TyG Index Association with Melanoma Mortality: NHANES 1999–2018 and Hospital Data

1999–2018 data from 21,360 participants in the National Health and Nutrition Examination Survey (NHANES) and 475 patients from three tertiary hospitals, used to investigate the relationship between the triglyceride-glucose (TyG) index and melanoma prevalence and mortality. The study employed weighted logistic and Cox regression, restricted cubic splines, and seven machine learning algorithms, with a ridge regression model achieving an AUROC of 0.85. The dataset was authored by Yao Liang and last updated in March 2026.

Use Cases

• Train machine learning models for melanoma risk prediction based on metabolic markers like the TyG index.
• Analyze the U-shaped association between the TyG index and all-cause mortality in melanoma patients.
• Conduct subgroup analyses to verify the robustness of associations across different demographic factors.
• Interpret predictive models using SHAP analysis to identify key factors like race, age, and serum phosphorus levels.

Strengths

• Includes a large, nationally representative sample of 21,360 participants from NHANES.
• Model performance is quantified with a ridge regression AUROC of 0.85.
• Analysis includes an external validation cohort of 475 patients from three tertiary hospitals.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the primary dataset is unknown, which may limit suitability assessment.
• The dataset is very small at 12.6 KB, indicating it likely contains summary or processed results, not raw patient-level data.

Provenance

Source

National Health and Nutrition Examination Survey (NHANES) and three tertiary hospitals.

Collection Method

Cross-sectional survey data and hospital-based patient records.

Time Range

1999–2018

Freshness

Last updated 2026-03-18 05:35:02

Geography

United States (NHANES) and hospital locations (unspecified).