JEADV
DOI: 10.1111/jdv.12769
ORIGINAL ARTICLE
Dermoscopic difficult lesions: an objective evaluation of reflectance confocal microscopy impact for accurate diagnosis B. Ferrari,1,* G. Pupelli,1 F. Farnetani,1 N.T. De Carvalho,1 C. Longo,2 C. Reggiani,1 G. Argenziano,2 G. Pellacani1 1
Department of Dermatology, University of Modena and Reggio Emilia, Modena, Italy Dermatology and Skin Cancer Unit, Arcispedale Santa Maria Nuova, Istituto di Ricovero e Cura a Carattere Scientifico-IRCCS, Reggio Emilia, Italy *Correspondence: B. Ferrari. E-mail: [email protected]
2
Abstract Background Early detection of melanoma is the main objective to ensure a high survival rate. In some cases melanoma diagnosis still remain difficult and this leads to unnecessary excisions. Objective The aim of this study was to detect the most relevant Reflectance confocal microscopy (RCM) features for the detection of dermoscopic difficult melanomas. Method A total of 322 lesions were selected from database and were evaluated on dermoscopy according to the 7-point checklist score, in blind from histological diagnosis. We classified the lesions into three categories: (i) ‘featureless’ lesions with score ranging between 0 and 2; (ii) ‘positive-borderline’ moles with score between 3 and 4 and (iii) ‘positive-clear cut’ lesions with score from 5 to 10. We evaluated confocal features of the ‘featureless’ lesions and of the ‘positive-borderline’ lesions. Evaluated confocal features were as follows: presence of pagetoid cells, cell shape (roundish or dendritic) and number (< 5 or >5 cells per mm2), overall architecture (ringed, meshwork, clods and non-specific pattern); architectural disorder, presence of cytological atypia (>5 cells per mm2) and cells arranged in nests. Results Among 322 lesions 70 were melanomas and 252 were nevi. According to the classification based on the 7point checklist score, 130 ‘featureless lesions’ (score 0–2) including six melanomas, and 102 ‘positive-borderline’ moles (score 3–4) including 17 melanomas, were identified. Round pagetoid cells >5 cells per mm2 and/or architectural disorder on RCM were found in all of six melanomas with featureless dermoscopy. Round pagetoid infiltration and five or more atypical cells at the DEJ were found in 16 positive ‘borderline melanomas’. Conclusions RCM represents a rapid non-invasive technique that can aid early diagnosis of dermoscopic difficult melanomas. Use of RCM on lesions with clinical and/or dermoscopic suspect of malignancy may reduce the number of unnecessary excision increasing the rate of accurate diagnoses. Received: 20 June 2014; Accepted: 9 September 2014
Conflicts of interest None declared.
Funding sources None declared.
Introduction The worldwide incidence of cutaneous malignant melanoma has been rising steadily over the past 30 years.1,2 As excision of thin melanoma guarantee a high rate of survival, compared with high risk of progression and a poor prognosis for advanced stages, early diagnosis still represents the most effective approach. The conventional clinical diagnosis of melanoma is based on some observable criteria that can be summarized into ABCDE
JEADV 2014
(asymmetry, irregular borders, multiple colours, diameter >6 mm, enlarging lesion).3 However, the ABCDE method only provides 64% accuracy, indicating a need for additional diagnostic tools.4 In the last decades, many efforts have been directed towards an early detection of melanoma and dermoscopy is worldwide employed as screening tool giving significant improvement of diagnostic accuracy.5 However, in some cases
© 2014 European Academy of Dermatology and Venereology
Ferrari et al.
2
melanoma diagnosis still remain difficult, due to lack of characteristic dermoscopic features, inducing a consistent rate of excision of benign lesions, corresponding to a number needed to excise (NNE) value that ranges from 8.7 to 29.4, according to the level of expertise and the setting.6 Reflectance confocal microscopy (RCM) is a second-level novel non-invasive technique that enables the in vivo visualization of the skin at quasi-histological resolution. In retrospective settings it showed to improve diagnostic accuracy when combined with dermoscopy.7,8 Moreover, prospective studies showed its capability to reduce the NNE both for highly suspicious lesions9 and for equivocal ones that may be referred to digital dermoscopic monitoring.10. As RCM features are strongly related with dermoscopic patterns,11 and the combination may help to increase the diagnostic accuracy, in this study we aimed to detect the most relevant RCM features for the detection of dermoscopic difficult melanomas. To increase objectivity, lesions were categorized according their dermoscopic score, as calculated by the 7-point checklist,12 and main RCM clues suggestive of melanoma diagnosis have been evaluated on a series of consecutive cases referred to RCM examination.
Materials and methods Study samples included all melanocytic lesions excised on the basis of equivocal clinical and/or dermoscopic features in 2010. Before excision, all lesions were recorded by means of digital dermoscopy and RCM at the Department of Dermatology, University of Modena and Reggio Emilia. Only lesions with high quality dermoscopic images, a complete set of confocal images and histopathology report available were included in the study. After dermoscopic evaluation of 7-point checklist score,12 single dermoscopic and RCM features were considered for lesions with a score of 4 or lower (corresponding to dermoscopic borderline or negative lesions).
histological diagnosis, by a dermatologist trained in dermoscopy and RCM (BF). Dermoscopic criteria
The 7-point checklist score was calculated for each case12 as well as the frequencies of each distinct dermoscopic feature accounting for the score. Afterwards, lesions were classified according the 7-point checklist score into three categories: ‘featureless’ lesions for score ranging between 0 and 2, ‘positive-borderline’ lesions for score between 3 and 4 and ‘positive- clear cut’ lesions for score from 5 to 10. Confocal criteria
Previously published RCM parameters demonstrated useful for melanoma detection have been selected. In the superficial layer it was evaluated the presence of pagetoid cells, the cell shape (roundish or dendritic) and their number (
DOI: 10.1111/jdv.12769
ORIGINAL ARTICLE
Dermoscopic difficult lesions: an objective evaluation of reflectance confocal microscopy impact for accurate diagnosis B. Ferrari,1,* G. Pupelli,1 F. Farnetani,1 N.T. De Carvalho,1 C. Longo,2 C. Reggiani,1 G. Argenziano,2 G. Pellacani1 1
Department of Dermatology, University of Modena and Reggio Emilia, Modena, Italy Dermatology and Skin Cancer Unit, Arcispedale Santa Maria Nuova, Istituto di Ricovero e Cura a Carattere Scientifico-IRCCS, Reggio Emilia, Italy *Correspondence: B. Ferrari. E-mail: [email protected]
2
Abstract Background Early detection of melanoma is the main objective to ensure a high survival rate. In some cases melanoma diagnosis still remain difficult and this leads to unnecessary excisions. Objective The aim of this study was to detect the most relevant Reflectance confocal microscopy (RCM) features for the detection of dermoscopic difficult melanomas. Method A total of 322 lesions were selected from database and were evaluated on dermoscopy according to the 7-point checklist score, in blind from histological diagnosis. We classified the lesions into three categories: (i) ‘featureless’ lesions with score ranging between 0 and 2; (ii) ‘positive-borderline’ moles with score between 3 and 4 and (iii) ‘positive-clear cut’ lesions with score from 5 to 10. We evaluated confocal features of the ‘featureless’ lesions and of the ‘positive-borderline’ lesions. Evaluated confocal features were as follows: presence of pagetoid cells, cell shape (roundish or dendritic) and number (< 5 or >5 cells per mm2), overall architecture (ringed, meshwork, clods and non-specific pattern); architectural disorder, presence of cytological atypia (>5 cells per mm2) and cells arranged in nests. Results Among 322 lesions 70 were melanomas and 252 were nevi. According to the classification based on the 7point checklist score, 130 ‘featureless lesions’ (score 0–2) including six melanomas, and 102 ‘positive-borderline’ moles (score 3–4) including 17 melanomas, were identified. Round pagetoid cells >5 cells per mm2 and/or architectural disorder on RCM were found in all of six melanomas with featureless dermoscopy. Round pagetoid infiltration and five or more atypical cells at the DEJ were found in 16 positive ‘borderline melanomas’. Conclusions RCM represents a rapid non-invasive technique that can aid early diagnosis of dermoscopic difficult melanomas. Use of RCM on lesions with clinical and/or dermoscopic suspect of malignancy may reduce the number of unnecessary excision increasing the rate of accurate diagnoses. Received: 20 June 2014; Accepted: 9 September 2014
Conflicts of interest None declared.
Funding sources None declared.
Introduction The worldwide incidence of cutaneous malignant melanoma has been rising steadily over the past 30 years.1,2 As excision of thin melanoma guarantee a high rate of survival, compared with high risk of progression and a poor prognosis for advanced stages, early diagnosis still represents the most effective approach. The conventional clinical diagnosis of melanoma is based on some observable criteria that can be summarized into ABCDE
JEADV 2014
(asymmetry, irregular borders, multiple colours, diameter >6 mm, enlarging lesion).3 However, the ABCDE method only provides 64% accuracy, indicating a need for additional diagnostic tools.4 In the last decades, many efforts have been directed towards an early detection of melanoma and dermoscopy is worldwide employed as screening tool giving significant improvement of diagnostic accuracy.5 However, in some cases
© 2014 European Academy of Dermatology and Venereology
Ferrari et al.
2
melanoma diagnosis still remain difficult, due to lack of characteristic dermoscopic features, inducing a consistent rate of excision of benign lesions, corresponding to a number needed to excise (NNE) value that ranges from 8.7 to 29.4, according to the level of expertise and the setting.6 Reflectance confocal microscopy (RCM) is a second-level novel non-invasive technique that enables the in vivo visualization of the skin at quasi-histological resolution. In retrospective settings it showed to improve diagnostic accuracy when combined with dermoscopy.7,8 Moreover, prospective studies showed its capability to reduce the NNE both for highly suspicious lesions9 and for equivocal ones that may be referred to digital dermoscopic monitoring.10. As RCM features are strongly related with dermoscopic patterns,11 and the combination may help to increase the diagnostic accuracy, in this study we aimed to detect the most relevant RCM features for the detection of dermoscopic difficult melanomas. To increase objectivity, lesions were categorized according their dermoscopic score, as calculated by the 7-point checklist,12 and main RCM clues suggestive of melanoma diagnosis have been evaluated on a series of consecutive cases referred to RCM examination.
Materials and methods Study samples included all melanocytic lesions excised on the basis of equivocal clinical and/or dermoscopic features in 2010. Before excision, all lesions were recorded by means of digital dermoscopy and RCM at the Department of Dermatology, University of Modena and Reggio Emilia. Only lesions with high quality dermoscopic images, a complete set of confocal images and histopathology report available were included in the study. After dermoscopic evaluation of 7-point checklist score,12 single dermoscopic and RCM features were considered for lesions with a score of 4 or lower (corresponding to dermoscopic borderline or negative lesions).
histological diagnosis, by a dermatologist trained in dermoscopy and RCM (BF). Dermoscopic criteria
The 7-point checklist score was calculated for each case12 as well as the frequencies of each distinct dermoscopic feature accounting for the score. Afterwards, lesions were classified according the 7-point checklist score into three categories: ‘featureless’ lesions for score ranging between 0 and 2, ‘positive-borderline’ lesions for score between 3 and 4 and ‘positive- clear cut’ lesions for score from 5 to 10. Confocal criteria
Previously published RCM parameters demonstrated useful for melanoma detection have been selected. In the superficial layer it was evaluated the presence of pagetoid cells, the cell shape (roundish or dendritic) and their number (
