【Abstract】 Objective To determine the associations of preoperative platelet-to-lymphocyte ratio (PLR) and derived neutrophil-to-lymphocyte ratio (d-NLR) with the prognosis of gastrointestinal stromal tumor (GIST). Methods GIST patients with surgical treatment from June 2005 to February 2015 in West China Hospital of Sichuan University were enrolled in the study. The results of blood routine tests of the patients within one week prior to surgery and their clinical data were extracted. The patients were divided into high-PLR/d-NLR (PLR＞153.075, d-NLR＞1.245) and low-PLR/d-NLR (PLR≤153.075, d-NLR≤1.245) groups according to the optimal cutoff values of the receiver operating characteristic (ROC) curves. Recurrence-free survival (RFS) rates were calculated using Kaplan-Meier method. COX regression analyses were performed to identify factors associated with RFS for GIST patients without imatinib treatment. Results Regardless of imatinib treatment, the patients with high PLR and d-NLR had shorter RFS than those with low PLR and d-NLR. Tumor diameter, location, mitotic counts, preoperative PLR and d-NLR were identified as factors associated with RFS in the univariate analyses. The multivariate analysis identified tumor diameter 〔≥5 cm, hazard ratio (HR ): 4.295, 95% confidence interval (CI ): 1.772-10.413, P =0.001〕, non-stomach (HR ：2.247, 95%CI : 1.200-4.209; P =0.011), mitotic counts (>5/50 HPF: HR ：4.678, 95%CI : 2.364-9.257; P HR ：2.549, 95%CI : 1.159-5.606; P =0.020) as independent factors predicting the prognosis of GIST. The patients with high PLR or high d-NLR had shorter RFS than those with low PLR/d-NLR. Conclusion Preoperative d-NLR is an independent predictor of RFS in GIST. PLR and d-NLR can be used in predicting the recurrence risk of GIST.
