A systematic review and meta-analysis identified 21 risk prediction models for sentinel lymph node biopsy positivity in melanoma,
In the study, published in JAMA Dermatology, investigators reported a pooled weighted C statistic of 0.78 (95% confidence interval [CI] = 0.74–0.81). However, significant heterogeneity was observed among the models (I² = 97.4%), indicating variability in performance across different patient populations.
The investigators analyzed 23 articles developing or validating risk models, including 20 external validations of eight models. The Memorial Sloan Kettering Cancer Center and Melanoma Institute of Australia models were the most frequently validated, with pooled C statistics of 0.73 (95% CI = 0.69–0.78) and 0.70 (95% CI = 0.66–0.74), respectively.
Models incorporating gene expression profiles demonstrated slightly higher discrimination (pooled C statistic = 0.83, 95% CI = 0.76–0.90) compared with those using only clinicopathologic features (pooled C statistic = 0.77, 95% CI = 0.73–0.81), though the difference was not statistically significant (P = .11).
Lead study author Bryan Ma, MD, of the University of Calgary, and colleagues emphasized the importance of individualized risk assessment in sentinel lymph node biopsy decision-making, "By providing an individualized and objective measure of predicted risk, these models can better guide informed consent," the study authors suggested.
The systematic review followed rigorous methodology, searching Embase and MEDLINE (through May 1, 2024) and adhering to CHARMS and TRIPOD reporting guidelines. Despite strong external validation of several models, the investigators underscored the need for head-to-head comparisons to determine the best-performing nomogram for clinical use.
No conflicts of interest were reported.