Comparison of the Glasgow Prognostic Score (GPS) and the modified Glasgow Prognostic Score (mGPS) in evaluating the prognosis of patients with operable and inoperable non-small cell lung cancer

Abstract

Purpose: The Glasgow Prognostic Score (GPS) and modified Glasgow Prognostic Score (mGPS) are shown to be reliable prognostic indexes in patients with operable and inoperable non-small cell lung cancer (NSCLC). Considering the difference between the two indexes lies in whether hypoalbuminemia without an elevated C-reactive protein (CRP) is associated with worse survival, this study aims to evaluate the prognostic performance of hypoalbuminemia in patients without an elevated CRP and to compare the prognostic value of GPS and mGPS in patients with operable and inoperable NSCLC.

Methods: The data of 2988 patients were retrospectively collected from the Shanghai Health Information Network. Univariate and multivariate Cox regression was performed to investigate the prognostic effect of albumin, CRP, GPS and mGPS. Restricted cubic spline was also performed to evaluate the relationship between albumin and hazard ratio. Kaplan-Meier survival curves were estimated and compared using the log-rank test. Additional discriminative ability of GPS and of mGPS was evaluated using the area under the curve and Harrell's concordance index.

Results: Hypoalbuminemia was associated with worse survival in both operable and inoperable patients without an elevated CRP. The Kaplan-Meier survival curve of hypoalbuminemic patients without an elevated CRP was more close to the curve of patients with an elevated CRP and a normal albumin than to the curve of patients with neither of these abnormalities. Multivariate analysis, AUC and C-index all indicated that GPS had a higher prognostic value than mGPS.

Conclusions: Hypoalbuminemia was associated with worse survival in patients with or without an elevated CRP. GPS was superior to mGPS in predicting survival in operable and inoperable NSCLC patients.

Keywords: Albumin; C-reactive protein; Glasgow Prognostic Score; Modified Glasgow Prognostic Score; Non-small cell lung cancer; Survival.

Fig. 1

Flowchart indicating participants remaining (NSCLC non-small cell lung cancer, CRP C-reactive protein)

Fig. 2

Spline curves of overall survival for baseline albumin concentrations in NSCLC patients, presented as log hazards (solid lines) and 95 % CI (dashed lines) in univariate Cox models. 35 g/L is the reference standard. The vertical reference line was set at 35 g/L; the horizontal reference line was set at 0. a Operable patients with CRP ≤ 10 mg/L. b Operable patients with CRP > 10 mg/L. c Inoperable patients with CRP ≤ 10 mg/L. d Inoperable patients with CRP > 10 mg/L. Multivariable-adjusted spline curves behave similarly with these unadjusted curves with wider confidence intervals

Fig. 3

The Kaplan–Meier curves of overall survival in NSCLC patients according to the pretreatment CRP and albumin concentration. a Operable patients. b Inoperable patients (CRP C-reactive protein, Alb albumin)