Colorectal cancer (CRC) ranks among the three most commonly diagnosed malignant tumors worldwide according to the World Health Organization (WHO) (1). Importantly, CRC is a heterogeneous disease with different molecular subtypes, variable clinical course, and prognosis. Optimal treatment requires comprehensive characterization of these features. Research to further develop the histopathological, molecular, and genetic characterization of CRC therefore builds essential knowledge to improve cancer therapy and survival.

The current gold standard to define CRC prognostic groups is the tumor node metastasis (TNM) classification published by the International Union for Cancer Control (UICC) (2). Extensive efforts have been made to aid the projection of post-operative survival and treatment response based on molecular alterations (3). Adverse molecular markers include BRAF mutations found in approximately 10% of patients and the CpG-island methylator phenotype (4, 5). KRAS and NRAS mutations correlate with poor response to drugs inhibiting the epidermal growth factor receptor (EGFR) (6). Microsatellite instability (MSI) is identified in 15–20% of CRC cases and is associated with favorable prognosis due to an increased anti-tumoral host immune response (7). Approximately 80% of MSI CRC arise in the sporadic setting due to hypermethylation of the MLH1 gene, while 20% are associated with germline mutations of the mismatch-repair (MMR) genes, such as MLH1, PMS2, MSH6, or MSH2, in Lynch syndrome. Patients without an identified germline defect in a DNA MMR gene but with MSI and loss of MMR protein expression are likely to have Lynch syndrome if other causes of MSI, such as methylation of the MLH1 promoter, are excluded (8). The suggested WHO terminology for these cases is “probable Lynch syndrome” (8). However, recent studies indicate that pathogenic somatic mutations in MMR genes can also underlie the development of MSI in a subset of cases of early onset CRC (9–11). This group of cases has been termed “Lynch-like” to underline the similar clinicopathological presentation with Lynch syndrome in absence of a proven germline-cause for MMR deficiency (12).

While the clinicopathological and molecular features of sporadic and hereditary CRC have been extensively studied, little is known about the occurrence of rare histopathological variants such as colorectal choriocarcinoma. Choriocarcinoma is a highly malignant neoplasm with trophoblastic differentiation. Gestational choriocarinoma most frequently occurs as a result of a molar pregnancy in pre-menopausal women and represents the vast majority of cases (13). Non-gestational choriocarcinoma can present as a component of ovarian and testicular germ cell tumors, while non-gestational, extra-gonadal choriocarcinomas are exceedingly rare (14). Discrimination of these forms of trophoblastic malignancy is of central importance as significant differences in the genetic origin, oncogenic driver mutations, immunogenicity, and sensitivity to chemotherapy exist (15). In particular, frequent targetable molecular alterations have been described in gestational disease, including an activation of the mitogen-activated kinase pathway (MAPK) through mutations in KRAS and BRAF oncogenes, overexpression of c-MYC, EGFR mutations, and activation of the mammalian target of rapamycin (mTOR) signaling network (16).

In adenocarcinomas of the gastrointestinal tract, choriocarcinomatous differentiation can take several forms. This ranges from the presence of individual beta human gonadotropin (β-HCG) positive malignant syncytiotrophoblastic giant cells in poorly differentiated adenocarcinoma over mixed tumors to pure choriocarcinoma (14). Previous genetic analyses have highlighted a superimposed pattern of genetic changes in adeno- and choriocarcinoma components of mixed tumors suggesting a common stem cell origin (17). However, the molecular pathogenesis and presence of targetable mutations in extra-gestational disease has remained obscure. MMR deficiency and MSI in colorectal choriocarcinoma have not been previously identified. Here, we report the comprehensive morphological, immunohistochemical, and molecular analysis of a colorectal choriocarcinoma in a patient with probable Lynch syndrome.

A 61-year-old Caucasian post-menopausal female with a history of stage I signet ring cell carcinoma of the stomach at age 36 was evaluated for a change in bowel habits and abdominal discomfort. There was a known family history of breast cancer in a second degree relative (aunt). Physical examination discovered a palpable resistance in the lower right quadrant. Colonoscopy identified a large stenosing mass in the cecum. An endoscopic biopsy was performed showing a poorly differentiated carcinoma with extensive necrosis. Computed tomography (CT) imaging confirmed the diagnosis of a 9.6-cm cecal mass (Figure 1A) and identified four hepatic metastases with a maximum diameter of 2.6 cm (Figure 1B). There was no evidence of a lesion involving the ovaries, the uterus, or a locoregional recurrence of gastric carcinoma. A right hemicolectomy with primary end-to-end anastomosis was performed. A diagnosis of primary colorectal choriocarcinoma in clinical stage IV was made. Pathological tumor stage according to the UICC TNM classification, 7th edition (2) was pT4a, pN0 (0/29), cM1 (HEPAR), L0, V1, Pn0, G3, and R0. Laboratory examinations showed significantly elevated serum β-HCG levels of 70.173 IU/ml (internal reference <2 IU/ml), a normal level of carcinoembryonic antigen (CEA; 1.9 μg/l; internal reference <5.0 μg/l), and CA19.9 (6.7 kU/l; internal reference <35 kU/l). The patient was first treated with oxaliplatin, capecitabine, irinotecan, and bevacizumab (XELOXIRI/bevacizumab) for two cycles according to the standard protocols (18). Rapid radiographic progression of disease and increasing serum levels of β-HCG were recorded under treatment. The therapeutic strategy was therefore adapted toward bleomycin/etoposide/cisplatin (BEP) which has shown efficacy in high-risk gestational and extra-gestational trophoblastic neoplasia (15). A partial remission [RECIST 1.1 (19)] with radiographic evidence of a decrease in the sum of the diameters of the hepatic lesions of over 50% and a decrease in serum β-HCG levels to 8 IU/ml (internal reference range <2 IU/ml) was documented. Close follow-up was maintained.

Fifteen months later, laboratory and radiographic evidence of intra-abdominal tumor progression with new peritoneal lesions was recorded. Third line therapy with five cycles of carboplatin and bleomycin was initiated, leading to stabilization of disease (Figure 1C). New peritoneal, splenic, uterine, ovarian, and rectal lesions were noted on follow-up studies 6 weeks later. Salvage therapy with ifosfamide and vinblastine was started leading to a partial remission [RECIST 1.1 (19)]. The patient recovered well and surgical tumor debulking was performed 2 months later with resection of peritoneal lesions, the rectosigmoid, spleen, and uterus. Following recovery, two cycles of paclitaxel monotherapy were administered. Further treatment was delayed as a result of an intravascular catheter infection. There was rapid progression of hepatic disease with hilar compression. New bone metastases were detected in the spinal column. Due to rapid deterioration, the patient was entered into best supportive care and deceased 34 months after diagnosis. An autopsy was not performed.

Gross examination of the resection specimen showed a poorly marginated tumor localized to the cecum and terminal ileum with serosal perforation (Figure 2A). Histopathological analysis revealed a solid carcinoma consisting of a dimorphic population of syncytiotrophoblastic giant cells and mononucleate trophoblastic cells arranged in a plexiform pattern with transmural infiltration of the colonic wall (Figure 2B) and multifocal venous invasion. Immunohistochemical analysis demonstrated strong, diffuse cytoplasmic reactivity for pan-cytokeratin (pan-CK) and β-HCG, while stains for the intestinal transcription factor CDX2 were negative (Figure 3, top). Additional analyses of oncogenes associated with choriocarcinomatous differentiation revealed an overexpression of p53, the p53-associated protein MDM2 (not shown), c-MYC, and p16. Immunohistochemical analysis of the MMR proteins MLH1, PMS2, MSH6, and MSH2, showed an isolated loss of PMS2 protein (Figure 3, bottom right). Antibodies used for immunohistochemical analysis and staining protocols are provided in Table 1.

Based on a previous history of gastric cancer at age 36 and an isolated loss of PMS2 protein expression by immunohistochemistry, the suspicion of Lynch syndrome was raised. This was further investigated by molecular studies of both tumors (Figure 4A). MSI analysis of the colorectal tumor showed shifted alleles in seven out of the nine microsatellite markers by multiplex PCR assay and high-resolution gel electrophoresis (instable: Bat-25, Bat-26, D2S123, D17S250, Bat-40, D13S153, D18S58; stable: D5S346, D10S197) (Figure 4B). This finding was consistent with an MSI-high genotype. The tumor MLH1 promoter region was un-methylated as revealed by bisulfite sequencing analysis (Figure 4C). Sequencing of BRAF hotspot regions in codons 600 and 601 showed no mutations. Using archival material of the gastric primary from 1985, analysis of six microsatellite markers (Bat-25, Bat-26, D2S123, D5S346, D10S197, and D18S5) was successful. All markers were stable, and there was no evidence of MSI. In consistency with these findings, we found maintained expression of MLH1, PMS2, MSH6, and MSH2 in the gastric carcinoma by immunohistochemistry. No BRAF mutations were identified. MLH1 promoter methylation analysis was not performed. Germline mutation analysis of the PMS2 gene was not possible due to applicable legal regulations.

To further expand the molecular characterization of colorectal choriocarcinoma, two independent targeted next-generation sequencing (NGS) assays (Oncomine™ Focus Assay and Ion AmpliSeq™ Cancer Hotspot Panel v2, both from Thermo Fischer Scientific, Waltham, MA, USA) were performed according to the manufacturer’s instructions. The Oncomine™ Focus Assay allows the concurrent analysis of DNA and RNA for detection of mutations, copy number variations, and gene fusions in a comprehensive panel of clinically actionable oncogenic driver genes (21). The Ion AmpliSeq™ Cancer Hotspot Panel allows the identification of somatic mutations in a 50 gene panel (22). For NGS analysis, all diagnostic slides of a given case were re-reviewed and representative blocks of formalin-fixed paraffin-embedded material with a tumor cell content exceeding 80% were selected. Areas of necrosis were excluded. Microdissection was not performed. Identical analyses were carried out using archival material of the gastric primary from 1985. While no relevant molecular alterations were identified by the Oncomine™ Focus Assay in the colorectal choriocarcinoma, targeted NGS using the Ion AmpliSeq™ Cancer Hotspot Panel revealed an ataxia telangiectasia mutated (ATM p.P604S) mutation both in the primary tumor and metastatic lesions at an allele frequency of 51.48 and 57.32%, respectively. An identical ATM p.P604S mutation was identified in the gastric primary from 1985 with an allele frequency of 77.2%.

Colorectal choriocarcinoma is a rare and aggressive cancer which is disproportionately common in young patients. The identification of this subtype may aid the selection of an appropriate therapy and support the use of β-HCG as a clinically actionable biomarker. In addition, an increased suspicion for the detection of hereditary cancer syndromes may be warranted.

Publication of this study was evaluated by the local ethics committee (Ethikkommission Nordwest- und Zentralschweiz) on April 11, 2016. According to the guidelines of the committee, a designated ethics vote and consent is not required for an anonymized case report.

VK and GC designed and coordinated the study, drafted the manuscript, performed the histopathological and immunohistochemical analysis, and critically discussed molecular data with KM, UG, and DZ. KM helped with rendering the diagnosis, performed and interpreted the molecular analysis, and critically reviewed the manuscript. KS treated the patient at Radio Onkologie Allschwil, helped with rendering the diagnosis, and critically reviewed the manuscript. UG and DZ performed MLH1 promoter methylation analysis and critically reviewed the manuscript. AS performed next-generation sequencing analysis. All the authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.