Loading ...

TOPLINE:

A newly developed risk-prediction tool for invasive melanoma demonstrated enhanced accuracy, identifying 74% of future cases by screening the top 40% of the predicted risk population.

METHODOLOGY:

• Researchers conducted a population-based prospective cohort study involving 41,919 participants without melanoma (mean age at baseline, 55.4 years; 55% women; 94% of European ancestry) in Queensland, Australia, from November 2010 to December 2011.
• The primary outcome was the first histologically confirmed invasive melanoma diagnosis over 10 years of follow-up.
• The analysis included 31 candidate predictor variables (27 phenotypic/clinical variables and four statistical interaction variables), identified from previously known risk factors and the literature. Predictors that appeared in at least 70% of the 50 imputed datasets were retained in the final risk-prediction model.
• Model discrimination and calibration were assessed using the concordance (C)-index and Brier scores. The model was internally validated using 1000 bootstrap samples to generate adjusted C-index scores. Researchers also calculated the best screening thresholds and compared performance with their seven‐factor tool.

TAKEAWAY:

• During 401,356 person-years of follow-up, 706 participants (1.7%) developed invasive melanoma, with a median time to diagnosis of 5.1 years.
• The final model retained 16 terms: 14 clinical predictors and two statistical terms (age squared and age-by-sex interaction). European ancestry, high nevus count, red hair color, and inability to tan were associated with the highest risks for invasive melanoma.
• The analysis determined that targeted screening of the 40% of participants with the highest risk scores would potentially identify 74% of all future invasive melanomas (number needed to screen, 32). Screening the top 50% of the participants would potentially identify 82% of future cases (number needed to screen, 36).
• Compared with the melanoma prediction tool that used seven predictors, the novel model identified 580 true invasive melanomas vs 545 cases down to the sixth decile and achieved better net reclassification at both the top 40% and top 50% risk thresholds.

IN PRACTICE:

“This cohort study has identified an improved tool that offers enhanced accuracy for predicting the future risk of invasive melanoma compared with existing tools,” the authors wrote. Referring to plans to validate the tool externally, they added, “in the meantime, we see strong merit in assessing the performance of this tool independently in other settings.”

SOURCE:

The study was led by David C. Whiteman, MBBS, PhD, QIMR Berghofer Medical Research Institute, Herston, Queensland, and was published online on September 10 in JAMA Dermatology.

LIMITATIONS:

A key limitation was the inability to validate the new prediction tool in an independent sample.

DISCLOSURES:

The study was supported by grants from the National Health and Medical Research Council (NHMRC) of Australia, Tour de Cure, and donations from The Great Priory of Queensland, Hand Hearts Pockets, and Brian and Merle Dwyer. Whiteman reported receiving grants from NHMRC Investigator Grant and serving as chair of the expert advisory committee for the Australian Roadmap for a National Targeted Skin Cancer Screening Program. The rest of the authors declared having no conflicts of interest.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.