Predictive value of preoperative peripheral blood neutrophil/lymphocyte ratio for lymph node metastasis in patients of resectable pancreatic neuroendocrine tumors: A nomogram-based study

Abstract

Background: Neutrophil-to-lymphocyte ratio (NLR) is one of the systemic inflammation markers, which has prognostic values in many types of tumor. However, hardly any research has reported the relationship between NLR and pancreatic neuroendocrine tumors (PanNETs). In this study, we aimed to evaluate the predictive value of the preoperative peripheral blood NLR on the clinical outcomes in patients of resectable PanNETs. Methods: Ninety-five cases of PanNETs registered in the First Affiliated Hospital of Zhejiang University between March 2009 and May 2016 and underwent pancreatic surgery were included in this study. Univariate and multivariate analyses were applied to identify the prognostic factors for PanNETs. Prognostic nomogram and its calibration curve then used R (version 3.3.2) to predict lymph node (LN) metastasis. Results: Among these 95 patients, 52 (54.7%) patients were diagnosed as grade 1 (G1) NET (mitotic count <2/10 HPF, Ki-67 ≤2%), 32 (33.7%) as G2 NET (mitotic count 2-20/10 HPF, Ki-67 3-20%), and 11 (11.6%) as G3 NEC (mitotic count >20/10 HPF, Ki-67 >20%). Increased NLR was found to relate with advanced T stage, LN metastasis, tumor thrombus formation, and advanced grade (p < 0.05 for all). Multivariate logistic regression was performed and indicated that NLR (HR 6.74; p = 0.02) was an independent prognostic factor for LN metastasis. Furthermore, we developed a nomogram based on the combination of NLR, T stage, and grade for LN metastasis with a good discrimination ability with the AUC (area under the curve) of 0.885. This nomogram showed larger AUC than those using NLR (0.725), T stage (0.808), or grade (0.708) alone as a prognostic factor, which means this system achieved a more optional performance in predicting clinical outcomes. Finally, the Kaplan-Meier curve indicated that the recurrence-free survival (RFS) of patients with high NLR (NLR >1.40, RFS 61.1 ± 4.4 months) decreased significantly as compared with those of low NLR (NLR ≤1.40, RFS 63.8 ± 2.9 month, p < 0.05). Conclusions: The preoperative NLR is a potential independent predictor for LN metastasis and RFS. Our nomogram highlighted the important role of NLR in prognosis, which might be considered as a convenient indicator for lymph node metastasis, especially during the initial diagnosis for resectable PanNETs.