Geospatial analysis of the distribution of uveal melanoma patients at a referral center in the United States.

Abstract

e21566 Background: Approximately 2500 new uveal melanoma (UM) cases (5.2 cases per million population) are diagnosed yearly in the United States. We previously identified three unique geospatial accumulations of UM in North Carolina, Alabama, and New York based on our clinical experience. Such unexpected findings prompted us to conduct a more systematic evaluation of geospatial accumulations of UM patients at our institution. Methods: Through the electronic medical record system, we identified patients seen at our institution with a diagnosis of UM. Identified living patients and family of deceased patients were either sent an e-mail or paper letter with an internet link to a survey. The survey asked for the patient’s name, date of birth, sex, date of primary UM diagnosis, address at time of UM diagnosis and number of years at that address, and all previous addresses and dates of residence prior to the UM diagnosis. Using these addresses, ArcGIS and R software was used to analyze the geospatial distribution of our UM patients. The data was supplemented with 2020 American Community Survey census estimates at the state, county, and census tract level to estimate population-based rates of UM. Results: We identified 2718 UM patients in our clinical database, and a total of 726 participants completed the survey with a total of 2939 US addresses. The majority of addresses were precise to the street address level (77.9%). At the time of UM diagnosis, patients had lived at their current address for a median of 16 years. Pennsylvania and the neighboring states expectedly made up the majority (63.6%) of our patients, but about 1/3 of our patients resided in other states at the time of their UM diagnosis (total of 42 different states). Several counties presented elevated rates of UM with ≥5 case per 100,000 persons. We observed a group of counties in central Pennsylvania, at the border of Pennsylvania and New York, and in upper New York by the Finger Lakes with notably elevated rates of UM, which was consistent with our clinical observations at our institution. Several neighboring counties in Kentucky and Virginia also presented elevated rates. Conclusions: Through this pilot project, we have demonstrated that it is possible to collect lifetime residency information from UM patients in order to map their geospatial distribution in the US. Furthermore, although institutional bias toward referring physicians should be considered, our geospatial analysis identified accumulations of UM in areas similar to our clinical concerns. The results of our study warrant further investigations into environmental influences on UM development. Our future goal is to extend our approach through national and international collaborations to increase the number of participants in the analysis and to minimize institutional bias.