BJD C L I N I C A L A N D LA B O R A T O R Y I N V E S T I G A T I O N S
British Journal of Dermatology
Blue-black rule: a simple dermoscopic clue to recognize pigmented nodular melanoma G. Argenziano, C. Longo, A. Cameron,* S. Cavicchini, J.-Y. Gourhant, A. Lallas,§ I. McColl,– C. Rosendahl,* L. Thomas,** D. Tiodorovic-Zivkovic, P. Zaballos and I. Zalaudek§§ Dermatology Unit, Medical Department, Arcispedale Santa Maria Nuova, Viale Risorgimento 80, 42100 Reggio Emilia, Italy *School of Medicine, The University of Queensland, Brisbane, 4006 Qld, Australia Dermatology Unit, Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, 20122 Milan, Italy Centre de Dermatologie, 77140 Nemours, France §Hospital of Venereal and Skin Diseases, 54643 Thessaloniki, Greece –Australian Institute of Dermatology, John Flynn Medical Centre, Gold Coast, 4224 Qld, Australia **Department of Dermatology, Lyon 1 University, Centre Hospitalier Lyon Sud, 69495 Pierre Be´nite, France Clinic of Dermatovenerology, Clinical Center of Nis, Medical Faculty University of Nis, 18000 Nis, Serbia Department of Dermatology, Hospital de Sant Pau i Santa Tecla, 43007 Tarragona, Spain §§Department of Dermatology, Medical University of Graz, 8036 Graz, Austria
Summary Correspondence Giuseppe Argenziano. E-mail: [email protected]
Accepted for publication 4 September 2011
Funding sources None.
Conﬂicts of interest None declared. DOI 10.1111/j.1365-2133.2011.10621.x
Background Dermoscopy improves melanoma recognition, but most criteria were described in the context of superficial spreading melanoma. Objectives To test whether pigmented nodular melanoma could be recognized dermoscopically by the presence of a combination of blue and black colour within the lesion. Methods Dermoscopic images of histopathologically diagnosed pigmented nodular tumours with no (or only minimal) flat component were evaluated for the presence of standard melanoma criteria and for the presence of a new feature named blue-black (BB) colour. Sensitivity, specificity, positive predictive value and negative predictive value were calculated for standard criteria and BB feature in relation to the diagnosis of melanoma and to diagnosis of malignancy. Results Of 283 lesions, 160 were malignant, including 78 (27Æ6%) melanomas, and 123 were benign. The BB feature and the standard criteria had 78Æ2% and 43Æ6% sensitivity for melanoma, respectively, whereas a combined method based on the presence of either the BB feature or one (or more) of the standard criteria reached 84Æ6% sensitivity, with 80Æ5% specificity and 93Æ2% negative predictive value. Sensitivity values for malignant lesions were only 24Æ4%, 56Æ9% and 60% for standard criteria, BB feature and the combined method, respectively. However, the combined method gave 91Æ9% specificity and 90Æ6% positive predictive value for malignancy. Conclusions Using a method based on the BB feature or one of the standard melanoma criteria, only 9Æ4% of positive pigmented nodular lesions were found to be benign and only 6Æ8% of negative lesions were found to be melanoma histopathologically.
Melanomas that do not fulfil the clinical ABCD criteria are difficult to diagnose. These include not only small, very early melanomas but also advanced nodular tumours that may mimic haemangioma, pyogenic granuloma, dermal naevus, blue naevus, dermatofibroma, or even seborrhoeic keratosis.1 Nodular melanoma (NM) is a biologically aggressive, potentially lethal tumour that grows rapidly.2 The standard definition of NM relies on histopathological criteria: a
melanoma growing vertically with only minimal horizontal spreading, and with intraepidermal growth not extending beyond the width of three rete ridges in any section.3 However, NM is usually defined based on clinical criteria, namely, a raised lesion, either pigmented or nonpigmented, with a minimal or no flat component. Dermoscopy has the potential to improve the preoperative recognition of ABCD-negative NMs but the current diagnostic
2011 The Authors BJD 2011 British Association of Dermatologists 2011 165, pp1251–1255
1252 Blue-black rule, G. Argenziano et al.
accuracy is quite disappointing, especially for nonpigmented melanomas. In a multicentre study, a simple model distinguishing amelanotic and hypomelanotic melanomas from all nonmelanomas had a sensitivity of 70% and a specificity of 56% in the test set.4 Various sensitive and specific features have been proposed for recognition of pigmented melanomas, but most of them were described in the context of superficial spreading melanoma, the most common type of melanoma, with no particular emphasis to the criteria specifically seen in pigmented NM.5 The aim of this study was to test whether pigmented NM could be recognized dermoscopically by considering the additional feature of blue and black colour within the lesion.
Materials and methods Cases were collected from the pathology databases of eight academic and private centres established in Australia, France, Italy and Spain. Lesions were included based on the following clinical presentation: a pigmented nodule with no (or only minimal) flat component. Each lesion was provided with clinical and high-resolution dermoscopic images. Patient age and sex, lesion location, histopathological diagnosis and Breslow thickness (in the case of melanoma) were collected in a data sheet. Ethics committee approval was waived. Two of us (D.T.-Z. and A.L.) evaluated each dermoscopic image on a computer screen for the presence of standard melanoma criteria and for the presence of a new feature named blue-black (BB) colour. This feature was defined as the presence of a combination of blue and black pigmented areas involving at least 10% of the lesion surface (Fig. 1). When the black component of the BB feature was represented by
clearly recognizable comedo-like openings (usually found in seborrhoeic keratoses) or lacunae (seen in haemangiomas), the BB feature was scored as absent (Fig. 2). Standard melanoma criteria were scored as present when at least one of the following was found: atypical network (including negative network), irregular streaks, regression structures, irregular brown globules, and irregular brown structureless areas. Bluewhite veil, black dots ⁄globules or blotches, vascular patterns and chrysalis structures (white shiny lines) were not included in the evaluation of standard melanoma criteria because they suggest melanoma only when associated with criteria more specific for the melanocytic nature of the lesion (network, streaks and brown globules).4–8 Evaluations were performed in a blinded fashion with respect to the histopathological diagnosis. The standard melanoma criteria and the BB feature were scored based on the agreement of the two observers. When no agreement was achieved by the two clinicians, a third observer (G.A.) was consulted. Sensitivity, specificity, positive predictive value and negative predictive value were calculated for the standard melanoma criteria and for the BB rule in relation to the diagnosis of melanoma and to diagnosis of malignancy. Diagnosis of malignancy was defined to include all lesions diagnosed by histopathological examination as malignant, i.e. melanoma, basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and Kaposi sarcoma.
Results We examined images of a total of 283 histopathologically proven pigmented nodular lesions (102 lesions imaged with
Fig 1. Six blue-black-positive nodular melanomas. Standard melanoma criteria such as atypical network, irregular streaks and irregular brown globules are not clearly visible. 2011 The Authors BJD 2011 British Association of Dermatologists 2011 165, pp1251–1255
Blue-black rule, G. Argenziano et al. 1253
Fig 2. Six blue-black-negative benign pigmented nodules. A bluish seborrhoeic keratosis (a) with black structures clearly recognized as comedolike openings. A bluish seborrhoeic keratosis (b) without black colour. A bluish haemangioma (c) with black structures clearly identified as lacunae. A pyogenic granuloma (d), a blue naevus (e) and a dermal naevus (f) exhibiting striking blue hue without black colour.
polarized light dermoscopy) from 280 patients (143 females; 51Æ1%). The mean patient age was 53Æ7 years (range 6–98; age not available in three patients). Lesions were located on the trunk (117 lesions), head ⁄neck (59 lesions), lower extremities (50 lesions), upper extremities (30 lesions), acral (19 lesions) and genital (three lesions) areas (body site not available in five lesions). Based on histopathological examination 160 lesions were classified as malignant tumours, comprising 78 (27Æ6%) melanomas, 75 (26Æ5%) BCCs, three (1Æ1%) SCCs and four (1Æ4%) Kaposi sarcomas. The remaining 123 lesions were benign, consisting of 57 (20Æ1%) melanocytic naevi (28 blue naevi, 22 dermal naevi, three combined naevi, three congenital naevi and one pigmented Spitz naevus), 26 (9Æ2%) haemangiomas ⁄angiokeratomas, 26 (9Æ2%) seborrhoeic keratoses, nine (3Æ2%) dermatofibromas and five (1Æ8%) pyogenic granuloTable 1 Diagnostic accuracy for melanoma and for any malignant lesion of three methods tested on 283 pigmented nodular lesions
mas. Among the 78 melanomas, 73 were primary lesions and five were cutaneous melanoma metastases. With the exception of one in situ melanoma arising in a pre-existing intradermal naevus, the melanomas were invasive (32 pure nodular melanomas and 39 melanomas with prevalent vertical growth), with mean Breslow thickness of 3Æ2 mm (range 0Æ5–17; thickness not available in one lesion). As shown in Table 1, presence of the BB feature had 78Æ2% sensitivity for melanoma compared with a sensitivity of 43Æ6% achieved by the standard melanoma criteria. Five of the 17 melanomas that were BB-negative were positive for standard melanoma criteria. A combined method based on the presence of either the BB feature or one (or more) of the standard melanoma criteria reached a sensitivity of 84Æ6%, with specificity of 80Æ5% and negative predictive value of 93Æ2%. The sensitivity for malignant lesions (including melanoma, BCC, SCC and
Sensitivity for melanoma Sensitivity for malignancy Specificity for melanoma Specificity for malignancy PPV for melanoma PPV for malignancy NPV for melanoma NPV for malignancy
Blue-black rule, n (%)
Standard criteria, n (%)
Combined method, n (%)
61 ⁄ 78 91 ⁄ 160 165 ⁄ 205 113 ⁄ 123 61 ⁄ 101 91 ⁄ 101 165 ⁄ 182 113 ⁄ 182
34 ⁄ 78 39 ⁄ 160 200 ⁄ 205 123 ⁄ 123 34 ⁄ 39 39 ⁄ 39 200 ⁄ 244 123 ⁄ 244
66 ⁄ 78 96 ⁄ 160 165 ⁄ 205 113 ⁄ 123 66 ⁄ 106 96 ⁄ 106 165 ⁄ 177 113 ⁄ 177
(78Æ2) (56Æ9) (80Æ5) (91Æ9) (60Æ4) (90Æ1) (90Æ7) (62Æ1)
PPV, positive predictive value; NPV, negative predictive value.
2011 The Authors BJD 2011 British Association of Dermatologists 2011 165, pp1251–1255
(43Æ6) (24Æ4) (97Æ6) (100) (87Æ2) (100) (82) (50Æ4)
(84Æ6) (60) (80Æ5) (91Æ9) (62Æ3) (90Æ6) (93Æ2) (63Æ8)
1254 Blue-black rule, G. Argenziano et al.
Kaposi sarcoma) was only 24Æ4%, 56Æ9% and 60% for standard criteria, presence of the BB feature and the combined method, respectively. The combined method gave specificity and positive predictive value for malignancy of 91Æ9% and 90Æ6%, respectively.
Discussion The most striking results of our study are the 90Æ6% positive predictive value for malignancy and the 93Æ2% negative predictive value for melanoma achieved using the combined method for the differentiation of pigmented nodular tumours. On one hand, only 9Æ4% of pigmented nodular lesions exhibiting either BB colour or any of the standard melanoma criteria were found to be benign histopathologically. On the other hand, the chance to leave a melanoma untreated is relatively low as only 6Æ8% of pigmented nodular lesions lacking both BB colour and the standard criteria were found to be melanoma histopathologically. The risk of leaving a melanoma untreated greatly increases in the absence of clinical ABCD criteria. Dermoscopy has proven to help clinicians particularly with very early melanomas that are too small to have developed the shape asymmetry, border irregularity and colour variegation that are typically found in larger lesions.9,10 The ABCD criteria can also be lacking in NMs that might instead score positive for the so-called EFG criteria.1 EFG refers to a recently developed mnemonic rule that outlines the most frequent clinical characteristics of NM, i.e. a lesion that is elevated and firm, with a history of rapid and progressive growth. However, the diagnostic accuracy of the EFG criteria is unknown, and clinicians may still miss this potentially lethal form of melanoma. The recognition of pigmented NM can also be challenging using dermoscopy, and few studies have attempted to address this problem. In a case series evaluating 11 thin NMs, most lesions had a homogeneous, disorganized, asymmetrical pattern or a featureless pattern with atypical vessels.11 Although many dermoscopic features seen in other melanoma subtypes were frequently absent in the lesions in that series, some features such as a blue-white veil, structureless areas and atypical vascular structures were often identified. In another study on pigmented NM, nonspecific global dermoscopic patterns, globules, blue-white veil, atypical vessels and structureless areas were frequently found in NMs and in nodular areas from superficial spreading melanomas.12 By definition, many features included in the standard melanoma criteria may not be expressed in NM because they relate to melanomas with a prominent superficial component. Atypical pigment network, irregular streaks, irregular brown globules, irregular brown structureless areas and regression features are histopathologically correlated with the presence of melanin within melanoma cells or melanophages that are confined to the epidermis and ⁄or the superficial dermis. In previous studies, blue-white veil and atypical vessels were predictive for thick melanoma, but their significance was strongly correlated with the presence of the basic melanocytic criteria.8
When the latter criteria, including pigment network, brown globules and streaks, are absent, the diagnostic relevance of the veil and the vessels is limited, as these two features can also be found in a variety of other tumours, both benign and malignant. Addition of the BB feature to the scoring criteria seems to have overcome this problem. The combination of blue and black colour means that the pigment is not only present in the mid-deep dermis (blue colour), as is sometimes seen in blue naevi and haemangiomas, but also is present in atypical melanocytes within the epidermis (black colour). Because black colour can also be found in seborrhoeic keratoses and haemangiomas, the BB feature was scored as negative when the black structures were clearly recognized as comedo-like openings or lacunae, the hallmarks for the diagnosis of seborrhoeic keratosis and haemangioma, respectively. Using this criterion, the presence of the BB feature showed 78Æ2% sensitivity and 80Æ5% specificity for melanoma. Sensitivity of the BB feature for any malignancy was only 56Æ9%, meaning that only a small proportion of BCCs, SCCs and Kaposi sarcomas was positive for this clue. The sensitivity of the standard criteria for melanoma was only 43Æ6%, confirming that these features are less likely to be seen in pigmented NM with no or only minimal flat component than they are likely to be seen in superficial melanomas. However, a combined method based on the presence of either the BB feature or one (or more) of the standard melanoma criteria reached a sensitivity of 84Æ6%, with specificity of 80Æ5%. In conclusion, the addition of the BB feature to the diagnostic armamentarium seems to be effective in improving the recognition of pigmented NM. To improve recognition of those melanomas that might still evade detection, further studies are needed, perhaps with inclusion of the EFG clinical criteria.
What’s already known about this topic? • Melanomas that do not fulfil the clinical ABCD criteria are difficult to diagnose. • Dermoscopy improves melanoma recognition, but most criteria were described in the context of superficial spreading melanoma.
What does this study add? • Using a method based on the blue-black feature or one of the standard melanoma criteria, only 9Æ4% of positive pigmented nodular lesions were found to be benign and only 6Æ8% of negative lesions were found to be melanoma histopathologically.
Acknowledgments We are indebted to Barbara J. Rutledge, PhD for editing assistance. 2011 The Authors BJD 2011 British Association of Dermatologists 2011 165, pp1251–1255
Blue-black rule, G. Argenziano et al. 1255
References 1 Chamberlain AJ, Fritschi L, Kelly JW. Nodular melanoma: patients’ perceptions of presenting features and implications for earlier detection. J Am Acad Dermatol 2003; 48:694–701. 2 Liu W, Dowling JP, Murray WK et al. Rate of growth in melanomas: characteristics and associations of rapidly growing melanomas. Arch Dermatol 2006; 142:1551–8. 3 Clark WH, From L, Bernardino EA, Mihm MC. The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res 1969; 29:705–27. 4 Menzies SW, Kreusch J, Byth K et al. Dermoscopic evaluation of amelanotic and hypomelanotic melanoma. Arch Dermatol 2008; 144:1120–7. 5 Argenziano G, Soyer HP, Chimenti S et al. Dermoscopy of pigmented skin lesions: results of a consensus meeting via the internet. J Am Acad Dermatol 2003; 48:679–93. 6 Marghoob AA, Cowell L, Kopf AW, Scope A. Observation of chrysalis structures with polarized dermoscopy. Arch Dermatol 2009; 145:618.
2011 The Authors BJD 2011 British Association of Dermatologists 2011 165, pp1251–1255
7 Zaballos P, Daufı´ C, Puig S et al. Dermoscopy of solitary angiokeratomas: a morphological study. Arch Dermatol 2007; 143:318– 25. 8 Argenziano G, Fabbrocini G, Carli P et al. Clinical and dermatoscopic criteria for the preoperative evaluation of cutaneous melanoma thickness. J Am Acad Dermatol 1999; 40:61–8. 9 Argenziano G, Zalaudek I, Ferrara G et al. Dermoscopy features of melanoma incognito: indications for biopsy. J Am Acad Dermatol 2007; 56:508–13. 10 Bono A, Tolomio E, Trincone S et al. Micro-melanoma detection: a clinical study on 206 consecutive cases of pigmented skin lesions with a diameter < or = 3 mm. Br J Dermatol 2006; 155:570–3. 11 Kalkhoran S, Milne O, Zalaudek I et al. Historical, clinical, and dermoscopic characteristics of thin nodular melanoma. Arch Dermatol 2010; 146:311–18. 12 Segura S, Pellacani G, Puig S et al. In vivo microscopic features of nodular melanomas: dermoscopy, confocal microscopy, and histopathologic correlates. Arch Dermatol 2008; 144:1311–20.