Prechemoradiotherapy systemic inflammation response index stratifies stage IIIB/C non-small-cell lung cancer patients into three prognostic groups: a propensity score-matching analysis

Abstract

Purpose: we explored the prognostic influence of the systemic inflammation response index (SIRI) on the survival outcomes of stage IIIB/C non-small-cell lung cancer (NSCLC) patients who underwent concurrent chemoradiotherapy. Methods: present propensity score-matching (PSM) analysis comprised 876 stage IIIB/C NSCLC patients who received 1-3 cycles of platinum-based doublets concurrent with thoracic radiotherapy from 2007 to 2017. The primary and secondary objectives were the relationships between the SIRI values and overall (OS) and progression-free survival, respectively. Propensity scores were calculated for SIRI groups to adjust for confounders and to facilitate well-balanced comparability between the SIRI groups by creating 1 : 1 matched study groups. Results: the receiver operating characteristic curve analysis identified an optimal SIRI cutoff at 1.9 for OS (AUC: 78.8%; sensitivity: 73.7%; specificity: 70.7%) and PFS (AUC: 80.5%; sensitivity: 75.8%; specificity: 72.9%) and we grouped the patients into two PSM cohorts: SIRI < 1.9 (N = 304) and SIRI >= 1.9 (N = 304), respectively. The SIRI >= 1.9 cohort had significantly worse median OS (P<0.001) and PFS (P<0.001) than their SIRI < 1.9 companions. The further combination of SIRI with disease stage exhibited that the SIRI-1 (IIIB and SIRI < 1.9) and SIRI-3 (IIIC and SIRI >= 1.9) cohorts had the best and worst outcomes, respectively, with SIRI-2 cohort (IIIB and SIRI >= 1.9 or IIIC and SIRI < 1.9) being remained in between (P<0.001 for OS and PFS, separately). In multivariate analysis, the two- and three-laddered stratifications per the 1.9 cutoffs and SIRI groups retained their independent significance, individually. Conclusions: the SIRI >= 1.9 independently prognosticated significantly worse OS and PFS results and plated the stage IIIB/C patients into three fundamentally distinct prognostic groups.