Dermatoscopy of basal cell carcinoma: Morphologic variability of global and local features and accuracy of diagnosis

Transcription

1 Dermatoscopy of basal cell carcinoma: Morphologic variability of global an local features an accuracy of iagnosis Davie Altamura, MD, a Scott W. Menzies, MBBS, PhD, b Giuseppe Argenziano, MD, c Iris Zalauek, MD, H. Peter Soyer, MD, FACD, e Francesco Sera, DStat, f Michelle Avramiis, BSc, b Kathryn DeAmbrosis, MBBS, e MariaConcettaFargnoli,MD, a an Ketty Peris, MD a L Aquila, Naples, an Florence, Italy; Syney an Brisbane, Australia; an Graz, Austria Backgroun: Early etection of basal cell carcinoma (BCC) is crucial to reuce the morbiity of this tumor. Objective: We sought to investigate the variability an iagnostic significance of ermatoscopic features of BCCs. Methos: We conucte retrospective ermatoscopic analysis of 609 BCCs an 200 melanocytic an nonmelanocytic lesions, an assessment of interrater reliability of ermatoscopic BCC criteria. Results: Lesions inclue nonpigmente (15.1%), lightly pigmente (33.2%), pigmente (42.7%), an heavily pigmente (9%) BCCs. Classic BCC patterns incluing arborizing (57.1%), blue/gray ovoi nests (47.5%), ulceration (39.2%), multiple blue/gray globules (26.1%), leaflike areas (15.9%), an spoke-wheel areas (9%) were significantly increase in pigmente BCCs compare with nonpigmente an heavily pigmente BCCs (P =.0001). Among nonclassic BCC patterns, we etecte short fine superficial (10%) an multiple small erosions (8.5%), an escribe two new patterns name concentric structures (7.6%) an multiple in-focus blue/gray ots (5.1%). Dermatoscopic features suggestive of melanocytic lesions (eg, multiple brown to black ots/globules, blue/white veillike structures, an nonarborizing vessels) were observe in 40.6% BCCs an significantly increase in heavily pigmente BCCs (P \.0001). Expert observers provie an accurate (sensitivity: 97%) an reliable (K: 87%) ermatoscopic iagnosis of BCC, although a significant ifference in terms of specificity (P =.0002) an positive preictive value (P =.0004) was foun. Arborizing, leaflike areas, an large blue/gray ovoi nests represente reliable an robust iagnostic parameters. Limitation: The stuy was retrospective. Conclusion: BCCs show a large spectrum of global an local ermatoscopic features; heavily pigmente BCCs show the most challenging combinations of ermatoscopic features. ( J Am Aca Dermatol 2010;62:67-75.) Key wors: basal cell carcinoma; ermatoscopy; early iagnosis of skin tumors. Basal cell carcinoma (BCC) is a slowly growing malignant epithelial skin tumor preominantly affecting mile-age an fair-skinne iniviuals. 1 Clinicopathologic appearances of BCC are various an inclue noular, superficial, morpheic, an pigmente variants. 1,2 Epiemiologic ata inicate that the global incience is increasing worlwie particularly in younger age groups an From the Department of Dermatology, University of L Aquila a ; Syney Melanoma Diagnostic Center, Syney Cancer Center, Royal Prince Alfre Hospital, Camperown, an Discipline of Dermatology, University of Syney b ; Department of Dermatology, Secon University of Naples c ; Department of Dermatology, Meical University of Graz ; Dermatology Research Center, The University of Queenslan, School of Meicine, Princess Alexanra Hospital, Brisbane e ; an Molecular an Nutritional Epiemiology Unit, Cancer Prevention an Research Institute, Florence. f Funing sources: None. Conflicts of interest: None eclare. Accepte for publication May 17, Reprint requests: Ketty Peris, MD, Department of Dermatology, University of L Aquila, Via Vetoio - Coppito 2, L Aquila, Italy. ketty.peris@univaq.it. Publishe online October 14, /$36.00 ª 2009 by the American Acaemy of Dermatology, Inc. oi: /j.jaa

2 68 Altamura et al JAM ACAD DERMATOL JANUARY 2010 BCC emerges as a growing public health problem. 3 The ifferential iagnosis between BCC an other skin lesions is of critical importance, an serious morbiity may result from an uniagnose tumor. Noninvasive proceures such as ermatoscopy, high-frequency ultrasoun, optical coherence tomography, an reflectance confocal microscopy have been evelope for the iagnosis of skin cancers. 4-7 Among these, ermatoscopy represents so far the iagnostic tool with the highest clinical impact in ermatologic practice to better ifferentiate benign from malignant skin lesions an to etect tumors in an early stage. 6,7 The ermatoscopic moel for the iagnosis of the pigmente variant of BCC is base on the absence of a pigmente network an the presence of at least one positive feature incluing: (1) ulceration (not associate with a recent history of trauma); (2) multiple blue/ gray globules; (3) leaflike areas; (4) large blue/gray ovoi nests; (5) spoke-wheel areas; an (6) arborizing CAPSULE SUMMARY (treelike). 8 However, BCC may exhibit a large variety of clinical an ermatoscopic characteristics that are the result of a wie range of combinations of histopathological features. 1,8 Different case reports escribe BCCs with unusual clinical an ermatoscopic aspects an recent stuies emphasize the possible role of aitional features such as short fine an multiple small ulcerations as further criteria for the ermatoscopic iagnosis of superficial BCC The aims of this stuy were to: (1) escribe the morphologic variability of the ermatoscopic features in a large set of BCCs; (2) efine the istribution of ifferent ermatoscopic features in specific types of BCC; an (3) evaluate the accuracy an reliability of ermatoscopic iagnosis an the iagnostic significance of BCC-specific patterns. METHODS Lesion selection an ermatoscopic equipment Dermatoscopic images of the skin lesions inclue in this stuy were from the igital atabases Classic basal cell carcinoma (BCC) patterns were significantly increase in pigmente BCCs compare with nonpigmente an heavily pigmente BCCs. Dermatoscopic features suggestive of melanocytic lesions were observe in 40.6% of BCCs an significantly increase in heavily pigmente BCCs. Expert observers provie an accurate (sensitivity: 97%) an reliable (K: 87%) ermatoscopic iagnosis of BCC. Arborizing, leaflike areas, an large blue/gray ovoi nests represente reliable an robust iagnostic parameters. Heavily pigmente BCCs show the most challenging combinations of ermatoscopic features. of the outpatient clinics of the Departments of Dermatology of the University of L Aquila (L Aquila, Italy) an of the tertiary referral center of the Syney Melanoma Diagnostic Center (Syney, Australia). Lesions were collecte between January 1991 an May 2007 using 3 pieces of ermatoscopic equipment: (1) SolarScan system (Polartechnics Lt, Syney, Australia); (2) Nevuscreen system (Arkè s.a.s., Avezzano, Italy); an (3) Dermaphot (Heine Lt, Herrshing, Germany). Large lesions present on the atabases but not completely comprise within the fiel of view were not inclue in the stuy. Demographic ata such as age, sex, an anatomic site of the lesions were also collecte for each patient. Dermatoscopic criteria of BCC In the first phase of the stuy, ermatoscopic images of 609 histopathologically proven BCCs were score for global an local ermatoscopic features as previously escribe. 8,16 Iniviual lesions were categorize accoring to the egree of ermatoscopic pigmentation as: (1) nonpigmente BCCs, characterize by the absence of ermatoscopic brown, black, blue, or gray pigmentation; (2) lightly pigmente BCCs, showing pigmente areas involving less than 30% of the lesion s surface; (3) pigmente BCCs, isplaying pigmente areas involving 30% to 70% of the lesion s surface; an (4) heavily pigmente BCCs, characterize by the presence of ermatoscopic pigmentation in more than 70% of the lesion s surface. Score colors inclue blue, black, gray, light brown, ark brown, an re. Re an/or white structureless areas were evaluate to efine the surface of nonpigmente regions. 8 Accoring to morphologic efinition, local ermatoscopic aspects that isplaye specific criteria to iagnose pigmente BCC were classifie as classic BCC patterns (ie, ulceration, multiple blue/gray globules, leaflike areas, large blue/gray ovoi nests, spoke-wheel areas, an arborizing ). 8 Local features not classifiable in this category but representing a possible variation on the theme of the above patterns were categorize as nonclassic BCC

3 JAM ACAD DERMATOL VOLUME 62, NUMBER 1 Altamura et al 69 Table I. Morphologic efinition of classic basal cell carcinoma patterns an nonclassic basal cell carcinoma patterns Dermatoscopic pattern Definition Diagnostic significance Classic BCC patterns Arborizing (treelike) Telangiectasia with istinct treelike ramifications 8 Pigmente BCC Large blue/gray ovoi Well-circumscribe, confluent or near confluent Pigmente BCC nests pigmente ovoi or elongate areas, larger than globules an not intimately connecte to pigmente tumor boy 8 Ulceration Absence of epiermis often associate with congeale Pigmente BCC bloo an without recent history of trauma 8 Multiple blue/gray Multiple globules (not ots) that shoul be ifferentiate Pigmente BCC globules from multiple blue/gray ots (melanophages) 8 Maple-leaflike areas Brown to gray/blue iscrete bulbous extensions forming leaflike pattern (never arising from pigmente network an from ajacent confluent pigmente area) 8 Pigmente BCC Spoke-wheel areas Nonclassic BCC patterns Short fine superficial Multiple small erosions Concentric structures Multiple in-focus blue/gray ots Well-circumscribe raial projections, usually tan but sometimes blue or gray, meeting at often arker (ark brown, black, or blue) central axis 8 Irregularly isperse, kinke vessels of small caliber an length ( # 1 mm) without clear-cut treelike ramifications (with relatively few branches); they appear sharply focuse an are mainly locate in white to re backgroun appearance 14 $ 5 Superficial erosion with maximum iameter # 1mm characterize by complete or partial absence of epiermis; they appear brown pigmente an ranomly locate throughout lesion surface 14 Irregularly shape globular-like structures with ifferent colors (blue, gray, brown, black) an arker central area Foci of multiple blue/gray ots that appear in focus at ermatoscopic examination; they may be ifferentiate from multiple blue/gray ots of melanophages because of lack of focus an pepper-like appearance in latter Pigmente BCC Superficial BCC (hypothesize early phase of arborizing ) Superficial BCC (hypothesize early phase of ulceration) Pigmente BCC (hypothesize early phase of spoke-wheel areas) Pigmente BCC (hypothesize early phase of multiple blue/gray globules) BCC, Basal cell carcinoma. patterns (Table I). Finally, ermatoscopic patterns more frequently observe in melanocytic lesions compare with pigmente BCCs were inclue in the category of melanocytic patterns. 8 Test set of lesions In the secon phase of the stuy, the interrater reliability of ermatoscopic features an iagnosis of BCC was assesse among 3 observers experience in ermatoscopic evaluation (G. A., I. Z., an H. P. S.). The test set comprise 300 skin lesions incluing 150 BCCs (38 pigmente, 38 heavily pigmente, 37 nonpigmente, an 37 lightly pigmente), 50 melanomas (meian Breslow thickness 0.4 mm; range mm), 50 melanocytic nevi, an 50 nonmelanocytic skin lesions, with a similar egree an istribution of pigmentation, ranomly selecte from our igital atabases (Table II). The test set was presente in a mixe orer to the observers, who score the images bline to the histopathologic iagnosis an without knowlege of any clinical ata of the patients an lesions. Observers were first aske to make the iagnosis (BCC vs not BCC) of each lesion; subsequently, if a BCC was iagnose, observers reporte the BCC ermatoscopic features categorize as classic an nonclassic BCC patterns. Statistical analysis Differences between proportions of contingency tables were assesse using the Chi-square test. Chisquare statistic for tren was use to test the null hypothesis of no association between proportion of

4 70 Altamura et al JAM ACAD DERMATOL JANUARY 2010 Table II. Diagnosis an frequency of lesions inclue in test set Diagnosis No. (%) BCCs 150 (50) Nonpigmente 37 (16.3) Lightly pigmente 37 (16.3) Pigmente 38 (12.6) Heavily pigmente 38 (12.6) Melanomas 50 (16.6) Superficial spreaing 31 (10.3) In situ 10 (3.3) Noular 9 (3.0) Melanocytic nevi 50 (16.6) Atypical 28 (9.3) Spitz/Ree 9 (3.0) Blue 5 (1.6) Dermal 5 (1.6) Compoun 3 (1) Nonmelanocytic lesions 50 (16.6) Seborrheic keratosis 20 (6.6) Actinic keratosis 12 (4.0) Dermatofibroma 10 (3.3) Hemangioma 4 (1.3) Squamous cell carcinoma 2 (0.6) Eccrine poroma 1 (0.3) Viral wart 1 (0.3) BCC, Basal cell carcinoma. variation an categorical variables. All P values cite are two-sie, an values of P less than.05 were consiere statistically significant. Sensitivity, specificity, positive preictive value, an negative preictive value for the ermatoscopic iagnosis of BCC were calculate for each observer as compare with histopathologic iagnosis. Classic an nonclassic BCC patterns were consiere as positive ermatoscopic features to iagnose BCC. The agreement between ratings mae by 3 observers on ermatoscopic iagnosis an specific patterns of BCC (interrater reliability) was estimate using Cohen kappa statistic with 95% confience intervals. 17,18 Populations with trait prevalence closer to 50% were consiere well-balance inices for more reliable K values. 1 RESULTS Dermatoscopic images of BCCs were from 609 iniviuals, incluing 357 male (58.6%) an 252 female (41.4%) patients with a meian age of 63 years (range: years). Iniviual lesions were locate on the face (173/609; 28.4%), back (168/609; 27.6%), front of the trunk (154/609; 25.3%), scalp (59/609; 9.7%), lower (21/609; 3.4%) an upper (17/609; 2.8%) extremities, an neck (17/609; 2.8%). The populations inclue preominantly non-asian or non-black-skinne iniviuals. Global ermatoscopic features Lesions inclue 260 of 609 (42.7%) pigmente, 202 of 609 (33.2%) lightly pigmente, 92 of 609 (15.1%) nonpigmente, an 55 of 609 (9%) heavily pigmente BCCs. The colors observe by ermatoscopy were blue (495/609; 81.3%), re (430/609; 70.6%), light brown (321/609; 52.7%), gray (299/609; 49.1%), ark brown (167/609; 27.4%), an black (65/609; 10.7%). Multiple colors ( $ 3) were ientifie in 276 of 609 (45.3%) lesions. Nonpigmente an lightly pigmente BCCs exhibite a shiny white to re backgroun in 245 of 294 (83.3%) lesions, an a white scarlike epigmentation appearance in 49 of 294 (16.6%) lesions. Heavily pigmente BCCs showe a multicolore backgroun in 37 of 55 (67.3%) lesions, an a single color backgroun in 5 of 55 (9.1%) lesions. Local ermatoscopic features Dermatoscopic evaluation showe the presence of classic BCC patterns in 583 of 609 (95.7%) lesions. In all, 499 of 517 (96.5%) pigmente BCCs ha the overall pattern of an absent network an one or more positive feature as previously escribe. 8 The most common pattern was the arborizing (treelike), ientifie in 348 of 609 (57.1%) BCCs. Large blue/gray ovoi nests were etecte in 289 of 609 (47.5%) lesions, ulceration was seen in 239 of 609 (39.2%) lesions, an multiple blue/gray globules were observe in 159 of 609 (26.1%) lesions. In all, 97 (15.9%) an 55 (9%) of the 609 BCCs showe leaflike areas an spoke-wheel areas, respectively (Fig 1). In all, 499 of 517 (96.5%) pigmente BCCs ha the overall pattern of an absent pigmente network an one or more positive feature as previously escribe. 8 Nonclassic BCC patterns were shown in 159 of 609 (26.1%) BCCs an inclue: (1) short fine superficial (61/609; 10%); (2) multiple small erosions (52/609; 8.5%); (3) concentric structures (46/609; 7.6%); an (4) multiple in-focus blue/- gray ots (not pepper-like) (31/609; 5.1%) (Fig 2). Among pigmente BCCs, the overall ermatoscopic appearance of 9 of 517 (1.8%) lesions showe the presence of one nonclassic BCC pattern in the absence of pigmente network an classic positive criteria. Dermatoscopic patterns inicative of melanocytic lesions were foun in 247 of 609 (40.6%) BCCs. In such cases the following features were etecte: (1) multiple brown to black ots/globules (132/609; 21.6%); (2) blue/white veillike structures (87/609; 14.3%); (3) pigmente network (11/609; 1.8%);

5 JAM ACAD DERMATOL VOLUME 62, NUMBER 1 Altamura et al 71 Fig 1. Classic basal cell carcinoma criteria inclue: arborizing (treelike) (/), large blue/gray ovoi nests (v), ulceration (m), multiple blue/gray globules (-), mapleleaflike areas (C), an spoke-wheel areas (t). (4) multiple blue/gray ots (pepper-like) (7/609; 1.2%); an (5) raial streaming or pseuopos (4/609; 0.6%). In aition, nonarborizing vascular features were observe in 70 of 609 BCCs (11.5%) an inclue otte vessels (45/609; 7.4%), linear irregular vessels (30/609; 4.9%), hairpin vessels (16/609; 2.6%), an comma vessels (6/609; 1%). Distribution of ermatoscopic patterns Classic BCC patterns were significantly more frequent in lightly pigmente an pigmente types of BCC (P =.0001) than in nonpigmente an heavily pigmente BCCs. In contrast, melanocytic patterns were significantly more frequent in heavily pigmente BCCs as compare with the other variants (P \.0001) (Fig 3). No significant ifference in terms of istribution of nonclassic BCC patterns in ifferent types of BCC was foun. In aition, the frequency of melanocytic patterns significantly increase with the increasing rate of the pigmentation of BCCs (test for tren P value \.0001). Table III summarizes the overall istribution of ermatoscopic patterns accoring to the ifferent types of BCC. Accuracy an interrater reliability of ermatoscopic iagnosis of BCC Sensitivity, specificity, positive preictive value, an negative preictive value for the ermatoscopic iagnosis of BCC obtaine by observers 1, 2, an 3 are reporte in Table IV. There was no significant ifference between the sensitivity an negative preictive value between observers 1, 2, an 3. A significant ifference was etecte between specificity an positive preictive value (P =.0002 an P =.0004, respectively). The interrater reliability of ermatoscopic iagnosis of BCC obtaine a Cohen kappa statistic value of.87 (95% confience interval: ). 17,18 Table V summarizes the overall prevalence of ermatoscopic BCC criteria obtaine by the observers an the interrater reliability of ratings. DISCUSSION In this stuy we investigate the ermatoscopic variability of global an local features of BCCs an analyze the interobserver agreement (interrater reliability) on patterns an iagnosis of this neoplasm. We valiate the ermatoscopic metho previously escribe to iagnose pigmente BCC, 8 an showe the presence of the 6 classic BCC ermatoscopic criteria in 96.5% of our set of pigmente BCCs. The frequency of multiple blue/gray globules, leaflike areas, an spoke-wheel areas was almost ientical to that reporte in the previous stuy, whereas we etecte a higher frequency of arborizing (treelike) an ulceration an a lower rate of large blue/gray ovoi nests. 8

6 72 Altamura et al JAM ACAD DERMATOL JANUARY 2010 Fig 2. Nonclassic basal cell carcinoma ermatoscopic criteria inclue: short fine superficial (m), multiple small erosions (/), concentric structures (v), an multiple infocus blue/gray ots (t). Fig 3. Two cases of heavily pigmente basal cell carcinomas (BCC) ermatoscopically characterize by prevalent istribution of brown to black ots an globules (/) an iffuse blue-white veillike structure (m). Interestingly, first lesion lacks pigmente network an has multiple blue/gray globules that woul satisfy classic pattern observe in pigmente BCC (t). Lesion selection coul account for these ifferences, as we inclue in our set both pigmente an nonpigmente types of BCC. BCC may also isplay ermatoscopic features commonly foun in melanocytic lesions such as brown to black ots/globules, blue/white veil, pigmente network, pseuopos, raial streaming, or a polymorphous vascular pattern. 9-12,16 In the current stuy, ermatoscopic features suggestive of melanocytic lesions were observe in 40.6% of BCCs an mainly inclue multiple brown to black ots/ globules, blue/white veillike structures, an vascular

7 JAM ACAD DERMATOL VOLUME 62, NUMBER 1 Altamura et al 73 Table III. Distribution of ermatoscopic patterns accoring to ifferent types of basal cell carcinoma BCC category Dermatoscopic patterns Nonpigmente, No. (%) Lightly pigmente, No. (%) Pigmente, No. (%) Heavily pigmente, No. (%) x2 test P value Test for tren P value Classic BCC patterns 84/92 (91.3) 199/202 (98.5) 254/260 (97.7) 46/55 (83.6) Arborizing (treelike) 77/92 (83.7) 127/202 (62.9) 126/260 (48.5) 18/55 (32.7) e e Large blue/gray 0/92 (0) 99/202 (49) 158/260 (60.8) 32/55 (58.2) e e ovoi nests Ulceration 45/92 (48.9) 93/202 (46) 90/260 (34.6) 11/55 (20) e e Multiple blue/gray 0/92 (0) 39/202 (19.3) 100/260 (38.5) 20/55 (36.4) e e globules Leaflike areas 0/92 (0) 24/202 (11.9) 59/260 (22.7) 14/55 (25.5) e e Spoke-wheel areas 0/92 (0) 17/202 (8.4) 34/260 (13.1) 4/55 (7.3) e e Melanocytic patterns 12/92 (13.0) 64/202 (31.7) 127/260 (48.8) 44/55 (80.0) \.0001 \.0001 Multiple brown to 0/92 (0) 31/202 (15.3) 87/260 (33.5) 14/55 (25.5) e e black ots/globules Blue/white veillike 0/92 (0) 8/202 (4) 45/260 (17.3) 33/55 (61.8) e e structures Nonarborizing vessels 11/92 (12) 30/202 (14.9) 27/260 (10.4) 2/55 (3.6) e e Pigmente network 0/92 (0) 3/202 (1.5) 5/260 (1.9) 3/55 (5.5) e e Multiple blue/gray 0/92 (0) 2/202 (1) 5/260 (1.9) 0/55 (0) e e ots (pepper-like) Raial streaming 0/92 (0) 1/202 (.5) 2/260 (.8) 1/55 (1.8) e e or pseuopos Nonclassic BCC patterns 17/92 (18.5) 63/202 (31.2) 65/260 (25.0) 14/55 (25.5) Short fine superficial 13/92 (14.1) 32/202 (15.8) 15/260 (5.8) 1/55 (1.8) e e Multiple small erosions 7/92 (7.6) 23/202 (11.4) 17/260 (6.5) 5/55 (9.1) e e Concentric structures 0/92 (0) 21/202 (10.4) 21/260 (8.1) 4/55 (7.3) e e Multiple in-focus blue/gray ots 0/92 (0) 12/202 (5.9) 15/260 (5.8) 4/55 (7.3) e e BCC, Basal cell carcinoma. Table IV. Sensitivity, specificity, positive preictive value, an negative preictive value for ermatoscopic iagnosis obtaine by observers 1, 2, an 3 in test set Observers No. 1 No. 2 No. 3 P value Sensitivity 95% 97% 95%.573 Specificity 96% 96% 87%.0002 PPV 96% 96% 88%.0004 NPV 95% 97% 94%.465 NPV, Negative preictive value; PPV, positive preictive value. features such as otte, linear irregular, hairpin, an comma vessels. Remarkably, we showe that the frequency of melanocytic patterns linearly increase with the pigmentation of the lesions, having a prevalent istribution in the heavily pigmente BCCs. These finings suggest that, among pigmente BCCs, the heavily pigmente variant represents the most ifficult type to be ifferentiate from both melanocytic nevi an melanomas, an confirm that the global aspect shoul always be evaluate in the ermatoscopic interpretation of the lesions. Local ermatoscopic features classifie in this stuy as nonclassic BCC patterns inclue short fine, which were observe in 10% of BCCs an have been interprete as an early variant of the arborizing vessels. In aition, multiple small erosions were etecte in 8.5% of BCCs an thought to represent the initial stage of clear-cut ulceration. Consistent with our observation, short fine an multiple small ulcerations ranomly locate in a shiny white to re structureless backgroun have been reporte as aitional ermatoscopic criteria for the iagnosis of superficial BCC. 14,15 Other nonclassic BCC patterns to our knowlege escribe for the first time in our stuy inclue concentric structures an multiple in-focus blue/gray ots, which were seen in 7.6% an 5.1% of the lesions, respectively. Concentric structures appeare as roun, concentrically overlappe

8 74 Altamura et al JAM ACAD DERMATOL JANUARY 2010 Table V. Overall observers prevalence inex of ermatoscopic basal cell carcinoma patterns an interrater reliability of ratings (K coefficient) Dermatoscopic BCC patterns Prevalence inex, % K coefficient 95% CI Large blue/gray ovoi nests ( ) Multiple blue/gray globules ( ) Short fine superficial ( ) Arborizing (treelike) ( ) Leaflike areas ( ) Multiple in-focus blue/gray ( ) ots Ulceration ( ) Multiple small erosions ( ) Concentric structures ( ) Spoke-wheel areas ( ) CI, Confience interval; BCC, basal cell carcinoma. structures that resemble the early stage of a spokewheel area lacking the characteristic peripheral raial projection. Furthermore, multiple in-focus blue/gray ots represente foci of sharply focuse ermatoscopic features that were ifferentiate from the pepper-like ermatoscopic appearance of melanophages an from black ots of melanocytic lesions. Base on our results, we hypothesize that the nonclassic BCC criteria escribe herein may further support the iagnosis of BCC, particularly in early lesions that may lack BCC classic patterns. The ermatoscopic moel propose by Menzies et al 8 to iagnose pigmente BCCs is inee not an effective iscriminator for lesions lacking pigmentation. In the current stuy we also examine the ermatoscopic features of 92 nonpigmente BCCs showing that the arborizing was the most common criterion, followe by ulceration, short fine superficial, nonarborizing vessels, an multiple small erosions. We believe that features such as ulceration or small erosions an arborizing or short fine superficial might be useful to iagnose nonpigmente variant of BCC, although further stuies are neee to valiate these observations. Our group previously emonstrate the substantial reproucibility of the ermatoscopic iagnostic criteria 19 propose by Menzies et al. 8 Interestingly, the K statistic coefficient is consiere an appropriate measure of the reliability of the clinicians ratings in areas such as iagnosis an the interpretation finings. 17,18 In the current stuy, all the observers provie high levels of iagnostic sensitivity for BCC, an the ermatoscopic iagnosis was consiere substantially reliable because of the 87% of iagnostic K agreement. However, the prevalence inex of ratings may influence the magnitue of K coefficient, an a more appropriate estimation of the reliability of finings can be obtaine with trait prevalence closer to 50%. 20 In fact, among ermatoscopic patterns, arborizing, leaflike areas, an large blue/gray ovoi nests seem to represent reliable an robust parameters to iagnose BCC, whereas short fine superficial appears as a less reproucible iagnostic criterion. In conclusion, we escribe a large spectrum of global an local ermatoscopic features of BCC, an we reporte the 3 most robust an reliable BCC parameters (ie, arborizing, leaflike areas, an large blue/gray ovoi nests). In aition, we emonstrate that heavily pigmente BCCs show the most challenging combinations of ermatoscopic features. REFERENCES 1. Roewert-Huber J, Lange-Asschenfelt B, Stockfleth E, Kerl H. Epiemiology an etiology of basal cell carcinoma. Br J Dermatol 2007;157(Suppl): Leiter U, Garbe C. Epiemiology of melanoma an nonmelanoma skin cancerethe role of sunlight. Av Exp Me Biol 2008;624: Telfer NR, Colver GB, Morton CA. Guielines for management of basal cell carcinoma. Br J Dermatol 2008;159: Kittler H, Pehamberger H, Wolff K, Biner M. Diagnostic accuracy of ermoscopy. Lancet Oncol 2002;3: Marghoob AA, Swinle LD, Moricz CZ, Sanchez Negron FA, Slue B, Halpern AC, et al. Instruments an new technologies for the in vivo iagnosis of melanoma. J Am Aca Dermatol 2003;49: Mogensen M, Jemec JB. Diagnosis of nonmelanoma skin cancer/keratinocyte carcinoma: a review of iagnostic accuracy of nonmelanoma skin cancer iagnostic tests an technologies. Dermatol Surg 2007;33: Ulrich M, Stockfleth E, Roewert-Huber J, Astner S. Noninvasive iagnostic tools for nonmelanoma skin cancer. Br J Dermatol 2007;157(Suppl): Menzies SW, Ingvar C, Crotty KA, McCarthy WH. Surface microscopy of pigmente basal cell carcinoma. Arch Dermatol 2000;136: Ferrari A, De Angelis L, Peris K. Unusual clinical an ermoscopic features in two cases of pigmente basal cell carcinoma. J Am Aca Dermatol 2005;53: Feler S, Rabinovitz H, Oliviero M, Kopf A. Dermoscopic ifferentiation of a superficial basal cell carcinoma an squamous cell carcinoma in situ. Dermatol Surg 2006;32: Feler S, Rabinovitz H, Oliviero M, Kopf A. Dermoscopic pattern of pigmente basal cell carcinoma, blue-white variant. Dermatol Surg 2006;32: Rossiello L, Zalauek I, Cabo H, Ferrara G, Gabriel C, Argenziano G. Dermoscopic-pathologic correlation in an unusual case of pigmente basal cell carcinoma. Dermatol Surg 2006;32:

9 JAM ACAD DERMATOL VOLUME 62, NUMBER 1 Altamura et al Zalauek I, Ferrara G, Broganelli P, Moscarella E, Morente I, Giacomel J, et al. Dermoscopy patterns of fibroepithelioma of pinkus. Arch Dermatol 2006;142: Giacomel J, Zalauek I. Dermoscopy of superficial basal cell carcinoma. Dermatol Surg 2005;31: Scalvenzi M, Lembo S, Francia MG, Balato A. Dermoscopic patterns of superficial basal cell carcinoma. Int J Dermatol 2008;47: Menzies SW, Ingvar C, Crotty K, McCarthy W. Frequency an morphologic characteristics of invasive melanomas lacking specific surface microscopic features. Arch Dermatol 1996;132: Fleiss JL. Statistical methos for rates an proportions. 2n e. New York: John Wiley an Sons; p Sim J, Wright CC. The Kappa statistic in reliability stuies: use, interpretation an sample size requirements. Phys Ther 2005; 85: Peris K, Altobelli E, Ferrari A, Fargnoli MC, Piccolo D, Esposito M, et al. Interobserver agreement on ermoscopic features of pigmente basal cell carcinoma. Dermatol Surg 2002;28: Hoehler FK. Bias an prevalence effects on kappa viewe in terms of sensitivity an specificity. J Clin Epiemiol 2000;53: