Diagnosis of primary biliary melanoma with distinct imaging features: a case report and literature review

Abstract

Primary biliary melanoma arises from proliferating melanocytes in the mucosal surface of the bile duct and is extremely rare. Since the vast majority of biliary melanomas represent metastases of cutaneous origin, accurate preoperative diagnosis of melanoma and exclusion of other primary sources are vital in cases involving primary lesions. Although melanomas with pigmented cells have typical signal characteristics, obtaining a non-invasive pre-treatment diagnosis remains difficult, due to their low incidence. Here, the case of a 61-year-old male Asian patient who presented with upper quadrant abdominal pain, swelling and jaundice for 2 weeks, and who was diagnosed with primary biliary melanoma following extensive preoperative blood analyses, computed tomography (CT) and magnetic resonance imaging (MRI), is described. Post-resection immunohistochemistry confirmed the diagnosis and the patient received six chemotherapy cycles of temozolomide and cisplatin, however, progression of multiple liver metastases was observed at the 18-month follow-up CT. The patient continued with pembrolizumab and died 17 months later. The present case of primary biliary melanoma is the first reported diagnosis based on typical MRI features and the exhaustive exclusion of a separate primary origin.

Introduction

Melanoma arises from proliferating melanocytes, a heterogeneous group of cells that are derived from the embryonic neural crest and specialized in synthesizing melanin.¹ The overwhelming majority of melanocytes are restricted to cutaneous sites, with small numbers scattered on mucous surfaces (nasal, oropharyngeal, respiratory, alimentary and genitourinary tract), the ocular region (uvea and retina) and the leptomeninges.² Based on the anatomic location, melanomas are clinically subdivided into several subtypes: cutaneous, acral, mucosal, uveal, and unknown.³ Mucosal melanoma is the rarest subtype, occupying only 0.8–3.7% of melanoma cases.⁴ Primary biliary melanoma arises from melanocytes in mucosal tissues lining the bile duct and is extremely rare.⁵

Primary biliary melanoma is a tumour with high-risk for recurrence and poor prognosis, thus, early diagnosis and appropriate therapies are urgently required. Previous studies have revealed that melanin synthesized by melanocytes may shorten T1 and T2 relaxation time, forming a distinctive appearance on magnetic resonance imaging (MRI).^6,7 This appearance combined with preoperative whole-body screening renders non-invasive diagnosis of primary biliary melanoma possible, which plays a critical role in assisting the therapeutic regimen formulation. Nevertheless, to date, there are no reports of a case that has been diagnosed prior to treatment, possibly due to the rarity of the disease. Here, the case of a patient with suggested preoperative diagnosis of primary biliary melanoma, evidenced by preoperative typical MRI features and whole-body screening, together with imaging features, histological characteristics, immunohistochemical profile and curative procedure, is described.

This study was approved by the Institutional Review Board of the Affiliated Hospital of Nantong University and written informed consent was obtained from the patient. The reporting of this study conforms to CARE guidelines.⁸

Case presentation

A 61-year-old Asian man (without any prior disease, special medication or drug allergy, or adverse life history) was admitted to the Gastrointestinal Department of the Affiliated Hospital of Nantong University, Jiangsu Province, China in July 2015, with upper quadrant abdominal swelling, pain and jaundice for approximately 2 weeks. On admission, physical examination of the abdomen revealed mild tenderness in the right upper quadrant, without rebound tenderness. Results of blood analysis indicated cholestasis and hepatic cell injury, evidenced by elevated total bilirubin (319.2 μmol/L), direct bilirubin (231.2 μmol/L), total bile acid (115.6 μmol/L), alkaline phosphatase (436 U/L) and gamma-glutamyl transferase (1509 U/L), representative of cholestasis; and raised aspartate aminotransferase (124 U/L) and alanine aminotransferase (307 U/L), reflective of hepatic cell injury. The value of tumour marker serum ferritin was elevated, at 860.7 ng/ml (normal range, 21.80–274.66), but values of other biomarkers, such as carcinoembryonic antigen (1.9 ng/ml) and carbohydrate antigen 19-9 (2.0 U/ml), were within normal range. Computed tomography (CT) imaging revealed a quasi-circular nodule of possible biliary origin, with slightly hypodense appearance on plain scan compared with adjacent muscle (Figure 1a), marked and mild heterogeneous enhancement in the arterial phase (Figure 1b), and uniform slightly reduced enhancement in the delayed phase (Figure 1c). The lesion was about 2.5 cm in diameter and located in the distal common bile duct, causing extrahepatic biliary dilatation (Figure 1d). On MRI (Figure 2), the MR cholangiopancreatography (MRCP) image (Figure 2a) clearly showed a space-occupying lesion with smooth margin in the distal common bile duct causing biliary dilatation. The lesion was hyperintense on T1-weighted imaging without signal reduction when fat was suppressed (Figure 2b), slightly hyperintense on T2-weighted imaging (Figure 2c), and hyperintense on diffusion-weighted imaging (Figure 2d) compared with the adjacent muscular tissue. After intravenous administration of gadolinium contrast agent, the lesion demonstrated slight heterogeneous enhancement during the arterial phase (Figure 2e) and progressive and more homogeneous enhancement surrounded by discontinuous enhanced biliary wall during the equilibrium phase (Figure 2f). Based on the particular signal characteristics of the entity, a senior diagnostician (JW) expressed a high suspicion of a melanoma. To rule out an alternative primary site, detailed ophthalmological, otorhinolaryngologic and dermatological examinations with CT enhancement of the neck, chest and pelvis were performed, and no other suspicious lesion was revealed. MRI did not detect any aberrant cerebral or ocular entity, and no mucosal lesions suggestive of melanoma were revealed by gastrointestinal endoscopy. The patient reported no previous sunburns, prolonged sun exposure, atypical nevus, or excised skin cancer, and no family history of skin cancers.

Figure 1.

Computed tomography (CT) images from a 61-year-old Asian male patient diagnosed with primary biliary melanoma: (a) CT plain scan image showing a round, slightly hypodense lesion (arrow) in the common bile duct; (b and c) arterial and delayed-phase CT scan images after injection of contrast agent, demonstrating arterial marked and mild heterogeneous enhancement, and uniform slightly reduced later enhancement; and (d) coronal reconstruction image showing a distal biliary lesion (arrow) causing extrahepatic bile duct dilatation.

Figure 2.

Axial magnetic resonance images from a 61-year old male Asian patient diagnosed with primary biliary melanoma: (a) magnetic resonance cholangiopancreatography image clearly showing a lesion (arrow) with smooth margin in the distal common bile duct causing biliary dilatation; (b) a T1-weighted fat-suppressed sequence image showing a biliary lesion (arrow) that is hyperintense versus the adjacent muscle; (c) a T2-weighted image showing the lesion (arrow) as slightly hyperintense versus the adjacent muscle; (d) diffusion weighted image showing the lesion (arrow) as hyperintense versus the adjacent muscle; and (e and f) arterial and delayed-phase dynamic contrast-enhanced fat-saturated T1-weighted images demonstrating arterial slight heterogeneous enhancement and later progressive and more homogeneous enhancement.

At 5 days after admission, the patient provided written informed consent to surgery and pylorus preserving pancreaticoduodenectomy was performed. The digestive tract was reconstructed using Child’s method, before which the patient had undergone percutaneous transhepatic cholangial drainage to reduce jaundice. The operation lasted 3 h with approximately 400 ml of blood loss. The tumour was completely resected with free surgical margins, and no evidence of regional lymph node metastasis was found in the six intraoperatively retrieved nodes. After surgery, the patient had an uneventful postoperative course. Macrography of the postoperative specimen disclosed a solitary, polypoid, 3 × 2.5-cm, firm, tan-red, pedunculated lesion located in the bile duct. Microscopically (Figure 3a), the lesion consisted of tumour cells of different sizes and dysplasia, with large, hyperchromatic and pleomorphic nuclei, and scattered pigmented cells. Immunohistochemical examination demonstrated tumour positivity for S100 protein, premelanosome protein (HMB-45) and melan-A (Figure 3b, c, and d), and negativity for pan cytokeratin (CK10, CK14, CK15, CK16, CK19 and Keratin 2; AE1/AE3 antibody clone), keratin 18 (CK18), synapsin (Syn), chromogranin A (CgA), leukocyte common antigen (LCA), desmin and alpha-smooth muscle actin. Supported by the histopathologic results and preoperative exhaustive exclusion of an alternative primary origin, a diagnosis of primary biliary melanoma was confirmed. The patient declined genetic analysis.

Figure 3.

Representative tumour histopathology photomicrographs from a 61-year old male Asian patient diagnosed with primary biliary melanoma: (a) haematoxylin and eosin-stained section showing pleomorphic malignant cells and some with pigmentation; (b) strong and diffuse positive staining for S-100 protein; (c) strong and diffuse positive staining for HMB-45; and (d) strong and diffuse positive staining for melan-A (all original magnification, × 200).

Postoperative chemotherapy was initiated, comprising six cycles of 200 mg/m²/day temozolomide (orally, days 1–5) plus 25 mg/m²/day cisplatin (intravenously, days 1–3), every 21 days. Follow-up CT re-examination 18 months later demonstrated progression of multiple liver metastases. From that time on, the patient was prescribed 2 mg/kg pembrolizumab, intravenously, every 3 weeks until the time of death, 17 months later.

Discussion

Primary melanoma in the common bile duct, first described by Zaide in 1963,⁹ is particularly rare. A review of the published literature searched in PubMed, using a series of search strings (‘biliary’ OR ‘bile duct’ AND ‘primary’ AND ‘melanoma’) with no date limit, revealed that only 15 cases, including the present case, have been reported to date (summarised in Table 1).^10–21 This tumour type appears to have a predilection for adult males (male: female ratio, 13:2) and a mean patient age at presentation of 47.9 years (range, 26–67; Table 1). The predominant clinical symptom was found to be jaundice, with the next most commonly reported symptoms being abdominal pain and dark urine. Almost all of the cases displayed elevated serum total bilirubin, representing cholestasis, and for some cases, increased alanine aminotransferase was reported, sensitively reflecting hepatic cell injury.

Table 1.

Summary of previously published reports of primary biliary melanoma.

Publication	Age/sex	Symptoms	Location	Therapy	Outcome
Zaide, 19639	47/male	Jaundice	Proximal common hepatic duct	Cholecystectomy and resection of common bile duct	Disease-free survival 6 months
Carstens et al., 198610	30/male	Indigestion, abdominal pain, and jaundice	Distal common bile duct	Whipple resection	Progressive disease with death 6 months after presentation
Deugnier et al., 199111	34/female	Abdominal pain and jaundice	Left hepatic duct	Left hepatectomy with partial right hepatic and common bile duct resection	Progressive disease with death 1.5 years after presentation
Zhang et al., 199112	58/male	Jaundice	Common bile duct and gallbladder	Whipple resection	Disease-free 2 years after presentation
Washburn et al., 199513	45/male	Month of itching and dark urine	Distal common bile duct and ampulla	Whipple resection	Disease-free survival 6 years after presentation
Washburn et al., 199513	43/male	Lethargy, itching, and dark urine	Common hepatic duct and right hepatic duct	Right hepatic lobectomy and cholecystectomy	Alive 11 months after presentation
Wagner et al., 200014	48/male	Right upper quadrant pain and jaundice	Common bile duct	Whipple resection	Progressive disease with death 9 months after diagnosis
González et al., 200115	67/female	Jaundice and abdominal pain	Proximal common hepatic duct	Extrahepatic bile duct resection	Disease-free survival 17 years after presentation
Bejarano et al., 200516	47/male	Jaundice and dark urine	Distal common bile duct	No surgery performed due to metastatic disease	Died 2 years after presentation
Hoshi et al., 200617	55/male	Jaundice	Common hepatic duct	Bile duct resection	Died 4 months after presentation
Agrawal et al., 201018	26/male	Abdominal pain	Intrahepatic biliary tract	Hepatic segmentectomy (II–III)	Disease-free survival 72 months after presentation
Smith et al., 201219	55/male	Jaundice	Distal common bile duct	Whipple resection	Disease-free survival 12 months after presentation
Addepally et al., 201620	52/male	Jaundice and abdominal pain	Common hepatic duct	Best care support	Died 3 months after presentation
Cameselle-García et al., 201921	51/male	Jaundice	Common hepatic duct	Cholecystectomy and common hepatic duct resection	Died 31 months after presentation
Present case	61/male	Upper quadrant abdominal pain and jaundice	Distal common bile duct	Pylorus preserving pancreaticoduodenectomy	Died 35 months after diagnosis

Because the vast majority of biliary melanomas represent metastases from cutaneous origin, suggestive diagnosis of melanoma and pretreatment exclusion of a different primary source is extremely significant for primary entities, particularly in patients suitable for thorough resection, as in the present case. CT is the preferred approach with the advantage of rapidly imaging multiple organs and systems, directly visualizing enlarged lymph nodes, estimating tumour bulk and better characterizing its relationship with neighbouring structures. Nevertheless, CT images do not enable further analysis of the distinct component of the melanoma. MRI may sometimes provide a suggestive non-invasive diagnosis, relying on paramagnetic metal bound to melanin, which has the characteristic of influencing T1 and T2 relaxation time and contributes to the typical hyperintensity on T1-weighted imaging and hypointensity on T2-weighted imaging.²² In the present case, since the entity appeared uniformly hyperintense in the T1 fat-suppressed sequence compared with adjacent muscle, it was first necessary to identify six elements, including gadolinium contrast agents, haemoglobin degradation products, fats, high-protein substances, melanin and minerals. Fat-suppressed images may be explicit in the case of fatty lesions. The possibility of the most common cholelithiasis, haemoglobin degradation products, or high-protein substances may be ruled out by diffusion-weighted imaging and contrast enhancement. On the basis of the exclusion method, the lesion would likely be a melanin-containing tumour, including melanoma or pigmented perivascular epithelioid cell tumour (PEComa). These two tumours rarely occur in the common bile duct, and PEComa is even rarer. In addition, differential diagnosis of the imaging manifestations of these two tumours is not well known, however, PEComa tends to display a female predominance of 7:1, and haematoxylin and eosin (H&E) staining usually reveals epithelioid cells arranged in a nested, alveolar, or trabecular pattern near blood vessels, without scattered pigmented cells.²³ Importantly, a suggestive diagnosis of melanoma is helpful for clinical treatment. A mass-forming type of cholangiocarcinoma is another common tumour that needs to be distinguished. However, not only the signal intensity of the present tumour, but also the pattern of enhancement, was inconsistent, with mainly moderate peripheral enhancement followed by progressive centripetal enhancement. It is relatively easy to note that the hypersignal on T1-weighted imaging was not particularly pronounced in the present case, as it appeared isointense compared with the pancreas. Melanin content is shown to correlate with hyperintensity on T1-weighted imaging, meaning that the signal intensity of melanoma may be determined by the amount of melanin within the lesion.²⁴ We hypothesized that there was not much melanin in the present case, which would explain why the T2-weighted imaging showed a slightly higher signal rather than the typical low signal. Based on the pre-treatment suggestive diagnosis, exhaustive screening with particular attention to the skin, eyes, anorectum and vagina, and general radiological examination, could be arranged to avoid the omission of another primary melanocytic cancer and contribute to formulating an optimal therapeutic regimen.

The diagnosis of melanoma was eventually verified by histopathological outcomes including H&E staining and immumohistochemical staining. Melanomas with unique melanin-pigmented cells are typically pathological manifestations, but ancillary immunohistochemical studies are still needed to obtain the final diagnosis. Three melanocytic differentiation markers, including S100 protein, a 21 kDa acidic calcium-binding protein, HMB-45, a marker of the cytoplasmic premelanosomal glycoprotein gp100, and melan-A, an antigenic target of cytotoxic T lymphocytes, are currently the most useful immunomarkers to identify melanocytes.²⁵ The most sensitive marker, S100 protein, has demonstrated a sensitivity of >89% in diagnoses of melanoma, but a relatively low specificity of 75–87%.²⁶ Melan-A has shown a sensitivity of 75–92% and a specificity of 95–100% for melanoma, while HMB-45 displays a sensitivity of 69–93% and a specificity of 56–100%.²⁶ Meanwhile, negativity for pan cytokeratin, CK18, Syn, CgA, LCA, desmin and alpha-smooth muscle actin ruled out other tumours in the present case, and further confirmed the diagnosis of melanoma.

The primary and metastatic entities of melanoma share consistent pathological features. Several characteristics to separate primary from metastatic melanoma in the bile duct have been previously summarised.^18,19 The most important characteristic is the absence of an identifiable primary origin, which requires a thorough exclusion of current, or history of, pigmented lesions. Secondly, the majority of primary biliary melanomas demonstrate common morphological characteristics of solitary, polypoid and intraluminal mass consistent with the present case, whereas metastatic tumours are more likely to appear as multiple, multifocal and flattened lesions. Finally, some investigators consider that the most essential histologic characteristic to distinguish a primary entity from a metastatic one is a junctional component,¹⁹ meaning the presence of microscopic clustered melanocytes at the in-situ mucosal–submucosal junction. A definitive in-situ junctional component was not identified in the present case, perhaps due to a sampling error, which may have prevented the identification of an in-situ component. However, junctional components have been reported in definitive cases of metastatic melanoma, because of neoplastic intraepithelial extendedness.²⁷

Although there is no established standard of care for primary biliary melanoma, radical surgery is recommended in those patients with a resectable entity and good status, such as in the present case. In unresectable or advanced cases, biliary drainage and immunotherapy should be adopted. Targeted therapy against KIT proto-oncogene, receptor tyrosine kinase-activating molecules is a recently emerging therapeutic strategy that has shown promise.²⁸ The prognosis of primary biliary melanoma remains poor, but it is difficult to ascertain a concrete range of survival duration due to low incidence rate and lack of long-term follow-up. The present patient survived 35 months after surgical resection, so we speculate that long-term survival is possible if surgical resection is achieved.

Conclusion

In summary, a suggestive non-invasive diagnosis may be obtained in cases of melanin-containing biliary melanomas, depending on their typical MRI signal characteristics. Typical pigmented tumour cells and appropriate immunohistochemical stain may certify the diagnosis. Tumour morphology and exclusion of a separate origin may provide clues to exclude metastatic melanoma.

Data availability statement

All data supporting this study are presented within the article.

CRediT authorship contribution statement

Xinhong Wang: Image analysis, Writing-Original Draft. Xiaofei Deng: Methodology, Writing-Editing. Zheng Shu: Conceptualization, Writing-Editing. Juan Wu: Supervision, Writing-Reviewing and Editing.