Computer-assisted melanoma diagnosis: a new integrated system.

In dermatology, attempts at synergy between man and machine have mainly been made to improve melanoma diagnosis. The aim of the present study was to test an 'integrated digital dermoscopy analysis' (i-DDA) system with a series of melanocytic lesions that were benign and malignant in nature, and to evaluate its discriminating power with respect to histological diagnosis. In a retrospective study we used an i-DDA system to evaluate a series of 856 excised, clinically atypical pigmented skin lesions (584 benign and 272 malignant). The system evaluated 48 parameters to be studied as possible discriminant variables, grouped into four categories (geometries, colours, textures and islands of colour) integrated with three personal metadata items (sex, age and site of lesion) and presence/absence of three dermoscopic patterns (regression structures, blue-white veil and polymorphic vascular structures). Stepwise multivariate logistic regression of i-DDA data selected nine variables with the highest possible discriminant power. At the end of the stepwise procedure the percentage of cases correctly classified by i-DDA was 89.2% (100% sensitivity and 40.8% specificity). The limitations of the study included those associated with a retrospective design and the 'a priori' exclusion of nonmelanocytic skin lesions. By incorporating numerical digital features with personal data and some dermoscopic patterns into the learning process, the proposed i-DDA improved the performance of assisted melanoma diagnosis, with the advantage that our results can be objectively repeated in any other clinical setting.

Interventional three-year longitudinal study of melanocytic naevus development in pre-school children in Dresden, Saxony.

Acquired melanocytic naevi (MN) are considered a risk factor for melanoma. Exposure to ultraviolet light (UV) is the major environmental factor for MN. UV protection is most critical in pre-school children. This 3-year interventional longitudinal study examined 395 3-year-old children attending daycare centres (DCC) in Dresden, Germany. Photo-skin type, eye and hair colour were recorded. DCC were randomly assigned to a control group and a behavioural intervention group. All children had a regular naevus check-up, including digital objective analysis with Dell'Eva-Burroni Dermoscopy Melanoma Image Processing Software (DB-MIPS) technology. Parents of children in the intervention group received additional guidance for sun-protection. The mean total MN counts of both groups at the start of the study period were 7.19 ± 4.55 (intervention) and 6.84 ± 4.63 (control), respectively. There was a significant increase in MN counts for both groups (mean 12.5 and 13.8). Subgroup analysis for skin type, eye colour, and hair colour did not demonstrate a significant influence on MN counts. The DB-MIPS integrated classifier revealed no risky lesions while analysing their patterns. Intervention did not reduce the number of newly acquired MN. MN counts in pre-school children were approximately 5 times higher than expected from previous large studies in Germany. This is the first study in pre-school children using objective digital image analysis of pigmented lesions. No atypical lesions were observed. New approaches to UV protection in pre-school children are now required.

Impact of digital dermoscopy analysis on the decision to follow up or to excise a pigmented skin lesion: a multicentre study.

BACKGROUND: The quality of early malignant melanoma (MM) diagnosis is dependent on the experience of dermatologists, tools like dermoscopy and histopathology, and awareness and education of the studied population. Does a higher rate of excision of pigmented skin lesions (PSL) increase the rate of detected melanomas? MATERIAL AND METHODS: The DB-MIPS objective tool, able to evaluate mathematical defined variables, has been used to verify the variability of measurements among PSL stored by five different centres located in Italy, Switzerland, and Germany. RESULTS: The objective analysis showed low differences in terms of moles' features among the different groups, arguing for robustness of the dermatological patient's PSL inspection. Differences in terms of false positives and predictive positive values have been detected. The tendency to follow up a lesion was proportional to the percentage of thin MM (<0.75 mm tumour thickness), while the interventism was proportional to the percentage of dysplastic moles. Similar percentage of thin melanoma has been observed in all the centres, indicating a standardization in early diagnosing among experienced dermatologists. The main difference among the centres was their mode of action, i.e. to follow up or remove suspicious PSL. CONCLUSION: Interventism depends neither on the geographic site nor on the features of the observed moles. Higher removal rates do not correspond to higher MM detections: this means that an in-depth knowledge of melanoma patterns is required and follow-up of suspicious moles is highly suggested.

Objective melanoma progression.

BACKGROUND/PURPOSE: Many aspects of the natural history of malignant melanoma (MM) are still unclear, specifically its appearance at onset and particularly how it changes in time. The purpose of our study was to retrospectively determine objective changes in melanoma over a 3-24-month observation period. MATERIALS AND METHODS: Our study was carried out in two Italian dermatology centers. Digital dermoscopy analyzers (DB-Mips System) were used to retrospectively evaluate dermoscopic images of 59 MM (with no initial clinical aspects suggesting melanoma) under observation for 3-24 months. The analyzer evaluates 49 parameters grouped into four categories: geometries, colors, textures and islands of color. Multivariate analysis of variance for repeated measures was used to evaluate the statistical significance of the changes in the digital dermoscopy variables of melanomas. RESULTS: Within-lesion analysis indicated that melanomas increased in dimension (Area, Minimum, and Maximum Diameter), manifested greater disorganization of the internal components (Red, Green and Blue Multicomponent, Contrast, and Entropy) and increased in clusters of milky pink color (Light Red Area). CONCLUSION: Analysis of the parameters of our model and statistical analysis enabled us to interpret/identify the most significant factors of melanoma modification, providing quantitative insights into the natural history of this cutaneous malignancy.

Evaluation of cutaneous melanoma thickness by digital dermoscopy analysis: a retrospective study.

Digital dermoscopy analysis (DDA) exploits computerized analysis of digital images and offers the possibility of parametric analysis of morphological aspects of pigmented skin lesions by means of integration with dedicated software. We conducted a study by DDA in 141 melanomas, with the aim assessing whether the numerical variables extrapolated by univariate logistic analysis could be used in a system of multivariate analysis to predict melanoma thickness before surgery. Melanoma images were evaluated for 49 DDA parameters. Logistic analysis was conducted to identify statistically significant variables. The leave-one-out method was used to evaluate the predictive representations of rules for stepwise logistic classification. The percentage of correctly classified cases was calculated by a classification matrix. Melanomas less than 1 mm had a smaller area, faded borders and were more symmetrical than melanomas greater than 1 mm. The latter had a bluer colour and more random disposition of elements. The accuracy was 86.5%. Specifically, 97 of 108 thin melanomas (specificity 89.8%) and 25 of 33 thick melanomas (sensitivity 75.7%) were correctly classified. In conclusion, the predictive value of DDA for melanoma thickness was quite good. Moreover, DDA allowed us to know objectively those dermoscopic features important in the differentiation between thick and thin melanoma. However, further studies should be performed in a prospective setting before the clinical application.

Objective follow-up of atypical melanocytic skin lesions: a retrospective study.

Various authors have suggested that information from longitudinal observation (follow-up) of dynamic changes in atypical melanocytic pigmented skin lesions (MPSL) could enable identification of early malignant melanoma escaping initial observation due to an absence of specific clinical and dermoscopic features. The aim of our retrospective study was to determine the existence of numerical variables regarding changes in MPSL that could be useful to differentiate early melanomas and atypical nevi. The study was carried out in two Italian dermatology Centres. Digital dermoscopy analyzers (DB-Mips System) were used to evaluate dermoscopic images of 94 equivocal pigmented skin lesions under observation for 6-12 months and then excised because of changes across time (29 melanomas and 65 nevi). The analyzer evaluates 49 parameters grouped into four categories: geometries, colours, textures and islands of colour. The ROC curve designed on the 49 digital dermoscopy analysis parameters showed good accuracy. At sensitivity (SE) = specificity (SP), it correctly classified 89.3% of cases. When objective pigmented skin lesion parameters were considered together with their objective changes over 6-12 months, a decisive increase in discrimination capacity was obtained. At SE = SP accuracy was 96.3%. Analysis of the parameters of our model and statistical analysis enabled us to interpret/identify the most significant factors of modification and differentiation of lesions, providing quantitative insights into the diagnosis of equivocal MPSL and demonstrating the utility of objective/numerical follow-up.

Digital dermoscopy in clinical practise: a three-centre analysis.

BACKGROUND: Early detection of cutaneous melanoma is the most important step to improve prognosis. In recent years objective techniques have been developed to improve diagnostics of pigmented lesions superior to the naked eye. Digital dermoscopy offers several advantages such as independence from the investigator, format ready for teledermatology and image storage, and comparability. SETTING: In a routine setting including three centres in Europe, i.e. Lugano and Locarno in Switzerland, and Dresden in Germany, the usefulness of digital dermoscopy in daily routine has been evaluated. PATIENTS AND METHODS: Digital dermoscopy was performed with the DB-MIPS system in 1308 patients, and 3544 pigmented lesions were evaluated. A total of 466 pigmented lesions were surgically removed and classified by histopathology. Among those 52 melanomas and 299 dysplastic, naevi could be identified. DB-MIPS system evaluated 35 variables of geometrics, colour, colour distribution and texture. Lesion identification is realized by clustering. Another important part of the analysis was to study the decision making, i.e. removal or follow-up of lesions. The tendency to follow-up was defined as the ratio of (number of suspicious lesions-removed lesions)/suspicious lesions. The interventism is defined as the ratio of the number of removed lesions to the number of suspicious lesions. RESULTS: The DB-MIPS system was robust and easy in handling in daily routine. The data analysis and digital imaging were realized in real time. Fifty-two melanomas could be identified including eight in situ melanomas and 33 T1a melanomas (thinner than 1 mm). Sensitivity of digital dermoscopy varied between 90% and 95% between the three centres. Specificity varied between 79.6% and 93.3%. The centres differed in the patient population, the mean diameter of pigmented lesions and decisions. Interventism varied between 36.3% and 76.0%, follow-up varied from 24% to 63%. Higher levels of clinician's false positive were associated with a higher percentage of interventism. CONCLUSIONS: Digital dermoscopy offers advantages for daily routine in detection of early melanoma. Sensitivity and specificity for early melanomas is high and thereby, the experienced dermatologist can be easily supported in daily routine of a pigment lesion clinic to improve diagnostics and hopefully prognosis in cutaneous melanoma.

Melanocytic skin lesions and pregnancy: digital dermoscopy analysis.

BACKGROUND: Very few studies have tried to clarify how pregnancy influences the morphology of pigmented skin lesions (PSL). Our purpose was to objectively determine, by digital dermoscopy analysis (DDA), any dermoscopic changes of acquired melanocitic nevi during pregnancy and after 1 year from delivery. METHODS: Thirty-five healthy pregnant women and 35 age-matched female controls were enrolled in the study. Nevi of pregnant women were analysed by DDA between 5 and 8 weeks of pregnancy, between 39 and 41 weeks of pregnancy and 12 months after delivery. Nevi of control women were analysed by DDA in a month of the year matching the period of recruitment of pregnant women and 21 months later. RESULTS: Multivariate analysis of variance (manova) for repeated measures revealed that dermoscopic variables SKIN-GREEN-AVERAGE, SKIN-BLUE-AVERAGE and CONTRAST changed during pregnancy but returned to non-significant values after a year from delivery. The variable ENTROPY showed significant differences between initial evaluation and 1 year after delivery. Finally, the variable VARIANCE OF BORDER GRADIENT showed a significant difference between the first and the last evaluations, in both pregnant and control subjects. CONCLUSIONS: The study showed that pregnancy leads to significant modifications in PSL, especially with regard to pigment network, globules and architectural order or disorder.

Comments on "A new algorithm for border description of polarized light surface microscopic images of pigmented skin lesions".

In this paper, discrepancies and reference inaccuracies in the paper (Grana et al., 2003) are pointed out. Specifically, it is demonstrated that the definitions of "lesion gradient" and "skin lesion gradient," widely used in a number of medical papers on computer analysis of pigmented skin lesions, are unambiguous, and that the "new algorithm for border description" described in the subject paper substantially relies on well-established concepts dating back over one decade ago.

Predominant dermoscopic patterns observed among nevi.

BACKGROUND: It has been clinically observed that patients' "normal" moles resemble each other. Whether this concept is applicable to dermoscopic practice has not been sufficiently studied. OBJECTIVE: To investigate whether physicians evaluating dermoscopic images would identify common dermoscopic profiles of nevi within individual patients. METHODS: Images of 205 nevi belonging to 18 patients were evaluated by 2 dermatologists for dermoscopic global pattern, color, and specific structures. We defined dermoscopic patterns as dominant if seen in >or= 40% of the patient's nevi; a minor pattern was defined as 20 to 39%. RESULTS: A dominant pattern was seen in 15 patients (83%). In 13 (72%) of the patients, >or= 80% of the nevi were classified into one, two, or three global patterns. The reticular global pattern was the most prevalent dominant pattern, seen in 9 patients (50%); the homogeneous pattern was the most prevalent minor pattern, seen in 16 patients (89%). CONCLUSION: Individuals tend to have one to three predominant dermoscopic nevus global patterns.

Digital surface microscopy analysis of conjunctival pigmented lesions: a preliminary study.

The objective of this study was to investigate whether digital surface microscopy (DSM) could be used for the follow-up and comparison of malignant and benign conjunctival pigmented lesions (CPLs). Thirty-nine CPLs [16 de novo malignant melanomas (MMs), one MM arising from primary acquired melanosis (PAM), six PAMs and 16 naevi] were digitally analysed and biopsied. All of the PAMs and 10 naevi, which had not been surgically excised, were followed up using DSM. Thirty parameters were evaluated grouped into four categories: geometry, colour, texture and islands of colour. None of the CPLs that were followed up, which comprised 10 naevocytic naevi and seven PAMs, showed any morphological change at DSM analysis, except for one PAM which developed an MM 1 year later. Of the geometric variables examined, the area, maximum diameter and minimum diameter showed significantly higher values in MMs compared with benign CPLs. With regard to the colour of CPLs, MMs were significantly darker and bluer than naevi. In the texture group, contrast was significantly higher in MMs. In the islands-of-colour group, the imbalance of blue-grey regions and the presence of dark areas were significantly higher in MMs. DSM greatly simplified the follow-up of CPLs, such as PAMs with atypia, by providing satisfactory quality images with high reproducibility; this technique is also easy to use and well accepted by patients. Moreover, this preliminary study allowed us to determine which objective variables could be important for distinguishing between benign CPLs and conjunctival MMs.

Inter- and intra-variability of pigmented skin lesions: could the ABCD rule be influenced by host characteristics?

BACKGROUND/PURPOSE: Many differences in color, shape and dimension exist between different moles even in the same individual. Major differences might be accounted for anatomical location, genetic factors and by environmental factors, mainly sunlight exposure. Therefore, it would be of great value, when evaluating skin lesions, to take into account the degree of intra- and inter-variability of several diagnostic parameters. In order to assess the morphologic and chromatic differences between lesions belonging to different patients and between lesions belonging to the same individual, we examined objective digital parameters obtained with dermatoscopic analysis, using the DBDermo MIPS system (BIO MIPS Engineering, S.R.L, siena, Italy). METHODS: The automatic classifier inside the software is based on a 'match by similarity' algorithm, based on the measurement of the Euclidean distances of all variables considered from the reference image. Two-hundred and four clinically benign pigmented lesion, belonging to 18 patients were examined, stored and automatically processed. For each lesion objective parameters related to geometry, color and texture were automatically evaluated. RESULTS: We found skin color (healthy skin) is objectively different from subject to subject and the lesion color is more similar among different lesions of the same patient than among lesions belonging to different individuals both in their darkest and slightly dark component. We also observed that lesion dimensions are individual correlates, i.e. the probability for a lesion to be large is higher when the other, in the same patient, is large. CONCLUSION: Many parameters of pigmented skin lesions evaluated by digital dermoscopy analysis are similar in the same patient and different from those belonging to different individuals. This indicates that, when considering a lesion, we should take into account the peculiar patient's characteristics.

Melanoma computer-aided diagnosis: reliability and feasibility study.

BACKGROUND: Differential diagnosis of melanoma from melanocytic nevi is often not straightforward. Thus, a growing interest has developed in the last decade in the automated analysis of digitized images obtained by epiluminescence microscopy techniques to assist clinicians in differentiating early melanoma from benign skin lesions. PURPOSE: The aim of this study was to evaluate diagnostic accuracy provided by different statistical classifiers on a large set of pigmented skin lesions grabbed by four digital analyzers located in two different dermatological units. EXPERIMENTAL DESIGN: Images of 391 melanomas and 449 melanocytic nevi were included in the study. A linear classifier was built by using the method of receiver operating characteristic curves to identify a threshold value for a fixed sensitivity of 95%. A K-nearest-neighbor classifier, a nonparametric method of pattern recognition, was constructed using all available image features and trained for a sensitivity of 98% on a large exemplar set of lesions. RESULTS: On independent test sets of lesions, the linear classifier and the K-nearest-neighbor classifier produced a mean sensitivity of 95% and 98% and a mean specificity of 78% and of 79%, respectively. CONCLUSIONS: In conclusion, our study suggests that computer-aided differentiation of melanoma from benign pigmented lesions obtained with DB-Mips is feasible and, above all, reliable. In fact, the same instrumentations used in different units provided similar diagnostic accuracy. Whether this would improve early diagnosis of melanoma and/or reducing unnecessary surgery needs to be demonstrated by a randomized clinical trial.

Automated diagnosis of pigmented skin lesions.

Since advanced melanoma remains practically incurable, early detection is an important step toward a reduction in mortality. High expectations are entertained for a technique known as dermoscopy or epiluminescence light microscopy; however, evaluation of pigmented skin lesions by this method is often extremely complex and subjective. To obviate the problem of qualitative interpretation, methods based on mathematical analysis of pigmented skin lesions, such as digital dermoscopy analysis, have been developed. In the present study, we used a digital dermoscopy analyzer (DBDermo-Mips system) to evaluate a series of 588 excised, clinically atypical, flat pigmented skin lesions (371 benign, 217 malignant). The analyzer evaluated 48 parameters grouped into 4 categories (geometries, colors, textures and islands of color), which were used to train an artificial neural network. To evaluate the diagnostic performance of the neural network and to check it during the training process, we used the error area over the receiver operating characteristic curve. The discriminating power of the digital dermoscopy analyzer plus artificial neural network was compared with histologic diagnosis. A feature selection procedure indicated that as few as 13 of the variables were sufficient to discriminate the 2 groups of lesions, and this also ensured high generalization power. The artificial neural network designed with these variables enabled a diagnostic accuracy of about 94%. In conclusion, the good diagnostic performance and high speed in reading and analyzing lesions (real time) of our method constitute an important step in the direction of automated diagnosis of pigmented skin lesions.

Digital dermoscopy analysis and artificial neural network for the differentiation of clinically atypical pigmented skin lesions: a retrospective study.

Noninvasive diagnostic methods such as dermoscopy or epiluminescence light microscopy have been developed in an attempt to improve diagnostic accuracy of pigmented skin lesions. The evaluation of the many morphologic characteristics of pigmented skin lesions observable by epiluminescence light microscopy, however, is often extremely complex and subjective. With the aim of obviating these problems of qualitative interpretation, methods based on mathematical analysis of pigmented skin lesions have recently been designed. These methods are based on computerized analysis of digital images obtained by epiluminescence light microscopy. In this study we used a digital dermoscopy analyzer with 147 clinically atypical pigmented skin lesions (90 nevi and 57 melanomas) to determine its discriminating power with respect to histologic diagnosis. The system evaluated 48 objective parameters used to train an artificial neural network. Using the artificial neural network with 10 variables selected by a stepwise procedure, we obtained a maximum accuracy in distinguishing melanoma from benign lesions of about 93%. Comparing this result with those of the many studies using classical epiluminescence light microscopy, it emerges that the method proposed is equal or even superior in diagnostic accuracy and has the advantage of not depending on the expertise of the clinician who examines the lesion.