Ambient Ultraviolet Radiation and Sebaceous Carcinoma Incidence in the United States, 2000-2016

Abstract

Sebaceous carcinoma (SC) is an aggressive skin tumor. Although ultraviolet radiation (UVR) is an important risk factor for some skin cancer types, no population-level study has evaluated for an association between UVR and SC risk. Herein, we examined satellite-based ambient UVR in relation to SC incidence using Surveillance, Epidemiology, and End Results 18 cancer registry data (2000-2016). There were 3503 microscopically confirmed cases of SC diagnosed during the study period. For non-Hispanic whites, there was an association between increasing ambient UVR and SC risk (incidence rate ratio [per UVR quartile] = 1.15, 95% confidence interval = 1.11 to 1.19; two-sided P < .001) including among individuals with and without putative Muir-Torre syndrome. In contrast, there was no association between ambient UVR and SC risk for other race and ethnicities. Our findings support a role for UVR in SC tumorigenesis, which suggests that photoprotection may reduce SC risk, particularly for high-risk populations (eg, Muir-Torre syndrome).

Figure 1.

Ambient ultraviolet radiation (UVR) and incidence rate ratios for sebaceous carcinoma in select subgroups. Incidence rate ratios (IRR) with 95% confidence intervals (CIs) for microscopically confirmed cases of sebaceous carcinoma (International Classification of Diseases for Oncology-3 code 8410/3) diagnosed in Surveillance, Epidemiology, and End Results cancer registries (2000–2016) with increasing UVR. UVR data were cloud-adjusted daily ambient irradiance (wavelength = 305 nm). Analysis includes cases of sebaceous carcinoma from San Francisco–Oakland, Connecticut, Detroit (Metropolitan), Iowa, New Mexico, Seattle (Puget Sound), Utah, Atlanta (Metropolitan), San Jose-Monterey, Los Angeles, rural Georgia, California (excluding San Francisco, San Jose-Monterey, and Los Angeles), Kentucky, Louisiana, New Jersey, and greater Georgia. Cases from Hawaii and Alaska were excluded from the analysis because they were outliers for ambient UVR. Incidence for each UVR quartile was compared with UVR quartile 1 to calculate the IRR. Models are adjusted for sex, age (<50 , 50–64, 65–79, and 80 years and older), diagnosis period (2000–2005, 2006–2010, 2011–2016), and registry volume. Individuals were designated as having putative Muir-Torre syndrome, a phenotypic variant of Lynch syndrome (OMIM: 120435), if they had SC plus one of the following Lynch syndrome cancers: colon, rectum, stomach, liver, biliary tract, urinary bladder, renal pelvis, ureter, small intestine, pancreas, ovary, endometrial.