Malignant melanoma arising from a perianal fistula and harbouring a BRAFgene mutation: a case report

Malignant melanomas of the anal region are a rare occurrence, accounting for only around 0.2-0.3% of all melanoma cases and 4% of anal cancers [1]. Diagnosis may be challenging, since perianal melanomas may be disguised as haemorrhoids [2] and more rarely as fissures [3] or perianal abscesses [4]. Prognosis is generally very poor [5], with five-year survival rates ranging between 6.7% and 16% [6, 7].

The BRAF gene encodes a serine/threonine kinase involved in the mitogen-activated protein kinase pathway (MAPK) [8]. Activating BRAF mutations are present in approximately 70% of cutaneous malignant melanomas [9] and 82% of melanocytic naevi [10]. More than 90% of BRAF mutations involve a single point mutation, T1799A, in codon 600 of exon 15, leading to a V600E amino acid substitution [9, 10]. However, several studies have shown that BRAF mutations are very uncommon in melanomas arising in sun-protected areas [11‐13]. These findings have suggested an association between the presence of BRAF mutations in malignant melanomas and ultraviolet (UV) light exposure.

The patient was a 38-year-old male with a history of beta-thalassaemia minor with pseudopolycythemia and microcytosis. In May 2002 he underwent surgery for a perianal fistula of four years of evolution. Pathologic examination of the resected specimen revealed an abscessed fistulous tract with the presence of a small core of hyperpigmented and hyperplastic basal melanocytes arranged irregularly and forming a lentiginous pattern (Figure 1). A nest of epithelioid cells with atypical anaplastic nuclei, prominent nucleoli and abundant melanin was clearly discernible in the underlying dermis (Figure 1). Besides, features relatively specific for melanoma including absence of maturation and presence of mitoses in deep cells were also visible (Figure 1). Finally, other secondary characteristic also suggestive of melanoma was the abundant melanic pigment throughout the lesion, both in the depth of the lesion and in the upper portion of the epidermis, including the stratum corneum or horny layer of the epidermis.

In June 2002, once the diagnosis of melanoma in perianal fistula (Clark level III) had been confirmed by the pathologist, the patient went through a surgical extension of excision margins (1 cm) and a sentinel lymph node biopsy (one right and two left inguinal nodes). Both the extended margins and the sentinel lymph nodes produced negative results for tumour cells.

In order to discard the possibility of metastatic melanoma, the patient was also subjected to an abdominal ultrasound examination and a whole body CT scan, both without relevant findings. Serum levels of S-100 protein, a widely used melanoma marker, were also negative. The patient went through his latest follow-up in December 2010, and at the time of this report he remains free of disease.

High quality sequencing was obtained in both the 5' and the 3' direction from the melanoma sample, the control DNA sample and melanoma cell line A375. The V600E mutation (T1799A) in exon 15 of the BRAF gene was present in the heterozygous state in the melanoma sample, but it was absent in non-tumour DNA obtained from peripheral blood of the same melanoma patient (Figure 2). As expected, the positive control sample, melanoma cell line A375, showed the V600E mutation in the homozygous state (Figure 2).

The development of a tumour inside a longstanding fistula is an extremely infrequent event. More frequent is the case of a tumour presenting as a fistula, that is, the fistula being a consequence of the cancerous process. Occasionally, the inflammation and necrosis caused by the development and progression of a tumour stimulates the formation of a fistula. However, there are instances in which the fistula is not a secondary manifestation of the neoplasia, and its formation predates the development of the tumour. In these cases, the neoplastic transformation arises inside the longstanding fistula, probably due to the latter's chronic inflammatory nature [14]. We have found several cases in the literature in which a variety of different tumour types have developed inside a fistula once the latter was already formed, such as squamous cell carcinomas [15‐17], basal cell carcinomas [18, 19] and mucinous adenocarcinomas [17, 20‐22].

As far as the authors are aware, the occurrence of a melanoma within a fistula has not been previously described in the literature. We have found two published cases of melanomas presenting with fistulous formations, one concerning a biliary tract fistula [23] and another one presenting as a urethral fistula [24], in which the appearances of the fistulas were a consequence of the cancerous process.

In our patient, the fistula melanoma was quite small (Figure 1) and it was found by chance while the pathologist examined the stained histological sections of the fistula. The melanoma was located in the depth of the subcutaneous tissue, next to the fistulous tract, the interior of which was already epithelialised. It was surrounded by abundant inflammatory response and next to a lesion showing atypical lentiginous melanocytic hyperplasia (Figure 3). The presence of lentiginous hyperplasia suggested that the melanocytic lesion originated inside the fistulous tract and that we were indeed in front of a primary melanoma and not a metastasis (already ruled out by the whole body CT scan). On the other hand, the opening of the fistulous tract in the perianal skin epidermis did not contain signs of inflammation or malignant tissue (Figure 3), which ruled out the possibility of the fistula being a complication or a secondary manifestation of a cutaneous melanoma [14]. Therefore, we believe that the fistula predated the melanoma, not the other way around, and that it arose as a complication of the inflammatory lesion of the fistula.

Anal melanoma is an uncommon disease that affects patients of advanced age, usually appearing in the sixth or seventh decade of life [25, 26]. However, Cagir and cols. found an increased incidence of anorectal melanomas in males aged 25 to 44 years living in the San Francisco area [27], and they proposed HIV infection as a risk factor for anorectal melanoma. But the HIV test in our patient (who was 38 years of age at diagnosis) was negative, and he had neither a previous history nor signs of other sexually-transmitted diseases.

The fact that this melanoma originated in a perianal location and inside a fistula also ruled out UV light exposure as a possible aetiological factor. Classic genetic mutations caused by UV exposure comprise C>T and CC>TT transitions, although UV exposure generates oxygen-free radicals that also damage DNA and give rise to other types of base substitutions [28]. Even though BRAF mutations occur more frequently in melanomas linked to sun exposure, the BRAF gene does not contain C>T or CC>TT transitions typical of UV-induced damage. The commonest BRAF mutation by far is T1799A (V600E), a T>A transversion present in more than 90% of all BRAF-mutated melanomas and naevi.

Due to the detection of the V600E BRAF mutation in this fistula melanoma, it was decided not to look for alterations in genes such as CCND1 or KIT, more frequently mutated in acral and mucosal melanomas [29], since concomitant mutations in BRAF and in genes such as KIT or CCND1 are extremely rare [29]. Regarding this, it has to be understood that, although this melanoma arose in the perianal region, it was not a case of mucosal melanoma, since it did not originate in the anal mucous membrane of the patient.

BRAF mutations, including T1799A, have occasionally been found in anal mucosal melanomas [30] and in other tumours not associated with UV exposure [9], such as thyroid cancer [31, 32], colorectal cancer [33‐35], pancreatic cancer [36, 37] or ovarian cancer [38, 39]. The development of all these tumour types may, in turn, be related to inflammatory processes [12], in the same way as cutaneous melanoma is associated with inflammation of the skin caused by UV exposure [40]. Accordingly, oxidative stress caused by inflammation may be in part responsible for mutations in the BRAF gene [12]. Since fistulae are environments that experience permanent inflammatory processes, the fact that the fistula melanoma reported here is carrying the T1799A transversion supports the hypothesis that oxidative damage caused by free radicals plays a role in the genesis of BRAF mutations.

Obviously, we are not presuming that all inflammation must cause BRAF mutations, but that persistent or chronic inflammatory processes in the absence of UV exposure (such as thyroiditis in thyroid cancer [31], chronic pancreatitis in pancreatic cancer [37], chronic bowel inflammation in colon cancer [41] and, in our case, a longstanding fistula in a melanoma), can probably make the BRAF gene more susceptible to genetic mutations via an as yet unidentified oxidative stress lesion.

As to the management of the perianal fistula melanoma, this case illustrates the benefits provided by the use of sentinel lymph node biopsy. Similarly to anorectal melanoma cases [42], the sentinel lymph nodes were located in the groin, not in the presacral region. As mentioned above, all three nodes dissected were free of tumour cells. After ruling out abdominoperineal resection [43], wide local excision (WLE) appears to obtain general good results when used as the initial treatment of choice in melanomas of the anal region [5, 26]. Our case can only corroborate this, since nine years after WLE our patient remains healthy and without signs of relapse.