LASSO-Driven Selection of Biochemical and Clinical Markers for Primary Resistance to PD-1 Inhibitors in Metastatic Melanoma.

Background and Objectives: Resistance to immune checkpoint inhibitors (ICIs) reduces treatment efficacy in 40-65% of patients. The ability to predict this at the outset of therapy could help optimise treatment selection and improve patient survival. The aim of this study was to identify factors associated with primary resistance to PD-1 inhibitors in metastatic melanoma and discover predictive markers. Materials and Methods: This retrospective study involved 110 patients with non-uveal metastatic melanoma treated with PD-1 inhibitors from 2016 to 2023. Demographic, clinical and haematological data were collected. LASSO regression was utilised to identify the best markers. Bootstrap resampling was performed for internal validation and to overcome overfitting. Results: Primary resistance occurred in 44.6% of the patients. The factors associated with resistance included elevated platelet-to-lymphocyte ratio (PLR), the presence of acral/mucosal melanoma, BRAF mutant disease, low globulin levels and ≥3 metastatic sites. An evaluation of the predictive capability of these variables showed robust discrimination, with an area under the receiver operating characteristic curve of 0.831. Conclusions: This study identified the key predictors of primary resistance to PD-1 inhibitors to be PLR, globulin levels, metastatic burden and melanoma subtype. These identified parameters may guide the early prediction of primary resistance to PD-1 inhibitors. Future work should externally validate the model and further explore robust strategies to overcome resistance.