Composite versus individual biomarkers for predicting clinical benefit to PD-1 blockade in NSCLC 109 3 ◀ Figure S2. Performance of PD-1T TILs and TIS as individual and as composite biomarker to predict disease control at 6 months (DC 6m) in NSCLC patients treated with PD-1 blockade. (A) Specificity correlating to a sensitivity of ≥90% for combinations with PD-1T TILs as predictive biomarker for DC 6m in the training cohort (n=27 or n=42). The grey dashed line indicates the prespecified specificity criterium of ≥50%. Different composite biomarkers were compared to the predictive performance of PD-1T TILs alone. P values were calculated by McNemar test. (B) Same plot as in A for combinations with TIS in the training cohort (n=27 or n=28). (C) PD1T TILs per mm2 in pretreatment samples from patients with DC 6m (n=12) and progressive disease (PD) (n=30) in the training cohort (n=42). Dashed line indicates a cut-off of 90 PD-1T TILs per mm2. Medians, interquartile ranges and minimum/maximum shown in boxplots, ***P<0.001 by Mann Whitney U-test. (D) TIS scores in pretreatment samples from patients with DC 6m (n=8) and PD (n=20) in the training cohort (n=28). Dashed line indicates a cut-off score of 6.65. Medians, interquartile ranges and minimum/maximum shown in boxplots, **P<0.01 by Mann Whitney U-test. (E) Receiver operating characteristic (ROC) curve for predictive value of PD-1T TILs for DC 6m in the training cohort (n=42) (AUC 0.82; 95% CI: 0.690.95) and validation cohort (n=61) (AUC 0.72; 95% CI: 0.57-0.87). (F) ROC curve for predictive value of TIS for DC 6m in the training cohort (n=28) (AUC 0.81; 95% CI: 0.65-0.98) and validation cohort (n=40) (AUC 0.57; 95% CI: 0.36-0.77) (G) Same plot as in C (PD-1T TILs) for patients with DC 6m (n=18) and PD (n=43) in the validation cohort (n=61), **P<0.01 by Mann Whitney U-test. (H) Same plot as in D (TIS) for patients with DC 6m (n=12) and PD (n=28), P=0.52 by Mann Whitney U-test.
RkJQdWJsaXNoZXIy MTk4NDMw