Spectral regions contributing to melanoma: a personal view.

Although human cutaneous melanoma is a complicated disease, the principal etiologic agent for its incidence in fair skin individuals is exposure to sunlight. In order to understand the epidemiology of melanoma - temporal effects, latitude effects, sunscreen effects, albino susceptibility, and differences from nonmelanoma skin cancer -one must approach the problem by obtaining clues indicating which wavelengths in sunlight are effective in inducing melanomas. One way is to use an animal model. At present, the only suitable model is a backcross hybrid of small tropical fish of the genus Xiphophorus, bred to have only one tumor suppressor gene. Single UV exposures to 7-d-old fish induce melanomas readily observable by 4 mo. The initial slopes of dose-response curves for exposures at 302, 313, 365, 405, 436, and 547 nm yield sensitivity as a function of wavelength. This action spectrum does not look like the spectrum for light absorption by DNA (mostly in the UVB), but has appreciable sensitivities in the UVA and visible regions, and looks like a direct effect of light on DNA plus a large indirect effect on DNA by absorption of light by the intracellular melanin. Because the UVB is only a fraction of solar irradiance, one may calculate that 90% of melanoma induction in humans arises from UVA and visible, assuming the human spectrum is similar to the fish spectrum. The implications of this calculation are that (i) depletion of stratospheric ozone will not affect melanoma incidence, (ii) an increase in sun exposure time as a result of using UVB sunscreens could increase the risk of melanoma, and (iii) the use of high UVA sun tanning devices could increase the risk of melanoma.