Search

• By retrospective study on 308 patients with aNSCLC treated with chemoimmunotherapy. • PFS was similar to that reported in clinical trials. • Number of metastatic sites, histology and SII were independent prognostic factors. • These factors might contribute to a highly discriminating prognostic model. Efficacy outcomes and prognostic factors of real-world patients with advanced non-small cell lung cancer (aNSCLC) treated with first-line chemoimmunotherapy are still limited. In the retrospective Spinnaker study, data was collected from patients in six United Kingdom and one Swiss oncology centres with first-line pembrolizumab plus platinum-based chemotherapy. Efficacy outcomes and potential prognostic factors were estimated aiming at developing a prognostic model. Three-hundred-eight patients were included, 32% ≥ 70 years, with ≥ 3 metastatic sites in 33%, brain or liver metastases in 10% and 12%, respectively. With a median follow-up of 18.0 months (mo.) (range, 15.9–20.1), median overall survival (OS) and progression-free survival (PFS) were 12.7 mo. (range, 10.2–15.2), and 8.0 mo. (range, 7.1–8.8), respectively. The neutrophils-to-lymphocytes ratio (NLR) and systemic immune-inflammatory index (SII) (i.e., NLR × platelet count) were both significantly higher in ECOG PS 1 (p = 0.0147 and p = 0.0018, respectively), underweight or normal body mass index (p = 0.0456 and p = 0.0062, respectively), ≥3 metastatic sites (p = 0.0069 and p = 0.112), pretreatment steroids (p = 0.0019 and p = 0.0017). By MVA, the number of metastatic sites ≥ 3 (p < 0.001 and p = 0.002), squamous histology (p = 0.033 and p = 0.013) and SII ≥ 1444 (p = 0.031 and p = 0.009, respectively) were associated with both worse OS and PFS and led to a highly discriminating three-class risk prognostic model. Real-world PFS with chemoimmunotherapy in aNSCLC patients is similar to that reported in clinical trials. A high number of metastatic sites, squamous histology and high SII are adverse prognostic factors that might contribute to a clinically useful prognostic model. [ABSTRACT FROM AUTHOR]