Imaging techniques for computer-assisted diagnosis of skin-pigmented lesions

Early diagnosis of cutaneous melanoma followed by prompt surgical excision of the lesion is, at present, the most reliable means to assure the best prognosis of survival to the patient. Recent studies have shown that only 50% clinical accuracy in diagnosis of early melanoma is usually achieved by visual inspection, whereas accuracy can increase up to 90% when diagnosis is performed by experienced clinicians. For this reason, there is increasing interest in the possibility of new technologies to be used to assist the clinician in performing diagnosis of melanoma. Since 1976 several attempts have been made to set up methods to quantitatively estimate the clinical features which are subjectively evaluated in making clinical diagnosis. In this paper some of these techniques are reviewed including photographic photometry, integrating sphere spectrophotometry, multispectral image spectrophotometry and color CCD- camera based techniques. The two latter image acquisition systems seem to provide similar sensitivity and specificity in discriminating melanoma from benign nevi. In an attempt to improve the diagnostic capability of a computer-assisted diagnosis of melanoma, we are studying the potential of the computer image analysis of the two different image acquisition systems. By combining the information derived from the image analysis of both methods, the reliability of a computer-assisted diagnosis of melanoma might be enhanced.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.