Figure 1. Pure and mixed desmoplastic melanomas (DMs) may be indistinguishable and may present with features that are commonly associated with benign lesions. A and B, Pure DM with a Breslow thickness of 7.2 mm on the glabella of an 81-year-old man. A, Clinical image reveals an irregular and ill-defined pink nodule. B, Under dermoscopy, atypical vascular structures are seen, including serpentine vessels (arrows) and vascular blush. C and D, A 2-mm mixed DM located on the back of a 30-year-old woman. A, Clinical image demonstrates a pink nodule. B, Under dermoscopy, a negative network is seen.
Figure 2. Pure desmoplastic melanomas (pDMs) can present as pigmented or amelanotic lesions. Although some pDMs lack an associated epidermal non-DM component, they usually reveal at least 1 atypical vascular structure. A, The clinical image of a 3.25-mm pDM on the back of a 65-year-old man. B, Dermoscopy revealed regression structures including scarlike areas and peppering (dashed square), which is the most frequent regression structure observed in pDMs. In addition, atypical globules (solid square) and crystalline structures (white arrows) were observed. C, Clinical image of an amelanotic, indurated, and ill-defined nodule located on the chest of a 67-year-old man. D, Under dermoscopy, subtle dotted vessels are seen (square). A biopsy was performed, revealing a 6.1-mm pDM with no associated epidermal component.
Figure 3. Mixed desmoplastic melanomas (mDMs) can reveal a greater variety of melanoma-specific structures under dermoscopy, which may facilitate their detection. A, Clinical image of a 1.2-mm mDM on the trunk of a 67-year-old man that presented as a pink and brown papule. B, Dermoscopy revealed atypical globules (solid square) and dotted vessels (dashed square). C, A 1.6-mm mDM that presented clinically as an erythematous lesion with a focal bluish discoloration on the arm of an 81-year-old woman. Palpation of the lesion revealed a firm component. D, Dermoscopy demonstrated atypical dots/globules (solid square), polymorphous vessels (dashed squares), off-center blotch, and blue-white veil (arrow).
Jaimes N, Chen L, Dusza SW, et al. Clinical and Dermoscopic Characteristics of Desmoplastic Melanomas. JAMA Dermatol. Published online January 16, 2013. doi:10.1001/jamadermatol.2013.2248.
eFigure 1. A 0,8mm mixed desmoplastic melanoma on the trunk of a 66-year-old male. A, Clinical evaluation revealed an outlier lesion with regular border and pink, light-brown and dark brown colors. B, Dermoscopy evaluation revealed crystalline structures (arrows), vascular blush and polymorphous vessels including dotted and serpentine vessels (square)
eFigure 2. Mixed desmoplastic melanoma 1,6mm on the trunk of a 43-year-old male. A, Clinically it presented as a pink nodule. B, Under dermoscopy atypical vascular structures were observed including dotted vessles, serpentine vessels and milky red globules (square).
eFigure 3. A 9mm pure desmoplastic melanoma on the face of a 85-year-old male. A, Clinical evaluation revealed an outlier pink nodule. B, Dermoscopy evaluation revealed crystalline structures (white arrow) and polymorphous vessels including dotted and serpentine vessels (black arrows).
eFigure 4. A, 2,55mm pure desmoplastic melanoma on the arm of a 48-year-old female that presented as an outlier pink nodule (A). B, Under dermoscopy focal peppering (dashed square) and polymorphous vessels (solid square) were observed. Some vessels might have been compressed during imaging.
Jaimes N, Chen L, Dusza SW, Carrera C, Puig S, Thomas L, Kelly JW, Dang L, Zalaudek I, Braun RP, Menzies SW, Busam KJ, Marghoob AA. Clinical and Dermoscopic Characteristics of Desmoplastic Melanomas. JAMA Dermatol. 2013;149(4):413-421. doi:10.1001/jamadermatol.2013.2248
Author Affiliations: Dermatology Service (Drs Jaimes, Dusza, and Marghoob and Ms Chen) and Department of Pathology (Dr Busam), Memorial Sloan-Kettering Cancer Center, New York, New York; Melanoma Unit, Dermatology Department, Hospital Clinic of Barcelona, IDIBAPS, and CIBER de Enfermedades Raras, Instituto de Salud Carlos III, Barcelona, Spain (Drs Carrera and Puig); Department of Dermatology, Lyon 1 University Center Hospitalier Lyon Sud, Lyon, France (Dr Thomas); Victorian Melanoma Service, Alfred Hospital, Melbourne, Victoria, Australia (Drs Kelly and Dang); Dermatology and Skin Cancer Unit, Arcispedale Santa Maria Nuova, Instituto di Ricerca e Cura a Carattere Scientifico (IRCCS), Reggio Emilia, Italy (Dr Zalaudek); Division of Dermatology, Medical University of Graz, Graz, Austria (Dr Zalaudek); Department of Dermatology, University Hospital, Zurich, Switzerland (Dr Braun); and The Sydney Melanoma Diagnostic Centre, Sydney Cancer Centre, Royal Prince Alfred Hospital and The Sydney Medical School, The University of Sydney, Sydney, Australia (Dr Menzies).
Objective To describe and analyze the clinical and dermoscopic characteristics of desmoplastic melanoma (DM) as a function of pathologic subtype and phenotypic traits.
Design Retrospective case series.
Setting Eight high-risk dermatology clinics.
Patients Patients with DM confirmed by histopathologic analysis whose records included a high-quality dermoscopic image.
Main Outcome Measures Clinical, dermoscopic, and histopathologic features of DM.
Results A total of 37 DM cases were identified. The majority of patients had fair skin, few nevi, and no history of melanoma. Lentigo maligna was the most frequent subtype of melanoma associated with DM. The most frequent clinical presentation of DM was a palpable and/or indurated lesion located on sun-exposed skin. Forty-three percent of cases were classified as pure DM, and 57% as mixed DM. Pure DM lesions were thicker than mixed DM lesions (4.10 vs 2.83 mm) (P = .22) and were less likely to have an associated epidermal non-DM component (63% vs 100%) (P = .004). Dermoscopically, DMs had at least 1 melanoma-specific structure, the most frequent being atypical vascular structures. Peppering was more frequently seen in pure DM (44% in pure DM vs 24% in mixed DM) (P = .29). In contrast, crystalline structures, polymorphous vessels, and vascular blush were more commonly seen in mixed DM.
Conclusions Though DM can be difficult to diagnose based on clinical morphologic characteristics alone, dermoscopy has proved to be a useful aid during the evaluation of clinically equivocal lesions or those lesions with a benign appearance. The most common dermoscopic clues observed in DMs included atypical vascular structures, peppering, and occasionally other melanoma-specific structures.
Desmoplastic melanoma (DM) is a rare variant of melanoma, making up less than 4% of all melanomas.1,2 The overall incidence rate of DM is 2.0 per million, with a peak of 15.2 per million for persons 80 years or older.3 Typically, DMs are diagnosed later in life than non-DMs. In addition, it is not uncommon for the definitive diagnosis of DM to be delayed because of difficulties in its initial clinical recognition.4
Often DM manifests features commonly associated with benign lesions. Based on primary clinical morphologic characteristics alone, it may be difficult for clinicians to recognize DM as a malignant neoplasm. However, dermoscopy may provide clues prompting a biopsy in these otherwise clinically benign-appearing lesions. Although dermoscopy has been shown to increase diagnostic accuracy for most melanoma subtypes, there are limited data regarding the role of dermoscopy in the diagnosis of DM.5,6 In fact, to our knowledge, there has been only 1 retrospective study that described the dermoscopic features of 6 cases of DM.7
Dermoscopy may be a useful aid during the evaluation of DM, and together with a clinical history and examination, it can guide the clinician toward performing a biopsy. The aim of the present retrospective study is to describe and analyze the most common phenotypic traits associated with DM and to evaluate its most frequent clinical and dermoscopic characteristics.
Cases of patients diagnosed with DM were retrospectively selected from 8 melanoma centers in New York, New York (Memorial Sloan-Kettering Cancer Center [MSKCC]), Barcelona, Spain (Hospital Clinic of Barcelona), Lyon, France (Hospitalier Lyon Sud), Graz, Austria (Medical University of Graz), Reggio Emilia, Italy (Arcispedale Santa Maria Nuova), Sydney, Australia (Royal Prince Alfred Hospital), Melbourne, Australia (Alfred Hospital), and Zurich, Switzerland (University Hospital). Cases were included only if there was a confirmed pathologic diagnosis of DM and a high-quality dermoscopic image available. Whenever possible, clinical images were also evaluated. The study was approved by the institutional review board at MSKCC.
All lesions were confirmed to be a DM by histopathologic analysis. In addition, cases were further categorized into 2 histopathologic subtypes (pure DM [pDM] and mixed DM [mDM]) based on the degree of desmoplasia present in the tumor, as described by Busam et al.8 Specifically, pDMs were defined as having more than 90% desmoplasia, and mDMs had less than 90% desmoplasia. The presence of an associated epidermal non-DM component was also recorded. Neurotropism was evaluated by means of conventional hematoxylin-eosin and, when available, with S100 stain.
Demographic, clinical, and histopathologic information was obtained from the patients' medical records. Patients were assessed for the number of nevi, skin type, personal and family history of melanoma, nonmelanoma skin cancers, and history of chronic sun damage. When clinical images were available, the lesions were evaluated for their primary morphologic characteristics (macule, papule, plaque, nodule), colors (skin color, pink, red, blue, white, brown, and black), borders, and anatomic location.
Dermoscopic images were assessed for the presence or absence of melanocytic structures, including pigment network (gridlike network consisting of pigmented lines and hypopigmented “holes”), aggregated globules (3 or more clustered, well-demarcated, round to oval symmetric structures larger than 0.1 mm; may be brown, black, and blue), streaks (radial projections at the periphery of the lesion extending from the tumor toward the surrounding normal skin; may present as pseudopods or radially streaming structures), and negative network (serpiginous interconnecting hypopigmented lines that surround irregularly shaped pigmented structures resembling elongated, curvilinear globules).9
In addition, vascular structures and the presence or absence of melanoma-specific structures were evaluated, including atypical network (increased variability in the width of the network lines, their color and distribution; hole sizes of increased variability), negative network, streaks, atypical dots and/or globules (multiple globules irregularly distributed within the lesion or asymmetrically located off center or focally at the periphery; not associated with the pigmented network), off-center blotch (off-center homogeneous areas of pigment that obscure visualization of any other structures; may be dark brown to black), peripheral tan structureless areas (structureless areas located at the periphery of the lesion larger than 10% of a lesion area), blue-white veil (confluent blue pigmentation with an overlying white “ground glass” haze), regression structures (ie, scarlike depigmentation lighter than the surrounding skin and appearing shiny white under polarized dermoscopy; and/or peppering, which consists of tiny blue-gray granules giving the appearance of a blue-white veil), crystalline structures (shiny, white linear streaks that are often oriented parallel or orthogonal to each other), and atypical vascular structures (including dotted vessels consisting of red dots of 0.01 to 0.02 mm, serpentine vessels consisting of linear irregular or undulating short vessels, polymorphous vessels consisting of a combination of 2 or more vessel morphologic characteristics, corkscrew vessels consisting of coiled or tortuous vessels, milky-red globules and/or vascular blush consisting of ill-defined globules with a milky-red color and ill-defined areas of milky-red color).9
Finally, the clinical diagnosis, previous treatments, and follow-up status were recorded for each case. Presence or absence of local recurrence or metastases was documented. Descriptive frequencies were calculated for all cases and for each category (pDM and mDM). P values were calculated based on the Fisher exact or t test.
A total of 37 cases of DM that met the inclusion criteria were identified. Of these, 22 patients (60%) were men, and 15 (41%) were women, with an average age of 69 years (age range, 30-89 years). All patients had fair skin, and more than half of patients (78%; n = 21) had few nevi (fewer than 20 nevi). Only 3 patients (10%) had a family history of melanoma, and 9 patients (25%) had a history of melanoma (Table 1). Although 42% patients had a history of nonmelanoma skin cancer (NMSC) (n = 11), patients with a pDM were less likely to have a history of NMSC than patients with mDM (23% [n = 3] vs 62% [n = 8]) (P = .11).
Eighty-nine percent of DM lesions developed on sun-exposed areas (n = 33). Clinically, most DMs presented as palpable and/or indurated lesions (87% [n = 27]) with irregular and ill-defined borders (64% [n = 23]) (Figures 1,2, and 3) (eFigures 1-4). The most common primary morphologic characteristics were plaques (30% [n = 11]), macules (22% [n = 8]), papules (16% [n = 6]), or nodules (16% [n = 6]). Overall, DMs had at least 2 colors (58% [n = 21]), with pink/red and brown being the most common (Table 2; Figures 1,2, and 3) (eFigures 1-4).
The prebiopsy clinical diagnoses and reasons for performing a biopsy were available for 11 cases. The prebiopsy clinical diagnoses reported included melanoma (n = 3), basal cell carcinoma (n = 3), melanoma vs dysplastic nevus (n = 1), basal cell carcinoma vs lichen planus–like keratosis (n = 1), squamous cell carcinoma (n = 1), seborrheic keratosis (n = 1), and cyst (n = 1). Reasons for suspecting a malignant neoplasm included the presence of melanoma-specific structures noted under dermoscopy, new lesion found on total body skin examination, and/or the lesion identified as an outlier. In addition, for at least 16 of the 37 patients, the diagnosis of a malignant neoplasm was not entertained, and the lesion underwent treatment with cryotherapy (n = 8), laser (n = 1), intralesional steroids (n = 1), or other treatment not specified (n = 6) prior to the final biopsy that disclosed the diagnosis of DM. These lesions were 8 pDMs and 8 mDMs, located on the head and neck (44% [n = 7]), lower extremities including soles (25% [n = 4]), and trunk (19% [n = 3]); less frequently on the back (6% [n = 1]) or upper extremities (6% [n = 1]). Clinically, these lesions appeared as macules and/or plaques with irregular and ill-defined borders, with 1 color in 44% of the cases (n = 7), mainly pink or light brown. The rest of the lesions had 2 or more colors, in particular pink and light brown. Dermoscopically, the predominant structures observed in these lesions were vascular blush (56% [n = 9]), polymorphous vessels (38% [n = 6]), peppering (38% [n = 6]) and asymmetric perifollicular hyperpigmentation (38% [n = 6]) (Figure 1C and D).
Sixteen of the 37 cases were classified as pDM (43%), and 21 were classified as mDM (57%). Even though there were no statistically significant differences between pDM and mDM by anatomic location, it was observed that pDMs were more frequent than mDMs on the trunk (31% [n = 5] vs 19.0% [n = 4]) and lower extremities (6% [n = 1] vs 0% [n = 0]) (P = .84). In addition, none of the pDMs developed on sun-protected sites, whereas 4 of the 21 mDMs developed on sun-protected sites, including 2 lesions on the sole and 2 on the lower back (Table 2).
Dermoscopically, 16 of the 37 lesions (43%) revealed melanocytic structures including globules (44% [n = 7]), pigment network (38% [n = 6]), pseudonetwork (25% [n = 4]), and negative network (6% [n = 1]) (Table 3; Figures 1D,2B, and 3B). None of the lesions presented homogeneous blue pigmentation or streaks. While 57% of the DMs lacked any of the aforementioned melanocytic structures [n = 21], they all revealed at least one melanoma-specific structure. The most frequent melanoma-specific structures observed were atypical vascular structures (81% [n = 30]), followed by regression structures (ie, peppering and scarlike areas), blue-white veil, atypical globules, and atypical network (Figures 1B and D, 2B and D, and 3B and D) (eFigures 1-4). Peppering (also known as granularity) was the most frequent regression structure observed and was more frequent in pDM than mDM (44% [n = 7] vs 24% [n = 5]) (Figure 2B). It was noted that within the group of mDMs, there was a greater variety of other melanoma-specific structures that were not manifested in the pDMs, including off-center blotch, negative network, and peripheral tan structureless areas. In addition, features associated with lentigo maligna (LM), including annular granular pattern (24% [n = 9]) and polygonal lines (11% [n = 4]), were also observed.
Fifteen of the dermoscopic images of lesions were acquired using polarized dermoscopy (PD), and 19 using non-PD. For the 3 remaining lesions, it was unknown what device was used to acquire the images. Of the 15 lesions evaluated with PD, 80% revealed crystalline structures (n = 12) (Table 3; Figure 2B) (eFigures 1 and 3). It was observed that crystalline structures, polymorphous vessels, and vascular blush were more commonly seen in mDM than in pDM.
Vascular structures were observed in 81% of the lesions (n = 30), in particular atypical vessels and/or vascular blush. When both groups of DM were compared, pDMs were less likely to present with polymorphous vessels (31% [n = 4] vs 53% [n = 9]) and vascular blush (54% [n = 7] vs 77% [n = 13]) than were mDMs (Table 3).
Overall, pDMs tended to be thicker tumors than mDMs (4.10 vs 2.83 mm) (P = .22) and were less likely to have an associated epidermal non-DM component (63% [n = 10] vs 100% [n = 20]) (P = .004). There was 1 mDM for which this information was unavailable. The most commonly associated epidermal non-DM component was LM (53% [n = 16]) followed by superficial spreading melanoma (SSM) (10% [n = 3]) (Table 4).
Once the diagnosis of DM was confirmed, the most common treatment for the DM was surgical (97% [n = 31]), including wide local excision (94% [n = 30]) or staged excision (3% [n = 1]). Two patients with pDM received adjuvant radiation therapy. The average length of follow-up for the 37 patients was 30 months (range, 1-96 months). Two patients with mDM (10%) presented with local recurrence, one after 24 months and the other after 40 months (Breslow thickness, 4 mm and not available, respectively). The patient with longer follow-up also developed metastases to left submaxillary lymph nodes 54 months after the initial diagnosis. In addition, another patient with mDM (Breslow thickness, 10 mm) developed bilateral inguinal lymph node metastases 20 months after the initial diagnosis. Two patients with pDM developed metastases (13%), one at 17 months, and the other at 9 months after the initial diagnosis (Breslow thicknesses, 12 mm and 6.1 mm, respectively). The patient with the 12-mm pDM developed metastasis to lymph nodes, and the patient with the 6.1-mm pDM developed metastasis to lymph nodes, liver, and lungs.
Desmoplastic melanoma is a relatively rare entity that often presents with features that are more commonly associated with benign lesions. Thus, based on clinical morphologic characteristics alone, DM can prove to be a challenging diagnosis. Even under histopathologic examination, DM can be confused with other entities such as scars, dermatofibromas, and desmoplastic nevus.4,10,11 Ferrara et al11 describe the dermoscopic patterns of 3 cases of desmoplastic nevus that all appeared clinically as small, flesh-colored papules characterized by a subtle light-brown network overlying a pinkish erythematous background. All nevi were devoid of asymmetry or melanoma-specific patterns. Based on its nondescript appearance and the difficulties in the initial clinical recognition, it is not surprising that DM is seldom suspected in its early stages.4 In addition, it is not uncommon for these lesions to be diagnosed as benign entities and get treated as such. In fact, in our study at least 10 cases (27%) had received previous treatments including cryotherapy, laser, and intralesional steroids, before the definitive diagnosis of DM was confirmed by biopsy.
To our knowledge, this is the largest series describing the clinical and dermoscopic characteristics of DM. In addition, this study attempts to describe the most common phenotypic traits present in patients with DM. Overall, our patients with DM had fair skin, and most of them demonstrated actinic damage, few nevi, and a negative personal and family history of melanoma (Table 1). Dermoscopically, our study demonstrates that while 57% of the DMs (n = 21) lacked melanocytic pigmented structures (ie, globules, pigment network, and pseudonetwork on facial skin), all cases of DM revealed at least 1 melanoma-specific structure, in particular atypical vascular structures, peppering, blue-white veil, atypical globules, crystalline structures, and atypical network (Figures 1B and D, 2B and D, and 3B and D) (eFigures 1-4). Furthermore, dermoscopic features of LM such as annular granular pattern and polygonal lines were seen in one-third of the cases. In fact, 83% of our cases had an associated epidermal non-DM (n = 30), with LM being the most common type, followed by SSM. Thus, we are of the opinion that since DMs can be associated with LM or SSM, all of these lesions should be palpated to rule out a dermal component of DM (Figure 3D).10,12
Previous studies have reported that DM lesions tend to grow slowly and are more frequently located on the head and neck, extremities, and trunk of fair-skinned elderly men. The male to female ratio has been reported to be 2:1, and the mean age at diagnosis is 66 years, which is older than the 60 years found for non-DM lesions.2- 4,13,14 Our data reflect a similar male to female ratio (2:1) and average age. In addition, we also found that DM has a predilection for sun-exposed areas (89% [n = 33]), in particular for the head and neck (49% [n = 18]). Only 4 DMs of the mixed type were found on sun-protected areas. Thus, our results are in accordance with other studies that have reported a link between DM, chronic actinic damage, and LM.2,3 We speculate that mDM is more likely to have risk factors and clinical and phenotypic characteristics similar to those of LM. In contrast to DM, desmoplastic nevus appears to be associated with younger age and development on non–sun-exposed areas. Thus, patient age and anatomic location seem to be important criteria in the differential diagnosis between benign and malignant desmoplastic neoplasia.11
The clinical and dermoscopic characteristics of DMs were described as a function of the pathological subclassification (pDM vs mDM), which is based on the extent of desmoplasia present in the invasive component. This classification appears to have prognostic and perhaps even therapeutic implications.15,16 Evidence suggests that pDMs are less likely to have regional lymph node involvement and are associated with a more favorable outcome. In contrast, mDMs have been associated with more locoregional recurrences.15
From a clinical perspective, pDMs and mDMs may be indistinguishable. Both types of DM can present as firm and indurated lesions with irregular and ill-defined borders (Figure 1A and C). While there are limited data regarding the role of dermoscopy in the diagnosis of DM,5,6 the present study supports its use during the evaluation of skin lesions because it provides additional information that will prompt the clinician to perform a proper biopsy. One retrospective study performed by Debarbieux et al7 describes the dermoscopic features of 6 DM cases. The authors reported that 3 lesions lacked melanocytic criteria but presented with ulceration, vascular structures, and regression structures including scarlike areas, peppering, and/or blue-white veil. The other 3 lesions presented with melanocytic features such as irregular pigment network or pseudonetwork in addition to white scarlike areas, peppering (2 lesions), blue-white veil (2 lesions), and vascular structures (2 lesions). These 3 lesions presented in association with a LM or SSM. Vascular structures were noted in 5 of the 6 lesions and consisted of serpentine vessels (also known as linear irregular vessels) and/or milky red areas. In accordance with Debarbieux et al,7 we found that 81% of DMs revealed vascular structures (n = 30), particularly atypical vessels and/or vascular blush. When DMs were separated into histopathologic subtypes, we observed that pDMs tended to present with monomorphous vessels (38% [n = 6] vs 35% [n = 5]), whereas mDMs tended to present with a polymorphous pattern (31% [n = 4] vs 53% [n = 9]) and vascular blush (54% [n = 7] vs 77% [n = 13]) (Table 3).
Although 38% of the pDMs in the present study lacked an associated epidermal non-DM component (n = 6), all of these cases revealed at least 1 atypical vascular structure (Figures 1B and 2D) (eFigures 1-4). Therefore, it may be hypothesized that since mDM lesions can reveal a greater variety of melanoma-specific structures under dermoscopy (eg, crystalline structures, polymorphous vessels, vascular blush, off-centered blotch, negative network, and peripheral tan structureless areas), they may be easier to detect than pDM lesions (Figures 1D and 3B and D). This in turn may result in greater delays in diagnosis and greater Breslow thickness at diagnosis for pDMs (4.01 mm vs 2.83 mm). We suggest that the most important indicator prompting the dermatologist to perform a biopsy to rule out a melanoma may be the presence of dermoscopic structures associated with the epidermal non-DM component. In cases in which there is no epidermal component, the presence of atypical vascular structures may lead to a heightened suspicion for malignant neoplasm (Figures 1B and 2D and eFigures 2 and 4).
Limitations of our study are that it is a retrospective study of cases that included high-quality dermoscopic images but did not always include clinical and demographic information. The histopathologic diagnosis of each DM lesion was based on the official pathology diagnosis from the corresponding high-risk dermatology center, and a second pathologist did not confirm the diagnosis of DM. Although dermoscopic differences between pDM and mDM were observed, the power of the study did not reach statistical significance owing to the small number of cases. However, we believe that it is still important to highlight these differences, although future larger databases will be required to validate our findings.
In conclusion, this study demonstrates that dermoscopy is a useful aid during the evaluation of clinically equivocal lesions or those with a benign appearance. Although DM can be difficult to diagnose based on clinical morphologic characteristics alone, clinical information and risk factors for DM need to be considered when evaluating indurated and firm lesions on sun-damaged skin, including male sex, older age, chronic sun exposure, and the presence of an associated LM. In addition, all lesions suggestive of LM should be palpated and evaluated with dermoscopy. Palpation can lead to the detection of a subcutaneous nodule, which may be another clue for the diagnosis of DM. Dermoscopy can provide additional clues that might heighten the clinical suspicion for DM and may guide the clinician to perform a biopsy on these otherwise benign-appearing lesions. Common dermoscopic structures in DM include atypical vascular structures, peppering, or other melanoma-specific structures.
Correspondence: Ashfaq A. Marghoob, MD, Department of Dermatology, Memorial Sloan-Kettering Cancer Center, 160 E 53rd St, New York, NY 10022 (firstname.lastname@example.org).
Accepted for Publication: October 3, 2012.
Published Online: January 16, 2013. doi:10.1001/jamadermatol.2013.2248
Author Contributions: Drs Jaimes, Chen, and Marghoob had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Jaimes, Chen, Braun, Busam, and Marghoob. Acquisition of data: Jaimes, Chen, Carrera, Puig, Thomas, Kelly, Dang, Zalaudek, Braun, Menzies, Busam, and Marghoob. Analysis and interpretation of data: Jaimes, Dusza, Carrera, Busam, and Marghoob. Drafting of the manuscript: Jaimes, Chen, Dusza, Carrera, Dang, and Busam. Critical revision of the manuscript for important intellectual content: Jaimes, Dusza, Puig, Thomas, Kelly, Zalaudek, Braun, Menzies, Busam, and Marghoob. Statistical analysis: Dusza. Administrative, technical, and material support: Jaimes, Dang, and Zalaudek. Study supervision: Busam and Marghoob.
Conflict of Interest Disclosures: None reported.