 (p53) (Fig. 2e, f), caudal-type homeobox 2 (CDX2) and special AT-rich sequence-binding protein 2 (SATB2) (Fig. 3), positive for cytokeratin 20 (CK20) (Fig. 3) and negative for cytokeratin 7 (CK7), androgen receptor (AR) (Fig. 3), gross cystic disease fluid protein 15 (GCDFP-15), estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor 2 (HER2). Collectively, these findings were indicative of intestinal rather than mammary gland differentiation of the tumor cells. Thus, this lesion seemed to be compatible with metastatic colon cancer, albeit its gross and histological appearance mimicking primary small intestinal cancer.
To investigate the origin of the cancer cells, we reviewed the preparation of the past surgical specimens. The immunohistochemical findings of the small intestinal, colon, and breast cancer are summarized in Table 1.
As a similar abnormal p53 immunophenotype was found between the small intestinal cancer (Fig. 2e, f) and the colon cancer ( Fig. 5e, f), we analyzed the TP53 mutation status of these two tumors by direct sequencing, as described previously with minor modifications [9][10][11]. Consistent with the immunohistochemical findings, both cancers harbored an identical missense mutation, which was located on the codon 211 of the TP53 gene (Fig. 6). Therefore, we concluded that the small intestinal cancer in the present case was a metastatic adenocarcinoma arising from a past colon cancer.
Additional molecular tests were then performed in order to check the status of colon cancer biomarkers in the relapsed lesion. A PCR-based RAS/BRAF genetic test revealed KRAS G12D mutation in the small intestinal tumor, whereas the microsatellite instability test rendered a negative result. Owing to these results, we decided to investigate the genetic status of the KRAS gene in the three cancers by direct sequencing [12]. Consistent with the previous molecular findings, both the small intestinal and colon cancer specimens harbored the G12D mutation, whereas the breast cancer specimen only harbored wild type alleles (Fig. 6).
The patient is alive and under watchful waiting, 18 months after the last surgery.

Discussion
Multiple cancers that are histologically similar can be a diagnostic problem, regardless of the detection time. Similarly, multiple advanced cancers in the same organ and/or system can make pathological examination difficult. Such morphological and anatomical similarities sometimes conceal the origin of the tumor cells. Fundamentally, confirmation of the identity of multiple cancers, including precise tumor stage and pathogenesis, is important to not only satisfy scientific interest but also provide practical information for future therapeutic Macroscopically, the small intestinal tumor in the present case was an ulcer-like lesion; histologically, it appeared as a conventional colonic adenocarcinoma. Although these findings were consistent with those of primary small intestinal cancer, its p53 immunophenotype and TP53 mutational type were identical to those of the past colon cancer. Thus, we confirmed that these two tumors are metachronous intestinal cancer lesions, but they originated from identical neoplastic clones.
P53, which is encoded by the TP53 gene, is associated with one of the ten canonical oncogenic signaling pathways [13]. The p53 pathway plays an important role in cell survival, proliferation, senescence, and apoptosis. Recently, a comprehensive genome analysis revealed that the TP53 gene mutation is the most frequent genetic alteration across various cancer types [14]. It is found in approximately 60% of colon cancers [13,15]. Notably, the known genetic alteration in the TP53 gene consists of several hotspots and a myriad of minor sequence variants [16,17]. This suggests that the occurrence of TP53 somatic mutations in multiple cancers is a potent genetic signature of the identical cancer clone.
To understand the relationship between the aberrant p53 expression and TP53 mutation, a high-grade serous carcinoma of the female genital tract was studied as a representative cancer model [18]. This kind of malignancy almost always shows an aberrant p53 expression, consistent with that of the mutated type of the TP53 gene. Diffusely positive p53 expression principally  corresponds to a missense mutation of the DNA-binding domains (exon 4-8) in the TP53 gene, whereas diffusely negative p53 expression largely matches the frameshift, nonsense, and splicing site mutations of the TP53 gene [19]. In addition, the third minor immunophenotype, cytoplastic p53 expression, probably occurs due to a mutation in the nuclear localization site. Collectively, the aberrant p53 immunophenotype patterns precisely predict TP53 mutation and its mutational type. p53 immunohistochemistry is an effective tool to distinguish TP53-mutated tumors; hence, an identical p53 pattern in the two tumors of the individual indicates probable monoclonal tumor origin [20]. However, the aberrant pattern of p53 expression is the only surrogate marker for the TP53 mutational test, and the variety of the aberrant pattern is limited. Therefore, confirmation of TP53 sequence analysis is desirable, and we successfully demonstrated the genetic link between the two intestinal cancers through the identical TP53 mutational pattern.
At present, therapeutic planning for colon cancer requires the status of several established predictive biomarkers, including, RAS genes, BRAF, microsatellite instability [1]. Cancers with mutation of RAS and BRAF genes were found to be resistant to anti-EGFR (epidermal growth factor receptor) therapy [21,22]. In contrast, cancers with microsatellite instability respond to anti-PDL1 therapy [23,24]. Therefore, confirmation of these statuses is essential to select optimal molecular therapeutic agents, especially in advanced and/or relapsed colon cancer. In the present case, small intestinal and colon cancer specimens harbored identical KRAS mutations, suggesting that finding aberrant predictive biomarkers is also a potent diagnostic strategy to determine whether the multiple cancers derived from a single clone.
Apart from the genetic findings, immunohistochemical analysis is also useful to detect the origin of the cancer. Classically, a combination of the cytokeratins 7/20 expression has been used for assessing cancers with uncertain primary site. Most colon cancers are CK20 positive and CK7 negative, whereas most breast cancers are CK20 negative and CK7 positive [25]. Interestingly, small intestinal cancers frequently express CK7 and lack CK20 [26], despite the intestinal-type morphology. Although this aberrant immunophenotype may help in predicting the origin of the intestinal-type adenocarcinoma, presence of atypical CK7 positive and/or CK20 negative patterns are also observed in approximately a quarter of mismatch repair deficient colon cancers [27].
Alternatively, the expression of lineage-specific transcriptional factors and biomarkers help to presume cellular differentiation in cancer. A homeobox protein, CDX2, also known as a representative regulator of intestinal differentiation [28,29] is a sensitive and specific marker of colorectal adenocarcinoma [30]. However, high grade, mucinous and/or mismatch repair deficient colonic adenocarcinomas are associated with negative CDX2 expression [27,30], which is a prognostic factor of colon cancer without metastatic lesions [31]. Similarly, a lack of CDX2 expression is sometimes observed in small intestinal adenocarcinomas [32]. Therefore, in cases with a low CDX2 expression in intestinal-type adenocarcinomas, we should be careful about the uncommon situation mentioned above.
Recently, SATB2 is emerging as a next-generation marker for gastrointestinal tract differentiation. In addition, this nuclear matrix protein is also a marker for osteoblastic differentiation [33] because of its ability to induce skeletogenesis [34]. The expression of SATB2 is more specific to adenocarcinoma of the lower gastrointestinal tract origin compared to the expression of CDX2 [35]. However, reduced SATB2 expression (similar to that of CDX2) has been reported in mismatch repair deficient colon adenocarcinomas [36] and small intestinal cancers [37]. Taking these evidences into consideration, it is difficult to completely distinguish between colon and small intestinal cancers solely by the expression statuses of CDX2 and/or SATB2.
In contrast, AR is an emerging biomarker for prostate cancers [38], salivary duct carcinomas [39], and breast cancers [40]. Expression of this male hormone receptor in these cancers arise from AR gene dysregulation, including mutation, amplification, and alternative splicing. Consequently, an abnormal AR expression leads to cancer cell proliferation [40] even in androgen depleted states [38]. Expression of AR is significantly associated with apocrine differentiation of salivary duct carcinomas [41] and breast cancers [42] indicating that AR is a surrogate marker for these histological types [43,44], and possess a promising therapeutic target [45].
Colorectal cancer is a leading lethal malignancy, and the most common type of cancer occurs in the gastrointestinal tract [46]. In contrast, small intestinal cancer comprises only a small fraction of human neoplasia [47]. Furthermore, metastatic lesions in the small intestine, especially the distal location, outnumber primary small intestinal cancer [1]. Considering these facts, we cannot confirm the diagnosis of primary small intestinal cancer, until the possibility of a metastatic lesion from another anatomical site is ruled out. This clinical information may help the physician to decide whether the small intestinal lesion is truly a primary cancer. In addition, gross and microscopic findings of the tumor are clues to Fig. 6 Genetic analyses of the present case. A DNA sequence analysis of TP53 exon 6 (upper panels) and KRAS exon 2 (lower panels) from the small intestinal, colon, and breast carcinoma lesions and normal intestinal tissue (non-tumor). In the sequences of colon and small intestinal carcinomas, missense mutations were detected at c.632C > T in the TP53 gene and at c.35G > A in the KRAS gene its origin. However, a single clinical or morphological feature does not enable us to determine the identity of multiple neoplastic lesions definitively [48], as in the present case. To understand the evolutional history of cancer in individual cases, we propose that a molecular test must be conducted during each pathological examination. We believe that the molecular signatures, which consist of the genomic alterations, could properly confirm the identity of multiple cancers.
In conclusion, mutation analysis is a potent diagnostic tool to identify whether a tumor specimen is primary or secondary, regardless of the morphological features. Currently, cancer genetic analyses using next-generation sequencing are essential to find actionable molecular targets. In addition to their use in therapeutic strategies, the cancer genomic data indicate traceable molecular signatures that identify cancer cells. In the future, genomic findings could assist in the pathological diagnosis of morphologically indeterminate cases.