Final Diagnosis -- Desmoplastic melanoma




Desmoplastic melanoma is a term first used by Conley in 1971 to describe a variant of melanoma characterized by a dermal spindle cell population in a background of abundant collagen. Eight years later, Reed and Leonard used the term desmoplastic neurotropic melanoma to describe desmoplastic melanomas with nerve infiltration. Although desmoplastic melanoma may arise de novo, it is most commonly associated with other types of melanoma, mainly lentigo maligna. Like lentigo maligna, desmoplastic melanoma occurs in older individuals, most commonly in the sixth decade of life, and has a slight predilection for males (male-to-female ratio 1.75:1) (5). Lesions are most often located on sun-exposed skin of the head and neck but can also be located on the trunk or extremities.

Clinically, desmoplastic melanoma poses great diagnostic difficulty. Lesions are often described as raised, indurated nodules that are most often amelanotic. Desmoplastic melanomas are amelanotic in 44% - 73% of cases (5). This contrasts significantly with conventional melanoma that is amelanotic in less than 10% of cases. Due to the benign appearance of these lesions, diagnosis is often significantly delayed, and disease is usually advanced at the time of biopsy. In a study by Quinn et al of 280 patients from the Sydney melanoma unit with desmoplastic or desmoplastic neurotropic melanoma, median Breslow thickness was 2.5 mm, and 89% of lesions were Clark level IV or V (5). Other authors report up to 98% of lesions with Clark level IV or V invasion at diagnosis.

Histologically, desmoplastic melanoma is characterized by spindle cells amidst a collagenous background extending into the reticular dermis or subcutaneous adipose tissue. There is often an accompanying atypical or malignant melanocytic proliferation in the epidermis, although this may not be present in de novo lesions. The dermal spindle cell population varies greatly in appearance. It may be hypocellular and bland with minimal mitotic activity raising a differential diagnosis of scar or fibromatosis. At the opposite extreme, there may be sufficient nuclear atypia and mitotic activity to mimic high-grade sarcoma. Additional variable findings include fine, dusty melanin pigment, myxoid areas, multinucleated cells, and lymphoid infiltration either in a perivascular or nodular distribution. The spindle cells exhibit Schwannian and fibroblastic differentiation microscopically, and ultrastructural studies have shown that these cells have the capacity to synthesize both collagen and basement membrane and may contain melanosomes. These findings are consistent with neuroectodermal origin. S100 is typically positive in desmoplastic melanoma with percentages ranging from 80%- 94% (3). Results of Melan-A, tyrosinase, and HMB-45 are variable. In the literature, HBM-45 is commonly reported to be negative in desmoplastic melanoma. Our case with weak S100 expression highlights the variability of these stains and the importance of performing a panel.

Although desmoplastic melanomas are typically diagnosed as advanced lesions, when matched stage for stage, they are generally considered to have survival rates similar to conventional melanomas. In Quinn's study, the overall 5-year survival rate was 75% (5). There was no difference in survival between patients with desmoplastic melanoma and those with desmoplastic neurotropic melanoma at 5 years; however, in the desmoplastic neurotropic group, 20% local recurrence was reported as compared to 5% in the non-neurotropic group. Reported metastatic rates range from 11-40% (4). Jaroszewski et al reported in their study of 59 patients with desmoplastic melanoma, that 16 patients developed systemic metastases (27%) and in 81% of cases, the metastases were pulmonary (4). Regional lymph node metastases occur in 4-15% of patients with desmoplastic melanoma (4,5), which is a significantly lower percentage than is seen with typical melanoma. The combination of increased risk of local recurrence, systemic metastases primarily to lung, and decreased risk of nodal involvement has led some investigators to suggest that desmoplastic melanomas mimic soft tissue sarcomas more than conventional cutaneous melanomas. Due to the fact that desmoplastic melanoma is often amelanotic, locally advanced, and diffusely infiltrating with neurotropism, it is often inadequately excised. Various authors have reported statistically significant decreases in local recurrence rates with two versus one centimeter resection margins.

In summary, this case highlights several key points. Most importantly, when faced with a spindle cell lesion of skin or soft tissue, the diagnosis of desmoplastic melanoma must be included in the differential. This is true even when the age, sex, and/or site of the lesion are not classic. As with our case, the clinical impression is often misleading, so the pathologist must maintain a high index of suspicion. The histologic appearance of desmoplastic melanoma varies greatly from case to case. Therefore, immunohistochemical stains play an important role in diagnosis. A single immunostain is often misleading, so a panel of melanocytic markers is indicated when evaluating melanocytic lesions.


  1. Barnhill, Raymond, Textbook of Dermatopathology, 1st ed., McGraw-Hill, 1998: 577.
  2. Elder DE, Murphy GF. Atlas of Tumor Pathology Melanocytic Tumors of the Skin. Washington, DC: AFIP, 1991: 166-173.
  3. Hoang, Mai et al. CD34 Expression in Desmoplastic Melanoma, Journal of Cutaneous Pathology, 2001 Nov; Vol 28(10) 508-512.
  4. Jaroszewski et al. The Clinical Behavior of Desmoplastic Melanoma, American Journal of Surgery, 2001 Dec; Vol 182(6) 590-595.
  5. Quinn et al. Desmoplastic and Desmoplastic Neurotropic Melanoma, Cancer, 1998 Sept; Vol 83(6) 1128-1135.
  6. Weedon, David, Skin Pathology, 2nd ed., Churchill Livingston, 2002: 823-824.

Contributed by Amy Davis, MD and Drazen Jukic, MD, PhD

Case IndexCME Case StudiesFeedbackHome