Chapter 5 184 Within the validation sets, the number and proportions of patients assigned to each classification group were 37 (38%)/32 (43%) resistant, 30 (31%)/19 (25%) intermediate and 31 (32%)/24 (32%) sensitive for set 1/set 2 respectively. Kaplan-Meier plots of OS by test classification resistant vs. not resistant and sensitive vs. not sensitive are shown for the validation sets in Figure 3A-D. In validation set 1, Figures 3A and 3B, patients classified as resistant had significantly worse OS than not resistant patients (HR=0.60; 95% CI: 0.37- 0.97, P=0.037) and patients classified as sensitive had significantly better OS than not sensitive patients (HR=0.56; 95% CI: 0.33-0.97, P=0.038). One year survival for the sensitive group was 65% and the corresponding median was 15.3 (95% CI: 8.8-undefined) months. In contrast, median OS was only 4.8 (95% CI: 2.9-9.3) months in the resistant group, with 29% OS at one year. PFS was numerically superior in the sensitive group and inferior in the resistant group, but the differences in outcome were smaller and did not reach statistical significance, see Supplemental Data: Results Supplementary Figures 1 and 2. Analysis of the subgroup of patients treated with nivolumab in third or higher line (n=40), showed similar behavior in OS and PFS, with resistant patients showing a trend to shorter outcomes (HR=0.49; 95% CI: 0.23-1.04, P=0.062 for OS and HR=0.50; 95% CI: 0.25-1.02, P=0.057 for PFS) and sensitive patients showing numerically longer survival (HR=0.48; 95%CI: 0.21-1.10, P=0.082 for OS and HR=0.62; 95% CI: 0.31-1.23, P=0.172 for PFS). Kaplan-Meier plots for this subgroup are shown in the Supplementary data. Results for validation set 2 are shown in Figures 3C and 3D. Patients classified as resistant had worse OS than not resistant patients (HR=0.39; 95% CI: 0.19-0.77, P=0.007). The comparison of OS between the sensitive group and the not sensitive patients yielded a HR of 0.58, but did not show a significant difference (P=0.179). However, for ternary test classifications, the sensitive group had longer OS than the resistant group (HR=0.41; 95% CI: 0.18-0.94, P=0.036). Full analysis for the sensitive/intermediate/ resistant classifications can be found in Supplemental Data: Results. Analysis of PFS showed only numerical differences between classification groups. As results were consistent across cohorts, within the limits of relatively small subgroup sizes, a pooled analysis of all patients treated in second line with nivolumab was carried out stratified by cohort (n=249). There was no indication of any correlation of PD-L1 expression with test classification (P=0.292, 0.810, 0.337 for ternary, resistant vs. not resistant, sensitive vs. not sensitive test classifications), although positive PD-L1 expression was a predictor of improved OS and PFS in the pooled analysis (HR=1.60; 1.01-2.54, P=0.046 for OS and HR=1.61; 1.07-2.44, P=0.023 for PFS). Indeed, analysis including test classification and PD-L1 expression demonstrated both to be independent predictors of PFS (see Supplementary data). Within the
RkJQdWJsaXNoZXIy MTk4NDMw