Introduction: Desmoplastic melanoma (DM) is a rare subtype of cutaneous melanoma, comprising about 2% of all melanoma cases.  Patients are typically older males (7th decade), and approximately 50% of cases occur in the head and neck.  Sentinel lymph node biopsy (SLNB) has become common in the treatment and staging of most subtypes of invasive head and neck melanoma.  However, the benefits, if any, of SLNB for DM are unclear given different patterns of failure.  We sought to evaluate the role of SLNB in the treatment of DM of the head and neck.

Methods: A retrospective review of a single institution prospectively collected melanoma database was reviewed from years 1995-2014.   DM patients of the head and neck were matched with head and neck melanoma patients of superficial spreading or nodular histology (controls).  Matches were performed based on age, gender, ulceration, and T-stage.  The prevalence of performing SLNB, as well as sentinel node positivity were compared.  Quantitative variables were compared using Wilcoxon rank sum tests, and categorical variables were compared using Fisher’s exact tests.  Multivariate binary logistic regression was performed to determine risk factors for recurrence.  Kaplan-Meier survival curves were compared using the log rank test.

Results: There were 32 head and neck DM patients matched to 60 controls.  Patient and tumor factors were similar between DM patients and controls, respectively: mean age (66 v 64 years old, p=0.54), male gender distribution (75% v 68%, p=0.63), mean tumor thickness (4.18 v 4.23 mm, p=0.77), ulceration (23% v 28%, p=0.62), and T-stage distribution (p=0.99).  The presence of mitoses was much more common in the control group (92% v 53%, p<0.001) while perineural invasion was more common in DM (40% v 7%, p<0.001). 

SLNB was performed in 50% of DM patients vs 60% of control cases (p=0.39), and was only found to be positive in 1/16 (6.3%) of DM vs 8/28 (22.2%) of controls (p=0.16).  Recurrence was less frequent in the DM group 11/32 (34%) vs control 33/60 (55%) (p=0.05).  Performing SLNB did not significantly alter the overall recurrence risk in either group: DM (38% v 32%, p=1.0), control (61% v 46%, p=0.30).  Regional recurrence was rare and unchanged regardless of whether DM patients underwent SLNB (13% v 13%, p=1.0).  On multivariate logistic regression the odds of recurrence was lower in the DM group compared to controls (OR 0.38, 95% CI 0.14-1.04, p=0.06).  The 5 year overall survival and disease free survival for DM patients was better than controls (OS 79% v 62%, p=0.06; DFS 70% v 42%, p=0.10). 

Conclusions: Desmoplastic melanoma is a distinct entity of melanoma with different risk of recurrence and better overall survival.  The benefits of sentinel node biopsy in this patient population remain questionable given the relatively low (6.3%) chance of positive biopsy, and unchanged risk of recurrence when performing sentinel lymph node biopsy.