Three months later she reported a 6-week history of a left-sided neck lump, clinically felt to be a 1 cm level 5 node. This was 20 cm from the primary BCC site at which there was a concurrent further recurrence of BCC larger than the original lesion (figure 1). Fine needle aspiration of the neck lump was inadequate. Chest x-ray was normal. A panendoscopy and excision biopsy was planned. Endoscopy was normal but complete excision was precluded as the 1.5 cm mass identified at the time of surgery was closely related to the accessory nerve and no clear surgical plane could be found. Incisional biopsy was taken. Histology showed a predominantly basaloid lesion with small areas of squamous differentiation; some neural tissue was present but no lymphoid tissue (figure 4). A left level 2–5 selective neck dissection was performed and at the same time the recurrent BCC widely excised (figure 3). The latter BCC histology was more aggressive in appearance than previously, as was the metastasis. None of the primary BCC specimens had shown squamous differentiation, lymphovascular or perineural invasion. Immunohistochemistry on all samples showed both the BCC and metastasis to be consistently positive to Ber EP4, Neural Cell Adhesion Molecule (NCAM), cytokeratin 5/6 and cytokeratin 34beta12, but negative to Epithelial Membrane Antigen (EMA), suggesting BCC rather than squamous cell carcinoma (SCC). Thirty-seven neck lymph nodes contained no disease.

This case is very unusual in several : the patient was female and the primary BCC was small, truncal and histologically not aggressive. In their review of 238 cases of metastases, Snow et al⁷ documented a primary BCC on the trunk in just 17% of cases; the mean size of trunk/extremity primaries that metastasised was 217 cm² compared to less than 2 cm² in the present patient. However, more similar to our patient, Tavin et al¹¹ documented six small primary BCCs that underwent multiple recurrences before metastasising. Others have documented that it is common for the primary BCC to have been resistant to treatment.¹ Our patient's primary BCC and the recurrent area after 2 years had been treated by curettage and cautery. Although the British Association of Dermatologists’ guidelines do not generally recommend curettage for recurrent lesions, it is a ‘generally good choice’ for small nodular lesions in low-risk sites such as this, and even in retrospect was an appropriate choice for a small BCC with no worrying original histological features.¹²

The presence of a metastatic mass of BCC unrelated to a lymph node but immediately adjacent to the cervical lymph node basin is the most unusual feature of this case. A review of 170 previously reported cases of metastasis found that lymphatic and haematogenous spread were equally prevalent, with lymph nodes, lungs and bone being the commonest site of metastases;⁵ this review also noted that subcutaneous metastases unrelated to lymph nodes were not uncommon. Interestingly, the authors of this review commented that their impression was that ‘metastatic spread to the pre-auricular, parotid or even the submandibular region might well have been subcutaneous metastases’ (rather than lymph node metastases as reported).⁵ Some skin metastases may actually be explained by direct spread from deeper structures—for example, a lesion reported as adjacent to and involving the parotid gland.⁷ The mechanism for subcutaneous and skin metastases has not been proven and could represent either lymphatic or haematogenous spread. Although occurring outwith lymph nodes, our patient's metastasis could not conceivably represent direct spread; its site in relation to local lymphatics, and its histological and immunohistochemical profile, made it unlikely to be anything other than an in-transit metastatic BCC. It is surprising, although reassuring, that all 37 lymph nodes in the neck dissection proved to be negative.

We have been unable to find any report that conclusively suggests the possibility of in-transit metastasis. One report of a subcutaneous tumour documented lymphoid tissue on the edge of a subcutaneous metastasis and, although the authors did not attach any great significance to this finding, they probably described a similar but just slightly more advanced equivalent to the situation in our patient.⁵ Our patient's metastasis fulfilled older criteria for diagnosis of metastatic BCC (skin rather than mucosal origin, distant to the primary site and similar histology). Immunohistochemical concordance could also reasonably be expected, as we documented. In-transit metastases have been described in most tumours in which lymphatic spread and lymph node metastases occur. Therefore, it is reasonable to assume that in-transit metastases should also occur in some patients with metastatic BCC but this has not previously been confirmed—our report strongly suggests that this has now been documented.

The treatment modality for in-transit metastasis of BCC is controversial. There are no clear guidelines as it is such a rare event. In our case, adjuvant radiotherapy for this area was considered but was felt not to be necessary as excision was complete and all regional lymph nodes were negative. The nearest comparable guidelines are for in-transit metastasis of malignant melanoma and standard guidelines do not recommend adjuvant radiotherapy.^{13 14}