Data_Sheet_1_Using Systemic Inflammatory Markers to Predict Microvascular Invasion Before Surgery in Patients With Hepatocellular Carcinoma.docx

BackgroundMounting studies reveal the relationship between inflammatory markers and post-therapy prognosis. Yet, the role of the systemic inflammatory indices in preoperative microvascular invasion (MVI) prediction for hepatocellular carcinoma (HCC) remains unclear.
Patients and MethodsIn this study, data of 1,058 cases of patients with HCC treated in the First Affiliated Hospital of Sun Yat-sen University from February 2002 to May 2018 were collected. Inflammatory factors related to MVI diagnosis in patients with HCC were selected by least absolute shrinkage and selection operator (LASSO) regression analysis and were integrated into an “Inflammatory Score.” A prognostic nomogram model was established by combining the inflammatory score and other independent factors determined by multivariate logistic regression analysis. The receiver operating characteristic (ROC) curve and the area under the curve (AUC) were used to evaluate the predictive efficacy of the model.
ResultsSixteen inflammatory factors, including neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, etc., were selected by LASSO regression analysis to establish an inflammatory score. Multivariate logistic regression analysis showed that inflammatory score (OR = 2.14, 95% CI: 1.63–2.88, p < 0.001), alpha fetoprotein (OR = 2.02, 95% CI: 1.46–2.82, p < 0.001), and tumor size (OR = 2.37, 95% CI: 1.70–3.30, p < 0.001) were independent factors for MVI. These three factors were then used to establish a nomogram for MVI prediction. The AUC for the training and validation group was 0.72 (95% CI: 0.68–0.76) and 0.72 (95% CI: 0.66–0.78), respectively.
ConclusionThese findings indicated that the model drawn in this study has a high predictive value which is capable to assist the diagnosis of MVI in patients with HCC.

研究背景
越来越多的研究揭示了炎症标志物与治疗后预后之间的关联。然而，全身炎症指标在肝细胞癌（hepatocellular carcinoma, HCC）术前微血管侵犯（microvascular invasion, MVI）预测中的作用仍不明确。
患者与方法
本研究收集了2002年2月至2018年5月于中山大学附属第一医院接受治疗的1058例肝细胞癌患者的临床数据。通过最小绝对收缩和选择算子（least absolute shrinkage and selection operator, LASSO）回归分析筛选出与肝细胞癌患者MVI诊断相关的炎症因子，并将其整合为「炎症评分」。通过联合该炎症评分与多因素logistic回归分析筛选出的其他独立危险因素，构建预后列线图模型。采用受试者工作特征（receiver operating characteristic, ROC）曲线及曲线下面积（area under the curve, AUC）评估模型的预测效能。
研究结果
经LASSO回归分析筛选出包括中性粒细胞与淋巴细胞比值、血小板与淋巴细胞比值等在内的16种炎症因子以构建炎症评分。多因素logistic回归分析结果显示，炎症评分（比值比（odds ratio, OR）=2.14，95%置信区间（confidence interval, CI）：1.63~2.88，P<0.001）、甲胎蛋白（alpha fetoprotein）（OR=2.02，95%CI：1.46~2.82，P<0.001）以及肿瘤大小（OR=2.37，95%CI：1.70~3.30，P<0.001）均为MVI的独立危险因素。基于上述三种因素构建用于MVI预测的列线图。训练组与验证组的曲线下面积分别为0.72（95%CI：0.68~0.76）与0.72（95%CI：0.66~0.78）。
研究结论
本研究结果表明，本次构建的模型具有较高的预测价值，可辅助肝细胞癌患者的MVI诊断。