33P Nomogram to predict survival of patients with unresectable melanoma receiving immune checkpoint inhibitors

40-60% of patients with advanced melanoma receiving immune checkpoint inhibitors (ICI) do not derive benefit. Baseline predictive and prognostic biomarkers for ICI are currently missing. Here we aim to identify clinicopathologic features using a random survival forest (RSF) develop nomogram PFS in this collective. We included stage IV unresectable first-line documented blood markers at the time therapy start. The variable importance was calculated screen predictors progression-free (PFS) an RSF. developed Cox regression prediction model based on RFS results established PFS. Performance evaluated calibration plots, Brier score, discrimination by Harrel C-Index AUC. also verified internal validation. Patients were divided into four subgroups quartiles total points (TP) nomogram. Kaplan Meier analysis performed. 296 median 25 months. Median age 67 (IQR:57, 77) years. RSF identified nine variables associated [LDH, type melanoma, neutrophile/lymphocyte rate (NLR), age, S100, number metastatic organs, presence liver or brain metastases, BMI] that used construct model. C-index 0.66, curves showed good correlation between predicted actual progression risks. below 63.88 (below first quartile TP, n=73) had 63 months (95% CI: 48-76), while group above 121.45 (above third n=75) 11 CI 7-19); HR=3.95 CI:2.63-5.94; p<0.001). patients. integrating markers, accuracy predicting can be risk stratification