1040P Nivolumab or atezolizumab in the second-line treatment of advanced non-small-cell lung cancer? Prognostic index based on real-world data

Abstract

The efficacy of nivolumab and atezolizumab in NSCLC was documented in prospective trials. We aimed to confirm the benefits and indicate prognostic factors in daily practice. This study was a retrospective analysis of patients (pts) treated in the daily practice in Poland. The median of progression-free (PFS) and overall survival (OS) were estimated using the Kaplan-Meier method. The log-rank test was used for comparisons. Cox regression analysis was used for multivariate analyses. SASS v.27.0 software was used. A total of 260 pts (ECOG 0-1) were qualified to nivolumab (134 pts) or atezolizumab (126 pts). 18% of pts ended the treatment before radiological assessment. At landmark time-points of 12 and 24 months 31 % and 9 % pts remained alive. Median PFS and OS were 3 (95% CI 2.57-3.42) and 10 months (95%CI 8.03-11.96) for overall population. For atezo and nivo mPFS were 3 vs 3 months (p=0.175), mOS - 8 vs 14 months (p<0,018). In a univariate analysis for all pts, sum of measurable lesions (SML) >100.5 mm (p<0.001), liver mets (p=0.024), bone mets (p<0.001), short response for chemo (<4 mths) (p<0.001), BMI <25.6 (p=0.039), platelets (PLT) >281.5 G/l (p <0.007), a neutrophil-to-lymphocyte ratio >3.68 (p = 0.0001), platelet-to-lymphocyte ratio >182.9 (p<0.001) and LIPI 2 score (p=0.028) had a negative impact on OS. Other factors were less relevant. In a multivariate analysis SML >100.5mm (p=0.007; HR=1.003, 95% CI 1.001- 1.005), PLT>281,5 G/l (p<0.001; HR=1.003, 95%CI 1.001-1.003) and bone mets (p<0.004; HR=1.58, 95% CI 1.04-2.38) were independent negative prognostic factors for OS. Drug type was not shown to affect OS. Based on analyses, a prognostic index was constructed with formula: PI = 0.002 x (Platelets) + 0.003 x (Bones) - 0.725 x (SML). After estimating regression coefficients for individual variables, points were assigned to them to obtain three prognostic groups. Median OS differed significantly-16 (95%CI 13.3-18.7), 7 (95% CI 4.83-9.17) and 4 months (95%CI 2.88-5.13), respectively. Nivolumab and atezolizumab produced similar clinical benefit, but the treatment outcomes were worse than in clinical trials. Prognostic indices may help to indicate subgroups likely to benefit.