【作者】 赵汗青；

【导师】 赵松；

【作者基本信息】 郑州大学 ， 外科学（专业学位）， 2023， 硕士

【摘要】 研究背景肺神经内分泌肿瘤(Lung Neuroendocrine Tumor,LNET)是起源于肺部神经内分泌细胞的一类上皮性肿瘤,约占原发性肺肿瘤的20%,可分为4种亚型:典型类癌(Typical Carcinoid,TC)、非典型类癌(Atypical Carcinoid,AC)、大细胞神经内分泌癌(Large Cell Neuroendocrine Carcinoma,LCNEC)和小细胞肺癌(Small Cell Lung Cancer,SCLC)。这类肿瘤在组织学及病理学上明显区别于非小细胞肺癌。由于LNET具有罕见性和异质性,故对于LNET的报道较少。目前,手术是早期LNET患者的首选,部分学者认为早期或部分可切除的LNET患者可以在手术中获益,但由于LNET的不同亚型存在较大的异质性,患者手术预后差异很大。目前由美国癌症联合委员会(American Joint Committee on Cancer,AJCC)所建立的TNM癌症分期系统被广泛用于肿瘤预后的预测,但未纳入患者年龄、性别、肿瘤分化程度、病理类型及治疗方式等其他重要的预后影响因素,对预测个体患者的预测效果欠佳。因此,需要建立新的临床预测模型以指导更恰当的治疗选择和更精确的预后评估。研究目的本研究的目的是分析研究影响可切除性肺神经内分泌肿瘤预后的临床病理因素,并借列线图(Nomogram)这个工具,构建个体化预测模型,以预测患者的总生存率(Overall Survival,OS)。并与第8版TNM分期系统进行比较。旨在为肺神经内分泌肿瘤患者提供更个体化的预后评估方法,在一定程度上为临床决策提供指导方案。研究方法从美国国立癌症研究所建立的监测、流行病学和最终结果(Surveillance,Epidemiology,and End Results,SEER)数据库中提取2010年至2019年间行手术切除治疗的并且经病理确诊为肺神经内分泌肿瘤患者的资料,包括年龄、性别、种族、婚姻状态、原发部位、偏侧性、分化程度、组织学类型、AJCC分期、手术术式、淋巴结阳性率、是否行放化疗和术后随访资料(如生存状态和总生存期)等。将患者随机分为两组,70%纳入训练集,30%纳入验证集,应用卡方检验对患者临床病理特征进行比较。在训练集中首先采用单因素和多因素Cox比例风险回归分析影响生存预后的危险因素。在多因素Cox回归分析的基础上建立列线图生存预测模型,绘制术后1、3、5年的总生存率(OS)的列线图。采用Bootstrap自由抽样法,重复抽样1000次得到校准曲线(Calibration curve),使用校准曲线评估预测模型的校准度。分别在训练集和验证集中计算受试者工作曲(Receiver Operating Characteristic,ROC)的曲线下面积(Area Under Curve,AUC)和C-指数(Concordance index,C-index)评估预测模型的区分度,使用决策分析曲线(Decision Curve Analysis,DCA)评价模型的临床实用性并与第8版美国癌症联合委员会的TNM癌症分期系统进行比较。基于该预后模型建立了相对应的风险分层系统,并应用Kaplan-Meier生存分析对其进行检验。以上数据分析及绘图均通过R软件(版本4.2.1)进行。研究结果从SEER数据库中共筛选出4074例符合条件的肺神经内分泌瘤患者。经过单因素与多因素Cox比例风险回归分析之后,年龄、性别、婚姻状态、组织学类型、T分期、淋巴结阳性率和是否化疗被证明是独立的预后因素。基于上述Cox回归分析的结果构建了可切除性肺神经内分泌肿瘤总生存率(OS)的列线图预测模型。校准曲线显示在训练集和验证集中该模型具有良好的预测校准度。进一步对回归模型区分度分析,通过C指数来评估列线图的预测准确性。在训练集中,预测模型的C指数分别为0.826(95%CI,0.810-0.842),第8版TNM分期系统的C指数为0.714(95%CI,0.690-0.738)。在验证集中,预测模型的C-指数分别为0.814(95%CI,0.789-0.838),第8版TNM分期系统的C指数为0.693(95%CI,0.654-0.733)。在训练集中比较了1、3、5年生存率的ROC曲线,结果表明,预测模型1、3、5年生存率的AUC分别为0.847、0.843和0.837,而第8版TNM分期系统的AUC分别为0.707、0.691和0.674,表明预测模型具有比第8版TNM分期系统更好的区分度,该结论在验证集也得到了验证。DCA曲线也显示在训练集和验证集中,预测模型的净获益率高于第8版TNM分期系统,具有更好的临床实用性。此外,通过对预测模型建立风险分层系统,可更好的区分高风险和低风险人群。研究结论本研究结果显示年龄、性别、婚姻状态、组织学类型、T分期、淋巴结阳性率和是否化疗是可切除性肺神经内分泌肿瘤患者术后总生存率的独立影响因素。与第8版肺癌TNM分期系统相比,本研究构建的预测模型在预测总生存率方面具有更好的准确性和临床适用性,将为临床工作者提供更加个体化的预后评估方法。更多还原

【Abstract】 BackgroundLung neuroendocrine tumor(LNET)is a type of epithelial tumor that originates from neuroendocrine cells in the lung and accounts for approximately 20%of primary lung tumors.Large Cell Neuroendocrine Carcinoma(LCNEC),and Small Cell Lung Cancer(SCLC).These tumors are clearly distinguished from non-small cell lung cancer histologically and pathologically.Because of the rarity and heterogeneity of LNET,there are fewer reports of lung NETs.Currently,surgery is the first choice for patients with early stage LNET,and some scholars believe that some patients with resectable or early stage LNET can benefit from surgery,but due to the large heterogeneity of different subtypes of lung NET,the prognosis of patients with surgery varies greatly.The TNM cancer staging system established by the American Joint Committee on Cancer(AJCC)is widely used for predicting tumor prognosis,but it does not incorporate other important prognostic factors such as patient age,gender,degree of tumor differentiation,pathological type and treatment modality,and has poor predictive effect on individual patients.Therefore,new clinical prediction models are needed to guide more appropriate treatment selection and more accurate prognostic assessment.ObjectiveThe purpose of this study was to analyze and study the clinicopathological factors affecting the prognosis of resectable lung neuroendocrine tumors and to construct an individualized prediction model to predict the overall survival(OS)of patients by means of the columnar nomogram(Nomogram)as a tool.It was also compared with the 8th edition TNM staging system.The aim is to provide a more individualized assessment method for patients with lung neuroendocrine tumors,and to some extent,to provide a guiding scheme for clinical decision making.MethodsInformation on patients treated with surgical resection and pathologically confirmed lung neuroendocrine tumors between 2010 and 2019 was extracted from the Surveillance,Epidemiology,and End Results(SEER)database established by the National Cancer Institute,including information on age,gender,race,marital status,primary site,laterality,degree of differentiation,histological type,AJCC stage,surgical procedure,positive lymph node rate,whether radiotherapy was administered and postoperative follow-up information(e.g.survival status and overall survival).The enrolled patients were randomly divided into two groups,with 70%included in the training set and 30%in the validation set,and the cardinality test was applied to compare the clinicopathological characteristics of the patients.The risk factors affecting survival prognosis were first analyzed in the training set using univariate and multifactorial Cox proportional risk regression.A column nomogram survival prediction model was developed based on the multifactorial Cox regression model to plot column nomograms of overall survival(OS)at 1,3,and 5 years after surgery.The calibration curve was obtained by Bootstrap free sampling method with 1000repetitions,and the calibration degree of the prediction model was evaluated using the calibration curve.Area Under Curve(AUC)of Receiver Operating Characteristic(ROC)and C-index were calculated in the training and validation sets,respectively,to evaluate the discrimination of the prediction model using the decision curve(DCA)to evaluate the clinical utility of the model and to compare it with the 8th edition of the American Joint Committee on Cancer TNM cancer staging system.A corresponding risk stratification system was established based on the prognostic model and tested using Kaplan-Meier survival analysis.The above data analysis was performed by R software(Version 4.2.1).ResultsA total of 4074 eligible patients with lung neuroendocrine tumors were screened from the SEER database.After univariate and multifactorial Cox proportional risk regression analysis,age,gender,marital status,histological type,T-stage,lymph node positivity rate and the presence or absence of chemotherapy were shown to be independent prognostic factors.A columnar nomogram prediction model for overall survival(OS)of resectable lung neuroendocrine tumors was constructed based on the results of the above Cox regression analysis.The calibration curves showed that the model had good predictive calibration in both the training and validation sets.Further regression model differentiation analysis was performed to assess the prediction accuracy of the column nomogram by C-index.In the training set,the C-index was0.826(95%CI,0.810-0.842)for the prediction model and 0.714(95%CI,0.690-0.738)for the version 8 TNM staging system,respectively.In the validation set,the C-index was 0.814(95%CI,0.789-0.838)for the prediction model and 0.693(95%CI,0.654-0.733)for the 8th TNM staging system,respectively.The ROC curves for 1-,3-,and 5-year survival were compared in the training set,and the results showed that the AUCs of the prediction model for 1-,3-,and 5-year survival were 0.847,0.843,and 0.837,respectively,while those of the 8th version of the TNM staging system were 0.707,0.691,and 0.674,respectively,indicating that the prediction model had better discrimination than the 8th TNM staging system,and the The DCA curves also showed that the net benefit rate of the prediction model was higher than that of the 8th TNM staging system in both the training and validation sets,which had better clinical utility.In addition,by establishing a risk stratification system for this model,a better distinction could be made between high-risk and low-risk groups.ConclusionThe results of this study showed that age,gender,marital status,histological type,T-stage,lymph node positivity rate and chemotherapy were independent factors influencing overall survival after surgery in patients with resectable lung neuroendocrine tumors.Compared with the 8th edition of the TNM staging system for lung cancer,the prediction model constructed in this study has better accuracy and clinical applicability in predicting overall survival,and will provide clinical practitioners with a more individualized approach to prognostic assessment.更多还原