| Title | Value of the clinical history for different users of dermoscopy compared with results of digital image analysis. | | Author(s) | Blum A, Hofmann-Wellenhof R, Luedtke H, Ellwanger U, Steins A, Roehm S, Garbe C, Soyer HP | | Institution | Department of Dermatology, University of Tuebingen, Liebermeisterstrasse 25, 72076 Tuebingen, Germany. ablum@derma.de | | Source | J Eur Acad Dermatol Venereol 2004 Nov; 18(6):665-9. | | MeSH | Adolescent
Adult
Aged
Aged, 80 and over
Algorithms
Chi-Square Distribution
Child
Child, Preschool
Clinical Competence
Comparative Study
Dermatology
Diagnosis, Computer-Assisted
Diagnosis, Differential
Female
Humans
Infant
Male
Microscopy
Middle Aged
Nevus, Pigmented
Reproducibility of Results
Retrospective Studies
Sensitivity and Specificity
Video Recording
| | Abstract | BACKGROUND: The clinical history of a given pigmented lesion could influence the therapeutic decision. Teledermatology and automated image analysis also hold great potential for revolutionizing dermatology services.
AIM: The aim of this retrospective study was to evaluate the diagnostic accuracy of users with different experiences in dermoscopy with and without information about patients and their history compared with classification by an automated analysing system.
SETTING: One hundred and fifty-seven dermoscopic images of pigmented lesions, taken and proved by histopathology at the Pigmented Lesions Clinic of the Department of Dermatology of the University Tuebingen, Germany, were included.
METHODS: All images were viewed by three investigators with different experience: excellent (A), average (B) and beginner (C). In the first dermoscopic classification, no information was available. After 3 months the same images were once more classified by the three investigators, now with the information about the patients and their history. The melanocytic lesions were tested by the Tuebinger Mole Analyser.
RESULTS: For user A the sensitivity, specificity and diagnostic accuracy revealed no improvement on including the history (81.3% to 84.4%, 94.6% to 92.3% and 92.0% to 90.7%), whereas user B clearly improved his results (75.0% to 87.5%, 76.9% to 88.5% and 76.5% to 88.3%). No change in the sensitivity was seen by user C (84.4%), but there was a clear improvement in the specificity (69.2% to 87.7%) and diagnostic accuracy (72.2% to 87.0%). Using the computer algorithm, a sensitivity of 100%, a specificity of 76.9% and a diagnostic accuracy of 81.9% were achieved.
CONCLUSIONS: The study revealed results relevant to the use of dermoscopy: (1) continuing dermoscopic education influences the diagnostic accuracy; (2) the history is helpful for averaged users and beginners in dermoscopy; (3) digital image analysis has the highest sensitivity, but a lower specificity compared to the clinicians; and (4) digital dermoscopy could be used for store-and-forward systems in teledermoscopy. | | Language | eng | | Pub Type(s) | Journal Article
| | PubMed ID | 15482291 |
|
