Demographic, Geographic, and Temporal Patterns of Melanoma Incidence in Texas, 2000-2018: a Bayesian Disease Mapping Analysis
Introduction
Melanoma is an aggressive and deadly form of skin cancer, with an estimated 97,610 new cases of cutaneous melanoma in the U.S. in 2023 (5th leading cancer diagnosis). Furthermore, melanoma is rising in incidence and is projected to be the second most common cancer by 2040. Diagnosing melanoma at its earliest stages significantly improves melanoma survival. We previously demonstrated that melanoma patients living in persistent poverty Texas counties experienced statistically significant higher rates of incidence-based melanoma mortality. This highlighted the importance of considering area-specific characteristics when implementing place-based interventions to facilitate early melanoma diagnosis and improve melanoma outcomes. However, melanoma is a relatively rare cancer which complicates developing reliable county-level estimates and projections. Disease mapping models, with more accurate estimated across space and time, can identify future high disease burden areas. Understanding where and when future high disease burden areas will occur can help researchers develop place-based interventions and allocate future prevention resources.
Methods
We used a retrospective study design to analyze 2000 – 2018 reported melanoma incidence from the Texas Cancer Registry and developed Bayesian spatial-temporal disease mapping models to provide annual county-level incidence-based melanoma rates. We first investigated the annual temporal trends of incidence-based melanoma rates (IMR) in Texas by age groups (18-29, 30-39, 40-49, 50-59, 60-69, 70-79, and >= 80), sex (Female and Male) and race/ethnicity (Non-Hispanic (NH) Whites, NH Blacks, Hispanics and others). Then, we calculated the county-level yearly expected cases using internal standardizations, accounting for the substantial differences in IMR by age, sex and race/ethnicity. With the expected cases, we developed the Bayesian inseparable spatial-temporal models to obtain estimated relative risk (RR) of incidence-based melanoma by county and year, adjusting for the difference in county population composition. Model-based RR provides more accurate and reliable estimates for understanding the geographic and temporal patterns of melanoma incidence in Texas.
Results
Females aged 18-39 have higher IMR compared to their male counterparts; however, the trend is reversed among those above 40 years old. NH White population in Texas still exhibit the highest IMR across all age groups, with most age groups showing a steady increasing trend as age increases, except for those aged 18-29 where IMR is decreasing. Panel plots of the county-level relative risk (RR) maps highlight the identified counties with the highest melanoma incidence.
Conclusion
In disease mapping analysis of rare diseases such as melanoma, a common challenge is to provide reliable estimates across space and time. Utilizing the Bayesian disease mapping approach, we can produce more accurate estimates to identify high disease burden areas and to inform place-based interventions and guide future resource allocations.