Evaluation and treatment of periocular melanomas can be difficult. Sentinel lymph node biopsy, introduced by Morton et al,1 is a mode of early detection of regional lymph node metastasis for many variants of solid tumors. Prognostically, these data correlate more closely with melanoma-related mortality than other histologic data.2 Perineural invasion of cutaneous eyelid melanoma is uncommon; neither Dr Char nor William Hoyt, MD (oral communication, May 2005) have seen a case without marked sensory asymmetry. Our case provides a cautionary note regarding both sentinel lymph node biopsy and physical findings associated with perineural invasion.
A 20-year-old woman had delayed diagnosis of a left medial eyelid malignant melanoma. The patient was examined elsewhere at age 14 years for a red, acneiform papule and was followed up for 6 years. During the past year, she experienced periocular shooting pain 1 to 2 times daily and received 2 intralesional corticosteroid injections without regression. The lesion was biopsied elsewhere, revealing melanoma with dimensions of 1.5 × 1.3 cm. Suboptimal processing limited histologic interpretation.
On our examination, there was a linear scar with mild erythema in the left lower eyelid. Sensation in the distribution of the fifth cranial nerve first division (V1) and the fifth cranial nerve second division (V2) was symmetric. Extraocular movements were intact. Results of a metastatic evaluation were negative. We performed a wide local excision of the lesion with sentinel lymph node biopsy. The sentinel node drainage was to the left submandibular node and the biopsy results were negative. The area was resected and the tumor appeared to be removed with clear margins. Pathological analysis revealed residual melanoma 5.3 mm in thickness. The lesion had invaded the subcutaneous component and was at Clark level V. Examination results of the inferior medial and superior lateral margins were negative, as were those of the deep margins. The inferolateral margin was 2 mm and showed a 1-mm area of a contiguous microsatellite, a moderate mitotic rate (5-10 mitoses per high-power field), and a moderate number of tumor-infiltrating lymphocytes.
Unfortunately, the patient continued to experience periocular pain following resection. Magnetic resonance imaging showed tumor involvement of the infraorbital nerve (Figure 1). Orbital computed tomography showed that the infraorbital V2 canal was 8 times the normal size and that the tumor extended at least to the orbital apex. Biopsy results of the infraorbital nerve were positive for melanoma (Figure 2). The patient underwent a wide resection of the melanoma with additional resection of the floor of the orbit and infraorbital nerve as well as partial maxillectomy.
T1-weighted magnetic resonance image of a coronal section showing melanoma involvement of the left infraorbital (V2) nerve. Arrow indicates the tumor on the nerve.
Hematoxylin-eosin–stained section of the left infraorbital (V2) nerve showing melanoma invasion (original magnification ×10 [A] and ×40 [B]).
Cutaneous tumors of the eyelid rarely spread via a perineural route.3 Pain is the most common symptom, accompanied by sensory loss in affected nerves, dysesthesia, and ophthalmoplegia.4 Sentinel lymph node biopsy is the best technique to predict local melanoma recurrence; however, in cases involving perineural spread, this procedure probably has much lower sensitivity. Our case is unusual in at least 2 respects. First, the initial sentinel lymph node biopsy results were negative. Second, the patient lacked the typical symptoms of perineural involvement. While this case is a rare example of malignant cutaneous melanoma both in its manifestation and clinical course, it underscores limitations of tumor staging using sentinel lymph node biopsy.
Correspondence: Dr Char, The Tumori Foundation, 45 Castro St, Suite 309, San Francisco, CA 94114 (email@example.com).
Financial Disclosure: None reported.
Turell ME, Char DH. Eyelid Melanoma With Negative Sentinel Lymph Node Biopsy and Perineural Spread. Arch Ophthalmol. 2007;125(7):983-984. doi:10.1001/archopht.125.7.983