Objective: To evaluate the prognostic significance of inflammatory biomarkers, prognostic nutritional index and clinicopathological characteristics in tongue squamous cell carcinoma (TSCC) patients who underwent cervical dissection.
Methods: The retrospective cohort study consisted of 297 patients undergoing tumor resection for TSCC between January 2017 and July 2018. The study population was divided into the training set and validation set by 7 :3 randomly. The peripheral blood indices of interest were preoperative neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), systemic inflammation score (SIS) and prognostic nutritional index (PNI). Kaplan-Meier survival analysis and multivariable Cox regression analysis were used to evaluate independent prognostic factors for overall survival (OS) and disease-specific survival (DSS). The nomogram's accuracy was internally validated using concordance index, receiver operating characteristic (ROC) curve, area under the curve (AUC), calibration plot and decision curve analysis.
Results: According to the univariate Cox regression analysis, clinical TNM stage, clinical T category, clinical N category, differentiation grade, depth of invasion (DOI), tumor size and pre-treatment PNI were the prognostic factors of TSCC. Multivariate Cox regression analysis revealed that pre-treatment PNI, clinical N category, DOI and tumor size were independent prognostic factors for OS or DSS (P < 0.05). Positive neck nodal status (N≥1), PNI≤50.65 and DOI > 2.4 cm were associated with the poorer 5-year OS, while a positive neck nodal status (N≥1), PNI≤50.65 and tumor size > 3.4 cm were associated with poorer 5-year DSS. The concordance index of the nomograms based on independent prognostic factors was 0.708 (95%CI, 0.625-0.791) for OS and 0.717 (95%CI, 0.600-0.834) for DSS. The C-indexes for external validation of OS and DSS were 0.659 (95%CI, 0.550-0.767) and 0.780 (95%CI, 0.669-0.890), respectively. The 1-, 3- and 5-year time-dependent ROC analyses (AUC = 0.66, 0.71 and 0.72, and AUC = 0.68, 0.77 and 0.79, respectively) of the nomogram for the OS and DSS pronounced robust discriminative ability of the model. The calibration curves showed good agreement between the predicted and actual observations of OS and DSS, while the decision curve confirmed its pronounced application value.
Conclusion: Pre-treatment PNI, clinical N category, DOI and tumor size can potentially be used to predict OS and DSS of patients with TSCC. The prognostic nomogram based on these variables exhibited good accurary in predicting OS and DSS in patients with TSCC who underwent cervical dissection. They are effective tools for predicting survival and helps to choose appropriate treatment strategies to improve the prognosis.
目的: 评估术前炎症生物标志物、预后营养指数和临床病理特征对舌鳞状细胞癌(tongue squamous cell carcinoma, TSCC)患者行根治性切除术后生存结局的预后价值, 并以此构建患者预后预测列线图模型。
方法: 回顾性收集2017年1月至2018年7月于北京大学口腔医院接受根治性肿瘤切除术的297例TSCC患者的病例资料, 随机按照7 :3比例分为训练集和验证集。分析患者术前全身炎症反应标志物中性粒细胞/淋巴细胞比率(neutrophil-to-lymphocyte ratio, NLR)、淋巴细胞/单核细胞比率(lymphocyte-to-monocyte ratio, LMR)、血小板/淋巴细胞比率(platelet-to-lymphocyte ratio, PLR)、系统免疫炎症指数(systemic immune-inflammation index, SII)、系统性炎症评分(systemic inflammation score, SIS)及预后营养指数(prognostic nutritional index, PNI)与TSCC患者术后总生存期(overall survival, OS)和疾病特异性生存期(disease-specific survival, DSS)的相关性。使用X-tile软件确定连续变量的最佳截断值作为分界点。采用Kaplan-Meier生存分析和多变量Cox回归模型分析影响TSCC患者的独立预后预测因素, 据此构建OS和DSS的生存相关列线图预测模型, 通过训练集和验证集进行模型内部交叉验证和外部验证, 具体通过一致性指数、时间依赖性受试者工作特征(receiver operating characteristic, ROC)曲线、曲线下面积(area under the curve, AUC)、校准曲线和决策曲线分析对列线图的准确性进行验证。
结果: 单因素Cox回归分析显示, TNM分期、T分期、N分期、分化程度、侵袭深度(depth of invasion, DOI)、肿瘤直径和治疗前PNI水平为影响TSCC预后危险因素; 多因素Cox回归分析显示, 治疗前PNI水平、N分期、DOI和肿瘤直径为患者5年OS或DSS的独立预后因素(P＜0.05)。治疗前N分期≥1、PNI≤50.65和DOI＞2.4 cm与较差的5年OS显著相关, 而N分期≥1、PNI≤50.65、肿瘤直径＞3.4 cm与较差的5年DSS显著相关。基于独立预后因素构建的TSCC术后患者OS和DSS的列线图预测模型的一致性指数分别为0.708 (95%CI, 0.625~0.791)和0.717 (95%CI, 0.600~0.834), 验证集验证结果显示, OS和DSS列线图预测模型的一致性指数为0.659 (95%CI, 0.550~0.767)和0.780 (95%CI, 0.669~0.890)。OS列线图模型和DSS列线图模型的1年、3年和5年的时间依赖性ROC分析(AUC分别为0.66、0.71、0.72和0.68、0.77、0.79)表明模型具有稳定的判别能力。校准曲线显示OS和DSS预测估值与实际观察值之间具有良好的一致性, 决策曲线分析反映模型具有较好的临床应用价值。
结论: 治疗前PNI、N分期、DOI和肿瘤直径可能对TSCC患者的OS和DSS有可靠的预测价值, 基于这些参数构建的预后预测列线图在预测TSCC根治性切除术后患者的OS和DSS方面表现出良好的准确性和有效性, 是评估生存结局的有效工具, 有助于选择有针对性的联合治疗来改善患者预后。
Keywords: Disease-specific survival; Nomogram; Overall survival; Prognostic nutritional index; Prognostic prediction model; Tongue squamous cell carcinoma.