早期胃癌淋巴结转移的列线图预测模型
Nomogram Model for Predicting Lymph Node Metastasis in Early Gastric Cancer
DOI: 10.12677/acm.2024.1451550, PDF, HTML, XML, 下载: 32  浏览: 86 
作者: 李瑞林*, 赵姣雅, 于照祥#:西安医学院研工部,陕西 西安
关键词: 早期胃癌淋巴结转移列线图Early Gastric Cancer Lymph Node Metastasis Nomogram
摘要: 目的:淋巴结转移(Lymph Node Metastasis, LNM)对于早期胃癌(Early Gastric Cancer, EGC)患者的预后有着至关重要的影响。因此需要对LNM患者进行彻底而精确的评估。确定影响LNM的因素,建立EGC患者LNM的预测模型。方法:收集并分析从监测、流行病学和最终结果(SEER)数据库下载的4749例EGC患者的临床信息和病理数据。基于7:3的比例,3324人被随机分为训练组,1425人被随机分配到测试组。根据训练集确定的影响LNM的因素,绘制并验证了列线图。结果:基于多因素分析,诊断时的性别、年龄、种族、分化程度、T期和肿瘤大小、远处转移、原发部位的手术、淋巴结捡取数目、原发肿瘤数目是EGC发生LNM的危险因素。此外,还绘制了列线图来预测EGC患者发生LNM的风险。肿瘤大小是LNM最重要的危险因素。对于训练集和测试集,列线图的受试者操作特征曲线下面积(AUC)分别为0.754和0.724。此外,校准曲线表明,LNM预测模型具有良好的一致性。结论:诊断时的分化程度、T期和肿瘤大小、远处转移、原发部位的手术、淋巴结捡取数目、原发肿瘤数目是EGC患者LNM的独立危险因素。基于上述危险因素,预测模型可能为EGC后续治疗方法的选择提供一些指导意义。
Abstract: Objective Lymph Node Metastasis (LNM) plays an important role in the prognosis of Early Gastric Cancer (EGC). Therefore, a thorough and precise assessment of patients with LNM is required. Determine the factors affecting LNM and establish a predictive model of LNM in EGC patients. Methods: The clinical and pathological data of 4749 patients with EGC downloaded from the Surveillance Epidemiology and Final Outcomes (SEER) database were collected and analyzed. Based on a 7:3 ratio, 3324 were randomly assigned to the training group and 1425 to the test group. Nomograms were plotted and validated based on the factors affecting the LNM identified from the training set. Results Based on multivariate analysis, the risk factors of LNM in EGC were sex, age, race, differentiation degree, T stage and tumor size, distant metastasis, primary site surgery, number of lymph nodes retrieved, and number of primary tumors. In addition, nomograms were plotted to predict the risk of developing LNM in patients with EGC. Tumor size is the most important risk factor for LNM. The area under the operator characteristic curve (AUC) of nomograms was 0.754 and 0.724 for the training and test sets, respectively. In addition, the calibration curve shows that the LNM prediction model has good consistency. Conclusion: Differentiation degree, T stage and tumor size at diagnosis, distant metastasis, primary site surgery, number of lymph nodes retrieved and number of primary tumors are independent risk factors for LNM in EGC patients. Based on the above-mentioned risk factors, the prediction model may provide some guiding significance for the selection of subsequent treatment methods for EGC.
文章引用:李瑞林, 赵姣雅, 于照祥. 早期胃癌淋巴结转移的列线图预测模型[J]. 临床医学进展, 2024, 14(5): 1275-1283. https://doi.org/10.12677/acm.2024.1451550

1. 介绍

胃癌(Gastric Cancer, GC),作为全球第三大最常见的癌症相关死亡原因,其风险指标包括幽门螺杆菌(H. pylori)感染、性别、饮食习惯、吸烟和家族史 ‎[1] 。可以使用胃萎缩的标志物(胃癌的前体病变)进行胃癌筛查,如血清胃蛋白酶原或血清生长素 ‎[2] ;或血清中针对Hp的抗体,这是GC的主要危险因素;或者使用内镜检查胃粘膜 ‎[3] 。

早期胃癌(EGC)被归类为局限于粘膜或粘膜下层的胃癌,与是否存在局部淋巴结转移无关(LNM) ‎[4] 。据调查,大约8.9%~15.8%的EGC病例有区域淋巴结转移,根治性手术治疗可达到足够的肿瘤清除和淋巴结清扫范围,复发率低,5年生存率为>90% ‎[5] 。与晚期胃癌相比,EGC有更好的机会通过手术成功切除,从而获得更好的生存状态 ‎[6] 。内镜切除适合于低LNM率的EGC,是EGC的首选治疗方法。内镜黏膜下剥离术(ESD)和内镜黏膜切除术(EMR)是ER的两种主要手术 ‎[7] 。晚期胃癌可以通过手术彻底切除,包括D2淋巴结切除术 ‎[8] 。尽管在过去的30年里,发达国家胃癌的发病率有所下降,但胃癌的总体预后仍然很差。晚期胃癌的五年存活率约为20% ‎[9] 。LNM对预后有很好的预测价值。因此,在EGC患者中,LNM的存在与否是一个需要全面评估的关键因素。

从监测、流行病学和最终结果(SEER)数据库中获得了大样本量EGC患者的相应临床病理信息,包括患者的临床病理参数和信息。可能与EGC患者预后相关的因素被纳入我们的研究,以探索它们的影响。据我们所知,很少有研究探索影响EGC患者LNM的因素。因此,我们绘制了一个预测模型,可以全面评估各种指标的影响,并为预测EGC患者的LNM提供了一个平台。

2. 材料和方法

2.1. 数据来源与患者选择

临床病理信息来自SEER数据库。排除的标准如下:1) 术前接受过新辅助治疗的患者;2) 残留胃癌患者;3) 临床病理资料不完整的患者;4) 未知淋巴结的回收;5) 活检未确诊为EGC的患者。

2.2. 统计分析

定量数据表示为均数 ± 标准差( x ¯ ± s ),并使用t检验。定性数据以率(n%)表示,并采用卡方检验或Fisher精确检验进行分析。在逻辑回归中,在单因素分析中存在显著差异的变量纳入多因素分析中,确定训练集的影响因素。所有统计分析均使用R软件(版本4.2.2)进行。双侧P < 0.05具有统计学显著差异,人口基线特征见表1

Table 1. Population baseline characteristics

表1. 人口基线特征

多因素Logistic回归分析显示,EGC确诊时的性别、年龄、种族、分化程度、T期和肿瘤大小、远处转移、原发部位的手术、淋巴结捡取数目、原发肿瘤数目是EGC患者LNM的影响变量(表2)。

Table 2. Significant indicators in multivariate Logistic regression analysis (P < 0.05)

表2. 多因素Logistic回归分析中有意义的指标(P < 0.05)

2.3. EGC患者预测模型的构建

构建了列线图预测模型(图1)。在模型中,每个变量的点数从0到100不等。每个指标都有其相应的评分,其中每个患者都有一个从相应的第一行得出的评分。总点是所有变量点的总和。然后,每个患者的总分对应于底部的概率,即LNM发生的概率。

Figure 1. Nomogram prediction model

图1. 列线图预测模型

2.4. 列线图的评价

(a) (b)

Figure 2. (a) Calibration curve for the training set, (b) Calibration curve for the test set

图2. (a) 训练集的校准曲线,(b) 测试集的校准曲线

(a) (b)

Figure 3. (a) DCA curve for the training set, (b) DCA curve for the test set

图3. (a) 训练集的DCA曲线,(b) 测试集的DCA曲线

Figure 4. The AUC of the training set is 0.754; the AUC of the verification set is 0.724

图4. 训练集的AUC为0.754;验证集的AUC为0.724

图2校准曲线显示,模型预测准确度较为不理想;而图3 DCA曲线显示,无论是训练集还是验证集,当阈值概率在0.2~0.5时,列线图预测模型净获益均高于全干预和全不干预模型,说明在一定范围内,列线图模型具有更好的临床适用度;图4 AUC提示构建的模型具有较好的区分度。

3. 讨论

GC在全球范围内具有重大影响。胃癌发生在胃粘膜的上皮细胞中,早期阶段也有发生血源性或LNM的趋势 ‎[10] 。随着对胃癌的认识越来越全面和深入,其发病率和死亡率逐年下降。根据《日本胃癌治疗指南(2018版)》 ‎[11] ,EGC可以通过EMR或ESD治疗。EMR主要适用于无溃疡且粘膜直径 ≤ 2 cm的粘膜癌。与EMR相比,ESD不受肿瘤大小或溃疡的限制,这有利于治愈性肿瘤切除 ‎[7] 。内镜下根治性切除的标准 ‎[12] :整体切除、肠分化型、病理T1a、肿瘤大小 ≤ 2 cm、手术切缘阴性(侧向和垂直)以及无淋巴血管侵犯。

LNM与EGC患者的不良预后有明显的相关性 ‎[13] 。LNM病的存在与否决定了治疗的选择。准确预测EGC患者是否存在LNM有助于选择最佳治疗方案,这在临床治疗过程中非常重要。因此,构建EGC患者的预测模型可能有助于发现那些有LNM倾向的患者,并延长术后生存时间。

列线图是复杂数学公式的图形表示。医学列线图使用生物和临床变量,例如肿瘤分级和患者年龄,来图形化地描述统计预后模型,该模型为给定个体生成临床事件的概率,例如癌症复发或死亡。诺模图有两种主要的使用方式。一种是图示式的,其中每个变量被单独列出,相应的点数被分配给给定的变量大小。然后,所有变量的累积分数与结果量表相匹配。或者,该公式包含在基于计算机或智能手机的计算器中,其中输入特定变量并计算事件的可能性 ‎[14] 。在我们的研究中,确诊EGC的性别、年龄、种族、分化程度、T期和肿瘤大小、远处转移、原发部位的手术、淋巴结捡取数目、原发肿瘤数目是LNM的影响变量(P < 0.05)。这些变量被用来构建预测模型。

在分类数据中,肿瘤大小是最重要的影响因素。一项研究指出miR-145-5p能够诱导胃癌的分化并影响胃癌的LNM,小于4 cm的早期癌症具有非常低的LNM率,可以评估局部切除 ‎[15] 。其他研究表明,分化程度和浸润深度的肿瘤是LNM的独立危险因素 ‎[16] 。Sekiguchi等人 ‎[17] 报道,具有大直径、深度和组织学类型的肿瘤被证实是LNM的独立影响因素。

此外,诊断时的年龄、分化程度、T分期、远处转移、原发部位的手术、淋巴结捡取数目、原发肿瘤数目也是LNM的独立影响变量。Gurzu等人 ‎[18] 发现,在较年轻的胃癌患者中,VEGF的表达更活跃,这增加了胃癌中肿瘤侵袭和LMN的概率。Bao等人 ‎[19] 认为,MDM4表达的增加可能与相关,并导致胃癌的生存状态较差,特别是在年轻患者中。Park等人 ‎[20] 的研究表明,年轻患者的胃癌组织比老年患者更具侵袭性,年轻EGC患者的LNM率高于其他患者,这可能与其肿瘤的恶性潜能较高有关。根治性手术切除和淋巴结清扫适用于深度浸润的胃癌。

4. 结论

性别、年龄、种族、分化程度、T期和肿瘤大小、远处转移、原发部位的手术、淋巴结捡取数目、原发肿瘤数目是EGC患者LNM的独立危险因素。在此基础上,建立了预测模型,用于预测EGC患者发生LNM的概率。受试者工作特性曲线和校准曲线也证明了模型的可信度和说服力。

参考文献

NOTES

*第一作者。

#通讯作者。

参考文献

[1] Karimi, P., Islami, F., Anandasabapathy, S., et al. (2014) Gastric Cancer: Descriptive Epidemiology, Risk Factors, Screening, and Prevention. Cancer Epidemiology, Biomarkers & Prevention, 23, 700-713.
https://doi.org/10.1158/1055-9965.EPI-13-1057
[2] Murphy, G., Kamangar, F., Dawsey, S.M., et al. (2011) The Relationship between Serum Ghrelin and the Risk of Gastric and Esophagogastric Junctional Adenocarcinomas. JNCI: Journal of the National Cancer Institute, 103, 1123-1129.
https://doi.org/10.1093/jnci/djr194
[3] Ren, J.S., Kamangar, F., Qiao, Y.L., et al. (2009) Serum Pepsinogens and Risk of Gastric and Oesophageal Cancers in the General Population Nutrition Intervention Trial Cohort. Gut, 58, 636-642.
https://doi.org/10.1136/gut.2008.168641
[4] Degiuli, M., De Manzoni, G., Di Leo, A., et al. (2016) Gastric Cancer: Current Status of Lymph Node Dissection. World Journal of Gastroenterology, 22, 2875-2893.
https://doi.org/10.3748/wjg.v22.i10.2875
[5] Zhang, C., Zhao, S. and Wang, X. (2022) Prognostic Nomogram for Early Gastric Cancer after Surgery to Assist Decision-Making for Treatment with Adjuvant Chemotherapy. Frontiers in Pharmacology, 13, Article ID: 845313.
https://doi.org/10.3389/fphar.2022.845313
[6] Thrift, A.P. and El-Serag, H.B. (2020) Burden of Gastric Cancer. Clinical Gastroenterology and Hepatology, 18, 534-542.
https://doi.org/10.1016/j.cgh.2019.07.045
[7] Facciorusso, A., Antonino, M., Di Maso, M., et al. (2014) Endoscopic Submucosal Dissection vs Endoscopic Mucosal Resection for Early Gastric Cancer: A Meta-Analysis. World Journal of Gastrointestinal Endoscopy, 6, 555-563.
https://doi.org/10.4253/wjge.v6.i11.555
[8] Marrelli, D., Polom, K., De Manzoni, G., et al. (2015) Multimodal Treatment of Gastric Cancer in the West: Where Are We Going? World Journal of Gastroenterology, 21, 7954-7969.
https://doi.org/10.3748/wjg.v21.i26.7954
[9] Bao, J., Nanding, A., Song, H., et al. (2016) The Overexpression of MDM4: An Effective and Novel Predictor of Gastric Adenocarcinoma Lymph Node Metastasis. Oncotarget, 7, 67212-67222.
https://doi.org/10.18632/oncotarget.11971
[10] Isozaki, H., Matsumoto, S. and Murakami, S. (2019) Survival Outcomes after Sentinel Node Navigation Surgery for Early Gastric Cancer. Annals of Gastroenterological Surgery, 3, 552-560.
https://doi.org/10.1002/ags3.12280
[11] Japanese Gastric Cancer A (2021) Japanese Gastric Cancer Treatment Guidelines 2018 (5th Edition). Gastric Cancer, 24, 1-21.
https://doi.org/10.1007/s10120-020-01042-y
[12] Conti, C.B., Agnesi, S., Scaravaglio, M., et al. (2023) Early Gastric Cancer: Update on Prevention, Diagnosis and Treatment. International Journal of Environmental Research and Public Health, 20, Article No. 2149.
https://doi.org/10.3390/ijerph20032149
[13] Zhao, B.W., Chen, Y.M., Jiang, S.S., et al. (2015) Lymph Node Metastasis, a Unique Independent Prognostic Factor in Early Gastric Cancer. PLOS ONE, 10, e0129531.
https://doi.org/10.1371/journal.pone.0129531
[14] Wang, Z.K., Lin, J.X., Li, P., et al. (2019) Higher Risk of Lymph Node Metastasis in Young Patients with Early Gastric Cancer. Journal of Cancer, 10, 4389-4396.
https://doi.org/10.7150/jca.30260
[15] Xiang, R., Rong, Y., Ge, Y., et al. (2021) Cell Differentiation Trajectory Predicts Patient Potential Immunotherapy Response and Prognosis in Gastric Cancer. Aging (Albany NY), 13, 5928-5945.
https://doi.org/10.18632/aging.202515
[16] Feng, F., Liu, J., Wang, F., et al. (2018) Prognostic Value of Differentiation Status in Gastric Cancer. BMC Cancer, 18, Article No. 865.
https://doi.org/10.1186/s12885-018-4780-0
[17] Sekiguchi, M., Oda, I., Taniguchi, H., et al. (2016) Risk Stratification and Predictive Risk-Scoring Model for Lymph Node Metastasis in Early Gastric Cancer. Journal of Gastroenterology, 51, 961-970.
https://doi.org/10.1007/s00535-016-1180-6
[18] Gurzu, S., Kadar, Z., Sugimura, H., et al. (2015) Gastric Cancer in Young vs Old Romanian Patients: Immunoprofile with Emphasis on Maspin and Mena Protein Reactivity. APMIS, 123, 223-233.
https://doi.org/10.1111/apm.12347
[19] Jiang, X.C., Yao, X.., Xia, H.B., et al. (2023) Nomogram Established Using Risk Factors of Early Gastric Cancer for Predicting the Lymph Node Metastasis. World Journal of Gastrointestinal Oncology, 15, 665-676.
https://doi.org/10.4251/wjgo.v15.i4.665
[20] Park, J.C., Lee, Y.C., Kim, J.H., et al. (2009) Clinicopathological Aspects and Prognostic Value with Respect to Age: An Analysis of 3,362 Consecutive Gastric Cancer Patients. Journal of Surgical Oncology, 99, 395-401.
https://doi.org/10.1002/jso.21281