Anorectal melanoma, the third most common site of mucosal melanoma, is a rare tumor with an incidence rate of 1.7 cases per million in the global population (1). Surgery, including abdominoperineal resection (APR) or wide local excision (WLE), is the primary treatment strategy for localized disease; however, these two surgical approaches result in similarly poor clinical outcomes, with five-year recurrence rates of 79 and 81% following APR and WLE, respectively (2). Furthermore, the five-year overall survival rate following curative resection ranges from 15 to 22%, with the majority of patients succumbing to recurrent disease. Recurrence predominantly occurs in the liver, followed by the lung and brain (3,4).

Gastrointestinal stromal tumors (GISTs), characterized by the expression of cluster of differentiation (CD) 117 and a high incidence of KIT or platelet-derived growth factor receptor α (PDGFRA) mutations, are uncommon tumors with an incidence range of 6–15 cases per million individuals (5). The majority of GISTs arise in the stomach and small intestine, and GISTs originating from the esophagus, colon and extra-gastrointestinal organs, such as the retroperitoneum and liver, are rare (6). To the best of our knowledge, only five cases of primary hepatic GIST have previously been reported in the English literature (7–11). The most common KIT mutation sites involve exon 11 (66.1%) followed by exon 9 (13%); however, 12% of cases exhibit no detectable KIT mutation and are therefore categorized as KIT wild-type GISTs (12).

The present study reports the case of a long-term survivor of anorectal melanoma who, seven years after APR surgery, developed a solitary rapidly growing tumor in the liver.

A 58-year-old Taiwanese male was diagnosed with malignant melanoma of the anorectum and underwent APR at Kaohsiung Medical University Hospital (Kaosiung, Taiwan) in 2001. Microscopically, the anorectal tumor was composed of spindle-shaped neoplastic cells beneath the squamous epithelium with brownish pigments and adjacent junctional activity. Furthermore, the tumor was immunohistochemically positive for human melanoma black (HMB-45) and S-100 (Fig. 2A–D). No surgical margin or lymph node involvement was apparent, therefore, stage I anorectal melanoma was diagnosed. The patient did not receive any adjuvant therapy and was regularly followed up by physical examination every two weeks at Kaohsiung Medical University Hospital for five years after surgery.

In 2009, after a further two years and at the age of 65 years, the patient presented to the Department of General Surgery at Kaohsiung Medical University Hospital with a two-month history of general malaise, loss of appetite and progressively worsening dull pain over the epigastrium. An appropriate physical examination revealed no abnormalities, with the exception of the APR surgical scar on the abdominal wall. However, a follow-up contrast-enhanced computed tomography (CT) scan revealed a 12×11-cm solitary tumor within the left lobe of the liver (Fig. 1A). Serum α-fetoprotein, carcinoembryonic antigen and carbohydrate antigen 19-9 levels were all within normal limits. Subsequently, a percutaneous CT-guided biopsy of the liver tumor was performed. Microscopically, the liver tumor was composed of spindle cells arranged in interlacing fascicles with hyperchromatic, round or oval to spindle pleomorphic nuclei. Immunohistochemical (IHC) staining revealed that the neoplastic cells were diffusely and strongly positive for CD117, but negative for HMB-45, S-100 melan-A and CD34. Furthermore, the mitotic activity of the tumor cells was five mitoses per 10 high-power fields. Retrospectively, positive CD117 staining was also observed in the anorectal melanoma and the basal layer squamous epithelium. The results of the histological and IHC staining of the anorectal melanoma are shown in Fig. 2A–D, and the liver tumor in Fig. 2F–I. In addition, polymerase chain reaction followed by direct sequencing was performed using DNA extracted from formalin-fixed, paraffin-embedded (FFPE) tissue, revealing that the anorectal melanoma and hepatic tumor were negative for BRAF^V600E (data not shown).

To confirm the diagnosis, transmission electron microscopy (TEM) of the FFPE-derived tumors was performed to determine the ultrastructure of the two tumors. Analysis of the anorectal tumor showed a number of electron-dense granules around the nucleus, which were considered to be atypical melanosomes, and inconspicuous microtubules within the cisternal space of the cytoplasm (Fig. 2E). These features were considered to correspond with the diagnosis of melanoma. Due to damage of the fine structures of the small paraffin-embedded tissue during formalin fixation and subsequent xylene treatment, artifacts were present in the sample and a collapsed cytoplasmic structure was observed. Thus, the quality of the liver TEM image was poor. Therefore, no villous cytoplasmic processes or dispersed intermediate filaments, which are considered to be characteristic ultrastructure features of GISTs, could be identified (13). However, hyperchromatic nuclei and a vesicular cytoplasm without melanosomes were present (Fig. 2J), indicating that the liver tumor may have a different histology from the prior melanoma.

Despite testing negative for KIT and PDGFRA mutations, the liver tumor was diagnosed as primary GIST of the liver based on its IHC profiling (diffusely strongly CD117-positive, and negative for HMB-45 and S-100), and TEM features when compared with anorectal melanoma, as well as the lack of other intra-abdominal tumors identified on abdominal CT imaging. Due to the presence of left portal vein invasion, the hepatic GIST was considered as unresectable, therefore, oral imatinib was administered at a dose of 400 mg/day for two months. However, a follow-up CT scan after two months of imatinib therapy demonstrated that the tumor had progressed to 22×13 cm in size (Fig. 1B). While awaiting the approval of sunitinib from regulatory authorities, systemic chemotherapy consisting of 50 mg/m² doxorubicin and 50 mg/m² cisplatin was administered every three weeks for five months. This regimen resulted in symptomatic improvement manifesting as a decrease in dull epigastric pain and tumor size by palpation after the first cycle of treatment. Thus, this chemotherapeutic treatment strategy was continued. A follow-up CT scan, conducted two months later, demonstrated that the tumor had stabilized with minimal regression to a size of 19×11 cm (Fig. 1C). The chemotherapy was continued for an additional three months, substituting with 37.5 mg/day sunitinib after disease progression. However, the patient did not respond to sunitinib and succumbed to the disease one month later.

The current study describes the case of a unique patient who was found to have a solitary hypervascular liver tumor that was subsequently diagnosed as a second primary GIST of the liver by IHC. The patient had undergone APR for a stage I anorectal melanoma seven years prior. Clinically, it is not possible to distinguish between a metastatic melanoma and a metastatic or primary GIST in the liver, as the two types of tumor can be hypervascular in the early arterial phase of a CT scan, and grow rapidly. Considering the history of anorectal melanoma and the absence of a KIT/PDGFRA mutation during genotyping, the diagnosis of GIST in the present patient may be disputable. However, if the patient did not have a history of mucosal melanoma, the tumor could be easily diagnosed as GIST based on the results of the CD117 staining, irrespective of its KIT/PDGFRA genotyping. Therefore, a final diagnosis of hepatic GIST was determined, based on the diffusely strong CD117 staining, negativity for HMB-45 and S100 (two positive markers noted in his anorectal melanoma) and the difference in ultrastructure between the two tumors, as noted by performing TEM.

Initially, the diagnosis of GIST was considered to be more favorable for the current patient compared with those exhibiting metastatic melanoma, with disease-specific median survival periods of 19.0 and 13.3 months, respectively (14). However, following the approval of imatinib in 2002, results indicated that metastatic or unresectable GIST was more treatable compared with metastatic melanoma, with a notable increase in survival to 38.4–60 months (according to the kinase genotype) (15). Imatinib markedly improved the survival of GIST patients, regardless of the KIT/PDGFR mutation status. Despite imatinib typically producing a poorer response in wild-type GIST compared with exon 11-mutated GIST, imatinib does provide a benefit to patients with wild-type GIST, exhibiting a disease control rate of >60% (15). Furthermore, in patients with wild-type GIST, Asian populations appear more likely to benefit from imatinib treatment compared with Western populations. In the CALGB 150105 study, clinical benefit rates were 78–95% in the Asian population and 64% in the Western population (15). Additionally, five-year survival rates of 60–70% and 40–45% were noted for Asian and Western populations, respectively (15–17). Although the differences may have been caused by the small-size bias of Asian studies, the possibility of higher pharmacokinetic parameters in Asian populations following fixed-dose imatinib treatment should also be considered. Despite the increased benefit in Asian populations, the current patient did not receive any benefit from imatinib treatment and the tumor progressed rapidly within two months, which indicates that the nature of primary hepatic GIST may be different from those arising in the alimentary tract.

To the best of our knowledge, to date, only five cases of primary hepatic GIST have been previously reported (Table I) (7–11). Among these five cases, two cases received imatinib therapy (8,9). One of the imatinib-treated cases lacked mutation analysis data, however, the patient exhibited a good response to imatinib, resulting in a disease-free status for three years (8). The other imatinib-treated case had no mutation of KIT exon 11 (only KIT exon 11 was analyzed) and responded well to imatinib, with a long-term survival period of nine years (9). As only a small number of cases of primary hepatic GIST have been reported thus far, the correlation between the KIT genotype and response to imatinib, a tyrosine kinase inhibitor, remains inconclusive; thus, additional studies are required to address the issue.

In conclusion, despite the poor response to imatinib of the primary hepatic KIT/PDGFRA wild-type GIST in the current patient, the present study highlights the importance of considering a second primary malignancy and performing immunohistochemical analysis upon the occurrence of a newly developed lesions in long-term remission cancer survivors. With the emergence of modern molecular technologies, such as next-generation sequencing, it may be possible to precisely determine genetic alterations that exist between the two tumors. This may aid in determining the nature of the second tumor (delayed recurrence versus second primary), as well as providing useful information to guide therapeutic treatment of the disease.