preview

Computer Aided Dermatology

Decent Essays
Open Document

The different methods that fall under computer-aided diagnosis are more variable in diagnosing different attributes of melanoma as opposed to physicians. In computer diagnosis, patients are able to undergo several, but not all types of noninvasive diagnoses such as teledermatology, mole mapping and monitoring, automated diagnosis of melanoma, computer vision for automatic diagnosis of melanoma, computer vision for noninvasive measurement of the depth of melanomas, and confocal microscopy of histopathology. (Elbaum, 2002) With the variability in types of diagnoses, individuals have the ability to choose what they want view and analyze in regards to their skin, whereas physicians are limited by the methods they utilize (ABCDE, “Ugly Duckling,” …show more content…

A popular technique utilized by dermatologists to diagnose melanoma includes dermoscopy, which is a “noninvasive technique for taking high – resolution images of skin lesions by making the superficial layers of the skin translucent.” (Dreiseitl, Binder, Hable, & Kittler, 2009) Dermoscopy is a technique that allows dermatologists to view lesions and moles at a larger magnitude yielding to a more accurate diagnosis through the ABCDE method and Glasgow 7 – point checklist. While dermoscopy has aided physicians in accurately diagnosing melanoma, computer – aided dermoscopy has been proven to have higher accuracy in diagnosing patients. In one study, expert physicians scored a sensitivity of 81% and specificity of 95%, whereas the computer aided dermoscopy scored 93% on sensitivity and 95% specificity. (Barzegari, Ghaninezhad, Mansoori, Taheri, Naraghi, & Asgari, 2005) The high sensitivity and specificity of the computer diagnosis is the result of the lack of bias computers have in comparison to physicians and dermatologists. In computer-aided dermoscopy, the computer compares the patient’s lesions and moles to a database of previously recorded and imputed parameters. Thus, computer diagnosis is accurate to the extent of the set parameters and is reproducible in results as opposed to trained physicians who have variable ranges for specificity and sensitivity due to bias. Similarly, in the dermoscopy of small lesions, expert dermatologists scored an average of 71% sensitivity and 49% specificity whereas the computer vision system achieved 98% sensitivity and 44% specificity. (Friedman, Gutkowicz-Krusin, Farber, Warycha, Schneider-Kels, Papastathis, Mihm, King, Prieto, Kopf, Polsky, Rabinovitz, Oliviero, Cognetta, Rigel, Marghoob, Rivers, Johr, Grant-Kels, & Tsao, 2008) Especially in smaller lesions, the computer vision system had a relatively

Get Access