Impact of modified Glasgow prognostic score on predicting prognosis and modification of risk model for patients with metastatic renal cell carcinoma treated with first line tyrosine kinase inhibitor.

Abstract

Intermediate risk group of the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) criteria is thought to consist of patients with different prognoses. This study investigated the impact of a pretreated modified Glasgow prognostic score (mGPS), which is defined on the basis of the pretreated serum albumin and C-reactive protein level, on predicting the prognosis of patients with metastatic renal cell carcinoma (mRCC) and its usefulness for the re-stratification of patients into a more improved risk model. One hundred ninety-six mRCC patients treated with first-line tyrosine kinase inhibitor (TKI) were retrospectively investigated. All patients were classified into either a high-mGPS or a low-mGPS group on the basis of mGPS score upon starting systemic therapy, the overall survival (OS) and cancer specific survival (CSS) rates in each group were compared. We use decision curve analysis and calculate C-index based on OS and CSS to compare IMDC+mGPS model and IMDC model. The categories of favorable, intermediate, and poor risk groups in the IMDC model were assessed in 32, 113, and 51 cases, respectively. The low- and high-mGPS groups consisted of 149 and 47 cases. The median OS in the high- and low-mGPS groups were 38.4 months and 5.6 months, and their median CSSs were 41.0 months and 5.6 months, respectively (P < 0.0001). Multivariate analysis showed that a high mGPS, multiple metastatic organs, and hypercalcemia were independent predictive factors for a worse OS (P = 0.0260). Next, we divided the intermediate risk group into two subgroups using the mGPS score. The OS and CSS for the high-mGPS subgroup were significantly worse than those for the low-mGPS one (P = 0.0024, median OS: 21.0 months and 33.7 months, P = 0.0007, median CSS: 21.0 months and 39.8 months), and there was no significant difference in OS between the high-mGPS subgroup in the intermediate risk group and poor risk group (P = 0.2250). The value of C-index based on OS at IMDC and IMDC+mGPS model were 0.6771 and 0.6967, and those based on CSS were 0.6850 and 0.7080, respectively. In decision curve analysis to evaluate the clinical net benefit using the IMDC+mGPS model compared to the IMDC model, there was no significant difference between the two groups. mGPS is useful for establishing a more improved prognostic model that is able to stratify mRCC patients treated with first-line TKI.