aEarly stage: SEER stages 0 (in situ), 1 (localized), and 2 (regional by direct extension only).

bLate stage: SEER stages 3 to 5 (regional) and stage 7 (distant).

cUnknown stage: SEER stage 9: unknown, unstaged, unspecified.

IRs for males appear much higher than females among 50+ years, reaching over 140 per 100,000 in the 70+ age group and over 180 per 100,000 in the 80+ age group by 2018 (). In contrast, among younger age groups (18-49), females showed slightly higher rates than males over time. In addition, melanoma incidence increased at a more rapid rate for males from 2009 to 2018, with an AAPC of 3.48% (95% CI 2.49-5.19) compared with 2.47% (95% CI 1.89-3.26) for females ().

Stratifying by race and ethnicity, non-Hispanic White individuals displayed the highest rates across all age groups, with noticeable increases over the study period for those aged 50 years and older. The older non-Hispanic White patients (70+ years) showed rates over 130 per 100,000 by 2018. For the 18‐29 age group, non-Hispanic White patients had a slight decrease in IRs from 16.5 per 100,000 in 2000 to 11.4 per 100,000 in 2018. Other racial and ethnic groups, such as non-Hispanic Black and Hispanic individuals, maintained much lower IRs mostly below 30 per 100,000 across all age groups and years ().

To facilitate comparison with national statistics, age-adjusted IRs (to the 2000 US standard population) are presented by stage at diagnosis, age, gender, and race and ethnicity in .

The incidence of early-stage melanoma has increased more rapidly than late-stage melanoma incidence (). Early-stage melanoma temporal trends mirrored those of overall cases, with the highest rates for non-Hispanic White males older than 70 years, reaching 114 cases per 100,000 by 2018. Other racial and ethnic groups displayed minimal variation over time, with rates remaining low throughout the study period (under 10 cases per 100,000 for non-Hispanic Black individuals and under 25 per 100,000 for Hispanic individuals). Late-stage melanoma IRs were lower than early-stage incidence across all demographics. The overall population showed rates below 15 cases per 100,000, with a slight increase over time in older males, peaking around 25 per 100,000 between 2011‐2014 (refer to additional details in Discussion). Non-Hispanic White patients had the highest rates (maximum 16 per 100,000 in the 80+ age group), while non-Hispanic Black and Hispanic patients had consistently low late-stage IRs (mostly below 5 cases per 100,000). The IRs for unknown stage cases remained relatively stable over time across all demographic groups.

When investigating stage by race and ethnicity, early-stage diagnoses predominated across all races and ethnicities. However, Hispanic and non-Hispanic Black individuals had proportionately more late stage at diagnosis melanomas than non-Hispanic White patients (). The AAPC for late-stage melanoma cases was similar for Hispanic (AAPC 2.38%, 95% CI 0.53-3.73) and non-Hispanic White patients (AAPC 2.66%, 95% CI –0.98 to 7.17) and higher for non-Hispanic Black patients (AAPC 5.79%, 95% CI 2.61-9.01; ).

Maps showing high- and low-melanoma incidence spatial clusters are presented in . When considering all melanoma cases, spatial clusters with significantly higher-than-average melanoma incidence (median IR 42 per 100,000, IQR 23-62) were primarily in northwestern Texas from 2000 to 2006, with a shift to central Texas between 2007 and 2015, and then to counties surrounding Dallas by 2018. Spatial clusters with lower-than-average melanoma incidence were clustered near southern and western Texas (median IR 11, IQR 0-24). The spatial patterns for early-stage and overall melanoma cases were similar with median IRs of 35 per 100,000 (IQR 19-52) in high-incidence spatial clusters, and 6 per 100,000 (IQR 0-17) in low-incidence spatial clusters. However, the spatial clusters for higher-than-average late-stage cases showed distinct patterns with localization to northwestern Texas from 2000 to 2007, a shift toward southeast Texas between 2010 and 2014, and a return to central-northern Texas in 2015.

Overlaying the 2018 spatial clusters with rural counties (hatched lines), we observed that clusters of higher-than-average late-stage melanoma incidence significantly overlap with rural areas (P=.05; ). Similarly, clusters of higher-than-average early-stage melanoma incidence appear to overlap with urban areas (P=.06; ). In contrast, clusters of lower-than-average incidence for overall, early-stage, and late-stage melanoma appear to overlap with persistent poverty counties (all P values were <.001; ).

Texas has the highest proportion of late-stage melanoma cases relative to the total number of reported melanoma cases in the contiguous United States and is the second most populated state. Identifying the regions and patient populations that disproportionately bear the burden of late-stage melanoma at diagnosis is crucial for developing targeted early detection efforts. Our spatial clustering analysis of Texas revealed that high-incidence clusters of early-stage melanoma were primarily localized in urban, well-resourced areas, whereas high-incidence clusters of late-stage melanoma were concentrated in rural areas . This disparity may be partly explained by the lower density of dermatologists in rural areas [,]. Patients with melanoma living in these areas may experience delayed diagnosis [], need to travel longer distances to receive surgical management, and have decreased melanoma-specific survival []. These findings, while observational, highlight the importance of addressing structural disparities in health care access. Therefore, any early detection intervention must be tailored to be feasible in rural, lower-resourced settings.

Melanoma IRs exhibit substantial variation by age, with older patients experiencing significantly higher rates compared with younger patients. Because the age distribution of the population often shifts over time, age-adjusted rates are commonly used to compare trends across different time periods and geographic regions. These adjustments typically use a fixed reference year for the age distribution, which may lack the robustness to fully capture ongoing demographic change. As the population continue to age and shift in age structure, a static reference year may obscure important trends and fail to accurately reflect the current risk landscape. In this study, we used a novel visualization approach using temporal heat maps that directly incorporate age as a variable, providing a more adaptive and precise method for identifying shifts in melanoma incidence trends over time. This visualization highlighted pronounced incidence variations, particularly among late-stage male and non-Hispanic White patients from 2010 to 2014. We identified two TCR data sources which may explain this variation: Texas Health Care Information Collection (THCIC) and eMaRC Plus (Centers for Disease Control and Prevention [CDC]). THCIC, established by the Texas legislature, collects data on health care activities in hospitals and health maintenance organizations []. During a pilot from 2010 to 2013, THCIC identified melanoma cases not otherwise reported to TCR, leading to the inclusion of 559 cases in the TCR database, most of which were categorized as unknown stage at diagnosis. The second data source, eMaRC Plus, a software developed by the CDC to receive and process Health Level Seven files from pathology laboratories, was used from 2010 to 2018. eMaRC Plus identified 8,786 cases, with 93.1% being early-stage and 6.8% unknown stage. The peaks in early-stage and unknown-stage incidence observed in 2011 and onward may be attributed to these two data sources. However, the increase in late-stage incidence from 2010 to 2014 could not be fully explained by these data sources, suggesting that temporary reporting inconsistencies may warrant further investigation to fully understand their impact on the identified temporal trends.

Despite these data source differences, descriptive trends also revealed that older non-Hispanic White men comprise the majority of late-stage melanoma cases at diagnosis. This specific demographic provides a clear target cohort for refining early melanoma detection efforts. Given that older men generally have lower rates of skin self-awareness [], promoting early detection through their family members or PCPs may offer greater opportunity. Our analysis of race and ethnicity revealed that, while most melanoma cases were diagnosed at early stages across all groups, Hispanic and non-Hispanic Black patients experienced proportionately more late stage at diagnosis melanomas than non-Hispanic White counterparts. Although the absolute number of melanoma cases is small compared with other cancers, these findings reinforce the concept that melanoma can impact individuals of all races and ethnicities.

This study has several strengths and limitations. A key strength is the innovative visualization approach, which enabled a comprehensive analysis of temporal trends, stratified by age, sex, race and ethnicity, and stage at diagnosis. The use of TCR data allowed for the inclusion of patient residential information at the time of diagnosis, as well as additional data captures from pilot studies that would otherwise be unavailable. Furthermore, the geospatial analysis provided a comprehensive examination of patterns across all cases and by specific stages, offering valuable insights that could guide future interventions and educational efforts. The study also presents several limitations, particularly the large proportion of cases with an unknown stage at diagnosis. More precise data on these cases would enable better stage classification and improve the stage-specific analyses. In addition, the TCR data has limitations in capturing SDoH, which, if included, could provide a deeper understanding of the health disparities associated with melanoma outcomes.

Our study provides valuable guidance for future early melanoma detection efforts. Such efforts must be feasible in rural, lower-resourced areas of the state and focus on patients at highest risk of late-stage melanoma at diagnosis. Multimodal approaches, which combine foundational dermatology training for interested PCPs [], telementoring to support PCP incorporation of skin cancer detection examinations into practice [], and efficient store-and-forward eConsults [] to reduce dermatology access gaps, offer promising pathways to improve early melanoma detection in low-resource settings.

This project is funded by CPRIT RP230036 (MPI: CB and KCN).

The data used in this study were obtained from the Texas Cancer Registry via a formal data request, accompanied by an institutional review board application. For inquiries regarding data access, please contact the Texas Cancer Registry at CancerData@dshs.texas.gov.

CB conceived and designed the analysis. KCN contributed to the acquisition of the data. KZ and CB contributed to data processing and curation and conducted the formal data analysis. KZ and MMT wrote the initial draft of the manuscript. CB and KCN supervised the study and acquired funding.

None declared.

Edited by Naomi Cahill; submitted 23.10.24; peer-reviewed by Paul J Gruenewald, Sean Young; final revised version received 26.02.25; accepted 26.02.25; published 02.05.25.

© Kehe Zhang, Madison M Taylor, Jocelyn Hunyadi, Hung Q Doan, Adewole S Adamson, Paige Miller, Kelly C Nelson, Cici Bauer. Originally published in JMIR Cancer (https://cancer.jmir.org), 2.5.2025.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Cancer, is properly cited. The complete bibliographic information, a link to the original publication on https://cancer.jmir.org/, as well as this copyright and license information must be included.