Risk factors and predictive models for early death in patients with advanced melanoma: A population-based study

Area of Science:

• Oncology
• Biostatistics
• Cancer Epidemiology

Background:

• Advanced melanoma (AM) carries a poor prognosis, with many patients diagnosed at late stages facing significant early mortality risk.
• Identifying independent risk factors is crucial for improving survival outcomes in AM patients.

Purpose of the Study:

• To identify independent risk factors for all-cause and cancer-specific early death in advanced melanoma patients.
• To develop and validate two nomogram models for predicting early mortality in advanced melanoma.

Main Methods:

• Utilized data from 2138 advanced melanoma patients (2010-2015) from the SEER database, split into training and validation cohorts.
• Employed logistic regression to identify risk factors and constructed nomograms. Model efficiency was assessed using ROC curves, calibration curves, and decision curve analysis.

Main Results:

• Significant independent risk factors identified include marital status, primary site, histological type, N stage, surgery, chemotherapy, and metastases (bone, liver, lung, brain).
• The developed nomograms demonstrated good predictive power (AUCs ranging from 0.740 to 0.759) and accuracy, validated across cohorts.
• Decision curve analysis confirmed the clinical utility and applicability of the nomogram models.

Conclusions:

• The validated nomograms provide a practical tool for predicting early death risk in advanced melanoma patients.
• These models can assist oncologists in timely intervention and the development of personalized treatment plans.
• Improved risk stratification through these nomograms may lead to better management and outcomes for advanced melanoma patients.