Validity of melanoma diagnosis in a community-based screening program.

Although screening for melanoma is intuitively attractive, evidence of the effectiveness of screening programs for skin cancer is lacking. Since 1990, the Lions Cancer Institute has conducted clinics in Western Australia in which volunteer plastic surgeons and dermatologists undertake full-body skin screens. Advertisements for attendees target people with risk factors for skin cancer. Each person screened between 1994 and 2002 (n = 7,436) completed a questionnaire including basic demographic information, on which the physician added provisional diagnoses. Attendees' details were linked with the Western Australian Cancer Registry to determine the number of diagnosed melanomas up to 1 and 2 years after screening. The positive predictive value of a screening diagnosis of "any lesion" at a particular body site was 1.5% and that of a screening diagnosis of "melanoma" was 10.0%. The 1-year specificity of the screening test ranged from 95.1% to 99.5%, and 1-year sensitivity ranged from 63.6% to 81.8%. Two-year sensitivity was lower. If body site was not taken into account, the sensitivities were higher and the specificities lower. Findings suggest that the validity of skin screening diagnoses in the general population is reasonable. Body site of the lesion should be taken into account when calculating validity of these diagnoses.