External validation in an Australian population of the EORTC-DeCOG nomogram predicting recurrence, distant metastasis and overall mortality in melanoma patients with positive sentinel lymph nodes

Abstract

BACKGROUND: Calculating an accurate prognosis for melanoma patients who have a positive sentinel node (SN) biopsy is important both for them and for their treating doctors to guide decision-making, particularly when adjuvant systemic therapy is being considered. The recently published EORTC-DeCOG nomograms aim to provide this via an online portal that predicts 5-year rates for recurrence, distant metastasis and overall mortality. The present study provides external validation of these nomograms. METHODS/MATERIALS: De-identified data from patients with a positive SN biopsy between 2003 and 2015 were extracted from the prospectively maintained Melanoma Institute Australia (MIA) research database. ROC-curves with C-statistics, regression co-efficients and Decision Curve Net Benefit analyses were performed using the integrated private validation portal on the nomograms' hosting platform (Evidencio). RESULTS: Complete data were available for 352 patients. The respective C-statistics for recurrence, distant metastasis and overall mortality nomogram validations were 0.68, 0.69 and 0.66. CONCLUSION: The performance of the nomograms in predicting recurrence and distant metastasis was similar in the MIA and the development populations, suggesting that they are robust. However, the overall mortality nomogram performance was significantly poorer in the MIA population (C-statistic 0.66) than in the original EORTC-DeCOG derivation cohort (C-statistic 0.70) and may therefore be less reliable for clinical use.