Supplementary Figure S6 from Biomarker Analysis and Treatment Dynamics Following Preoperative Ipilimumab plus Nivolumab in Locally Advanced Urothelial Cancer from the Phase IB NABUCCO Study

Multiomics predictive model for response to ICB using pre-treatment biomarkers. a. Fitted logistic regression statistic with baseline TGFβ, TMB and PD-L1 TPS for ICB response prediction; columns show the t-statistic and p-value for each biomarker; b. Receiver-operator-characteristic (ROC) curve for the predictive performance of the multiomic model; a leave one patient out method was used to quantify the predictive potential of the model. The ROC curve performs better than a random prediction when the area under the curve >0.5. c. A plot of the pair-wise correlation between all biomarker pairs based on the Pearson correlation coefficient. Orange-red colors represent a negative correlation, and blue-dark blue colors represent a positive correlation.