结直肠癌同时性肝肺转移的预后因素分析及预后预测模型构建

Prognostic factors and construction of prognostic prediction model for simultaneously diagnosed liver and lung metastases from colorectal cancer

背景与目的:结直肠癌(CRC)可发生远处器官转移,导致预后不良,其中以肝脏和肺转移最为常见。然而,相较于CRC单纯肺转移或肝转移,同时性肝肺转移(SLLMCRC)的报道甚少。因此,本研究探讨SLLMCRC患者的预后相关因素并构建预后预测模型,为诊疗方案的选择及疗效评估提供参考。方法:在SEER数据库中提取2010—2019年诊断为SLLMCRC患者的病理资料,通过纳入和排除标准,最终筛选出800例符合要求的患者。根据7∶3比例随机分成建模集(560例)和验证集(240例),利用Cox比例风险回归模型筛选出SLLMCRC患者总体生存(OS)的独立危险因素,利用Fine-Gray竞争风险模型筛选出SLLMCRC患者肿瘤特异性生存(CSS)的独立危险因素。根据独立危险因素,以此来构建预测SLLMCRC患者OS和CSS的列线图模型。分别采用一致性指数、受试者工作特征(ROC)曲线、校正曲线对构建模型进行可靠性验证。结果:建模集与验证集患者基线因素差异均无统计学意义(均P>0.05)。年龄(50~69岁,HR=1.39,95%CI=1.07~1.81;≥70岁,HR=1.94,95%CI=1.46~2.58)、原发灶手术(HR=0.67,95%CI=0.48~0.95)、CEA水平(HR=1.39,95%CI=1.04~1.87)、化疗(HR=0.22,95%CI=0.18~0.28)是SLLMCRC患者OS的独立影响因素(均P<0.05);SLLMCRC患者年龄越大的OS率越差,而行原发灶手术、CEA阴性、接受化疗则有更高的OS率。年龄(50~69岁,HR=1.05,95%CI=1.01~1.12;HR=1.17,95%CI=1.02~1.35)、区域淋巴结清扫数目(HR=0.67,95%CI=0.48~0.90)、化疗(HR=0.45,95%CI=0.34~0.61)是SLLMCRC患者CSS的独立影响因素(均P<0.05);SLLMCRC患者的年龄越大CSS率越低,而区域淋巴结清扫数目多、接受化疗则有更高的CSS率。对基于以上因素构建的列线图预测模型的验证结果显示,建模集1、2、3年OS的ROC值分别为0.643、0.587和0.591;验证集分别为0.631、0.623和0.628。建模集1、2、3年CSS的ROC值分别为0.607、0.610和0.681;验证集分别为0.624、0.618和0.624。OS和CSS的校准曲线相对接近理想45°参考线。结论:年龄、原发灶手术、CEA水平、区域淋巴结清扫数目、化疗情况是SLLMCRC患者的预后密切相关,放疗可能无法让SLLMCRC患者获益。构建的预测模型具有较高的准确性和可靠性,为临床医生关于诊疗方案的选择及评估疗效提供了证据支持。

Background and Aims:Colorectal cancer(CRC)can metastasize to distant organs,leading to poor prognosis,with liver and lung metastases being the most common.However,simultaneous diagnosed liver and lung metastases from colorectal cancer(SLLMCRC)are rarely reported compared to isolated lung or liver metastases.Therefore,this study was conducted to explore the prognostic factors for patients with SLLMCRC and to develop a prognostic prediction model to provide reference for the selection of treatment plans and evaluation of therapeutic efficacy.Methods:Data of patients diagnosed with SLLMCRC from 2010 to 2019 were extracted from the SEER database.After applying inclusion and exclusion criteria,800 eligible patients were selected.These were randomly divided into a modeling cohort(560 cases)and a validation cohort(240 cases)in a 7∶3 ratio.The Cox proportional hazards regression model was used to identify independent risk factors for overall survival(OS)of SLLMCRC patients,while the Fine-Gray competitive risk model was employed to identify independent risk factors for cancer-specific survival(CSS).Prognostic nomograms for predicting OS and CSS were constructed based on these independent risk factors.The reliability of the models was validated using the consistency index,ROC curve,and calibration curve.Results:There were no statistically significant differences in baseline factors between the modeling and validation cohorts(all P>0.05).Age(50-69 years:HR=1.39,95%CI=1.07-1.81;≥70 years:HR=1.94,95%CI=1.46-2.58),primary tumor resection(HR=0.67,95%CI=0.48-0.95),CEA level(HR=1.39,95%CI=1.04-1.87),and chemotherapy(HR=0.22,95%CI=0.18-0.28)were independent factors affecting SLLMCRC patients OS(all P<0.05).The older age of SLLMCRC patients,the lower the OS rate,while having primary tumor resection,negative CEA result,and receiving chemotherapy result in higher OS rate.Age between 50-69 years(HR=1.05,95%CI=1.01-1.12),number of regional lymph nodes removed(HR=0.67,95%CI=0.48-0.90),and chemotherapy(HR=0.45,95%CI=0.34-0.61)were independent factors affecting CSS(all P<0.05).Older age correlated with lower CSS rate,while more extensive regional lymph node removal and receiving chemotherapy correlated with higher CSS rate.The nomogram validation showed that the 1,2,and 3-year OS ROC values for the modeling cohort were 0.643,0.587,and 0.591,respectively;for the validation cohort,these values were 0.631,0.623,and 0.628.The 1,2,and 3-year CSS ROC values for the modeling cohort were 0.607,0.610,and 0.681,respectively;for the validation cohort,these values were 0.624,0.618,and 0.624.The calibration curves for OS and CSS were relatively close to the ideal 45°reference line.Conclusion:Age,primary tumor resection,CEA level,number of regional lymph nodes removed,and chemotherapy are closely related to the prognosis of patients with SLLMCRC.Radiotherapy may not benefit SLLMCRC patients.The constructed prediction model has high accuracy and reliability,providing evidence to support clinical decision-making and evaluation of treatment efficacy for clinicians.