Abstract
BACKGROUND: The inflammation-based Glasgow prognostic score (GPS) has been demonstrated to be prognostic for various tumors. We investigated the value of the modified GPS (mGPS) for the prognosis of patients undergoing curative resection for colorectal liver metastases (CRLM). METHODS: A total of 343 patients were enrolled onto this study. The mGPS was calculated as follows: mGPS-0, C-reactive protein (CRP) ≤10 mg/L; mGPS-1, CRP >10 mg/L and albumin ≥35 g/L; and mGPS-2, CRP >10 mg/L and albumin <35 g/L. Prognostic significance was retrospectively analyzed by univariate and multivariate analyses. RESULTS: Of the 343 patients, 295 (86.0 %) were assigned to mGPS-0, 33 (9.6 %) to mGPS-1, and 15 (4.4 %) to mGPS-2. The median disease-free survival of patients with mGPS-0, -1, and -2 was 18.3, 15.5, and 5.2 months, respectively. The median cancer-specific survival (CSS) of patients with mGPS-0, -1, and -2 was 89.5, 62.2, and 25.8 months, respectively. The CSS of patients with mGPS-0 was significantly longer than that of patients with mGPS-2. Multivariate analysis revealed a significant association between cancer-related postoperative mortality and mGPS and carcinoembryonic antigen level. CONCLUSIONS: The preoperative mGPS is a useful prognostic factor for postoperative survival in patients undergoing curative resection for CRLM.