Pattern analysis: a two-step procedure for the dermoscopic diagnosis of melanoma

Abstract

Pattern recognition has historically been used by clinicians and histopathologists to differentiate benign lesions from malignant neoplasms. A similar process has been found useful with dermoscopy and has been termed “pattern analysis.” In 1987 the Austrian group of Pehamberger, Steiner, and Wolff published their classic article on pattern analysis in the Journal of the American Academy of Dermatology.1 On the basis of 3,000 pigmented lesions, the Viennese group demonstrated that pattern analysis of specific dermoscopic features permitted a distinction between the various types of pigmented skin lesions, specifically between benign and malignant melanocytic lesions. Since then, the concept of classic pattern analysis has been confirmed and redefined continuously by experts from all over the world. Special references include the Consensus Meeting in 1990,2 an article by Kenet et al3 in the Archives of Dermatology in 1993, an educational CD in 2000 by Argenziano et al,4 and the Consensus Net Meeting in 2000. This article primarily utilizes the terminology and descriptions from the Consensus Net Meeting of 2000 (http://www.dermoscopy. org), the Atlas of Dermoscopy, and the educational compact disc by Argenziano et al.4 There are two steps in the process of pattern analysis (Fig 1). The first step is to decide whether the lesion is melanocytic or nonmelanocytic (Fig 2). A modification of the Stolz algorithm4–8 is used for this first step, and a determination is made for the presence of aggregated globules, a pigment network (excluding fingerprint-like network), or branched streaks. If this is the case, the lesion should be considered a melanocytic lesion. If not, one should evaluate the lesion for the presence of homogeneous steel-blue areas. If so, the lesion should be considered as a blue nevus (unusual exceptions are a pigmented basal cell carcinoma or a metastatic melanoma). The lesion should then be evaluated for the presence of moth-eaten borders, fingerprinting, comedolike openings, milialike cysts, and fissures. If so, the lesion is suggestive of either a solar lentigo or a seborrheic keratosis. One then looks for red or red-blue to black lagoons, and if these structures are present, the lesion should be considered a hemangioma or an angiokeratoma. Finally, the lesion is evaluated for leaflike structures, arborizing telangiectasias, spoke-wheel-like areas, and gray-blue ovoid nests. If present, the lesion is likely a basal cell carcinoma. If all the preceding questions were answered with “no,” the lesion should be considered a melanocytic lesion. Step 2 in the pattern analysis process is the differentiation of benign melanocytic lesions from melanomas. The overall general appearance of Color, Architectural order, Symmetry of pattern, and Homogeneity (CASH) are important components in distinguishing these two From the Pigmented Skin Lesion Unit, Department of Dermatology, University Hospital Geneva, and the Departement Hospitalo-Universitaire Romand de Dermatologie et Venerologie, Geneva/Lausanne, Switzerland; Skin and Cancer Associates, Plantation Florida and Department of Dermatology, University of Miami School of Medicine, Miami, Florida; and the Oncology Section, Ronald O. Perelman Department of Dermatology, New York University School of Medicine, New York, New York. Address correspondence to R. P. Braun, Department of Dermatology, University Hospital Geneva, 24, rue Micheli-du-Crest, CH–1211 Geneva 14, Switzerland. E-mail address: braun@melanoma.ch Figure 1. Two-step procedure for dermoscopic diagnosis of pigmented skin lesions.