Variation in dermoscopic features of basal cell carcinoma as a function of anatomical location and pigmentation status

Dear Editor, Detecting basal cell carcinomas (BCCs) early and at relatively small sizes may expand therapeutic choices and enable less invasive treatment options. To this end, dermoscopy is a useful tool to diagnose BCC at early stages with high sensitivity and specificity.1 Superficial BCCs have been shown to be independently associated with location on the trunk and extremities and nodular BCCs have been independently associated with a head/neck location.2,3 Few studies have evaluated dermoscopic variation in BCC morphology as a function of anatomical location and pigmentation status.4,5 Herein, we evaluate BCC dermoscopic morphology based on pigmentation status and anatomical location. We conducted a retrospective analysis of patient demographics and dermoscopic and clinical features of consecutive histopathologically proven BCCs that had been biopsied between 1 December 2007 and 1 March 2014 at the Pigmented Lesion Clinic of Skin and Cancer Associates, Plantation, FL, U.S.A. Collision and recurrent tumours, and cases with poor image quality were excluded. Clinical and dermoscopic images were taken in vivo with a Nikon Coolpix 4500 camera Nikon USA, Inc., Melville, NY, U.S.A. and DermLite II Pro HR and DermLite II Fluid dermoscopes, 3Gen Inc., San Juan Capistrano, CA, U.S.A. at ×10 magnification. Dermoscopic images were routinely taken via nonpolarized and polarized contact and noncontact dermoscopy. Each lesion was classified by anatomical site as: head/neck, upper extremities (UE), trunk, and lower extremities (LE). Lesions were assessed for pigmentation and for the presence/absence of dermoscopic features. Of 479 consecutively biopsied BCCs evaluated, 392 histopathologically proven BCCs on 348 unique patients were included in the study (87 BCCs were excluded because of poor image quality). Tumours were on the head/neck for 169 (43%) lesions, trunk for 129 (33%), LE for 54 (14%) and UE for 40 (10%). The average patient age was 64·9 years (range 50·0–79·8). Pigmented lesions varied significantly across anatomical sites, with the highest frequency on the UEs (65·0%) and lowest frequency on the head/neck (49·1%). Compared with other anatomical sites (head/neck, UE and trunk) combined, BCCs on the LEs had a higher frequency of polymorphous vessels (74% vs. 44%), thin serpentine vessels (81% vs. 63%), shiny white blotches (85% vs. 46%), long white strands (76% vs. 45%), erythema (87% vs. 59%) and erosions/ulcers (69% vs. 35%) (P ≤ 0·01 for all comparisons). BCCs on the LE had the lowest frequency of arborizing vessels (24% vs. 78%) (P ≤ 0·0001) when compared with all anatomical sites. Overall, while 264 lesions contained arborizing vessels, 84 contained additional vessel morphologies resulting in the polymorphous attribute. Clinically, 35·2% BCCs were pigmented, and 57·7% were pigmented under dermoscopy. Compared with pigmented BCCs, nonpigmented BCCs were more likely to have arborizing vessels (77% vs. 59%), erythema (57% vs. 46%), shiny white blotches (48% vs. 34%) and long white strands (47% vs. 33%) (P < 0·05 for all) (Table 1). Dermoscopy features as a function of anatomical location and pigmentation Other is a combination of upper extremities and trunk. Pigmentation status is based on dermoscopy. Bold values indicate significance of P < 0·01 by the Fisher's two‐tailed exact test when compared with pigmented counterpart. Dermoscopy features as a function of anatomical location and pigmentation Other is a combination of upper extremities and trunk. Pigmentation status is based on dermoscopy. Bold values indicate significance of P < 0·01 by the Fisher's two‐tailed exact test when compared with pigmented counterpart. BCCs on the LE have been associated with female sex, a younger age at diagnosis, and a superficial histopathological subtype.6 While studies examined dermoscopic morphology of BCC of the head/neck vs. truncal location,5 data are limited on dermoscopic features of BCCs on the LE. In corroboration with Suppa et al.,5 we found that BCCs of the head/neck frequently presented with arborizing vessels and BCCs on the trunk frequently presented with leaf‐like structures, concentric structures, erosions/ulcers and polymorphous vessels. It is important to note that our analysis included the full spectrum of BCC histopathological subtypes (including infiltrative and morphoeiform). In addition, the present study evaluated all BCCs with nonpolarized and polarized images.4,5 By including nonpolarized and polarized images for every lesion, we characterized the full spectrum of dermoscopic features in BCC. The dermoscopic picture of BCCs on the LEs described in this study is strikingly different from those on the head/neck. Our study has established that shiny white streaks (SWS), polymorphous vessels and ulceration/erosions should raise suspicion for BCC.7 This is an important finding because polymorphous vessels and ulceration are features highly associated with melanoma.8 Our experience and the findings show that the combination of dermoscopic features of polymorphous vessels, ulcerations/erosions and SWS, identified via polarized light, can tilt the differential towards BCC, especially in LE lesions. In the present study, the BCC subtype was predominantly nodular, then infiltrative, superficial and other. The subtype distribution reflects an older population with sun‐damaged skin presenting to the study clinic. Our sample size precluded us from meaningfully exploring the relationship between BCC subtype and anatomical location. Other limitations include the use of images from a single practice and a retrospective study design. Despite these limitations, we demonstrate that BCCs have varying presentations depending on anatomical location and pigmentation status, with a novel description of the unique presentation of BCCs on the LEs. Funding sources: this research was funded in part through the NIH/NCI Cancer Center Support Grant P30 CA008748.