Charlson合并症指数对结直肠癌手术患者预后的预测价值

Predictive value of Charlson comorbidity index in the operative prognosis of colorectal cancer

目的探讨Charlson合并症指数(CCI)对结直肠癌手术患者预后的预测价值。方法采用回顾性队列研究方法。收集2013年1月至2019年2月西安交通大学第一附属医院收治的1337例行手术治疗的结直肠癌患者临床病理资料;男774例,女563例;年龄为62(22~80)岁。所有患者进行CCI评估。观察指标:(1)结直肠癌手术患者的临床病理特征。(2)随访及生存情况。(3)影响结直肠癌手术患者预后的相关因素分析。(4)基于CCI列线图预测模型的构建与评价。采用电话或门诊方式进行随访,了解患者生存情况。随访时间截至2020年3月。正态分布的计量资料以x±s表示,组间比较采用t检验。偏态分布的计量资料以M(范围)或M(Q1,Q3)表示,组间比较采用Mann-Whitney U检验。计数资料以绝对数或百分比表示,组间比较采用χ^(2)检验。等级资料比较采用非参数秩和检验。采用Kaplan-Meier法计算生存率并绘制生存曲线,采用Log-Rank检验进行生存分析。采用COX比例风险模型进行单因素和多因素分析。将独立危险因素引入R 4.0.4软件进行列线图建模,并绘制受试者工作特征曲线(ROC),以曲线下面积(AUC)评价列线图预测模型的区分度。计算一致性指数并绘制校准曲线图评价列线图预测模型的一致性。结果(1)结直肠癌手术患者的临床病理特征。1337例患者中,1041例CCI≤3,296例CCI≥4。1041例CCI≤3患者的年龄,吸烟史(无、已戒烟、吸烟),R0切除(否、是),肿瘤分化程度(低分化、中分化、高分化),临床TNM分期(Ⅰ~Ⅱ期、Ⅲ~Ⅳ期),术前癌胚抗原分别为61(53,68)岁,717、43、281例,12、1029例,123、859、59例,666、375例,3.22(1.84,7.75)μg/L;296例CCI≥4患者的上述指标分别为70(61,75)岁,217、19、60例,43、253例,48、237、11例,102、194例,5.55(2.43,17.64)μg/L,两者上述指标比较,差异均有统计学意义(Z=-10.50,χ^(2)=7.34、104.51,Z=-2.31,χ^(2)=82.14,Z=-5.78,P<0.05)。(2)随访及生存情况。1337例患者均获得随访,随访时间为31(1~84)个月。1337例患者中,1024例生存,313例死亡。1041例CCI≤3和296例CCI≥4患者的1、3、5年总生存率分别为94.8%、85.5%、80.1%和73.6%、46.9%、34.0%,两者生存情况比较,差异有统计学意义(χ^(2)=181.93,P<0.05)。(3)影响结直肠癌手术患者预后的相关因素分析。单因素分析结果显示:年龄,吸烟史(吸烟),肿瘤位置(降结肠-乙状结肠、直肠乙状结肠交界-直肠),R0切除,肿瘤分化程度(中分化、高分化),临床TNM分期,术后放化疗,术前癌胚抗原,CCI是影响结直肠癌手术患者预后的相关因素[优势比=1.76,0.71,0.72、0.61,0.08,0.39、0.13,3.02,0.60,2.41,4.96,95%可信区间(CI)为1.39~2.23,0.53~0.93,0.52~0.99、0.47~0.78,0.06~0.11,0.30~0.50、0.05~0.31,2.39~3.81,0.48~0.76,1.92~3.01,3.97~6.20,P<0.05]。多因素分析结果显示:年龄>60岁,临床TNM分期为Ⅲ~Ⅳ期,术前癌胚抗原>5μg/L,CCI≥4是影响结直肠癌手术患者预后的独立危险因素(优势比=1.29,1.88,1.77,2.84,95%CI为1.00~1.65,1.45~2.44,1.40~2.23,2.20~3.67,P<0.05);肿瘤位置(降结肠-乙状结肠、直肠乙状结肠交界-直肠),R0切除,肿瘤分化程度(中分化、高分化),术后放化疗是影响结直肠癌手术患者预后的独立保护因素(优势比=0.71、0.72,0.27,0.50、0.25,0.56,95%CI为0.51~0.98、0.56~0.93,0.19~0.37,0.38~0.65、0.10~0.62,0.44~0.70,P<0.05)。(4)基于CCI列线图预测模型的构建与评价。根据多因素分析结果,纳入年龄、肿瘤位置、R0切除、肿瘤分化程度、临床TNM分期、术后放化疗、术前癌胚抗原、CCI,构建预测结直肠癌患者术后生存列线图模型。年龄>60岁评分为1.0分,肿瘤位于近端结肠为18.0分、远端结肠为9.0分,未达到R0切除为53.0分,肿瘤低分化为62.0分、中分化为31.0分,临床TNM分期为Ⅲ~Ⅳ期为32.0分,术后未行放化疗为26.0分,术前癌胚抗原每升高100μg/L为4.6分,CCI每增加1为12.6分。根据每项评分,加和得总分后评估患者1、3、5年总生存率。绘制ROC曲线评价列线图模型的预后预测能力,AUC为0.75(95%CI为0.71~0.79,P<0.05)。列线图模型的一致性指数为0.80(95%CI为0.77~0.82)。校正曲线显示该列线图模型的预测生存率和实际生存率有较好一致性。结论年龄>60岁,临床TNM分期为Ⅲ~Ⅳ期,术前癌胚抗原>5μg/L,CCI≥4是影响结直肠癌手术患者预后的独立危险因素;肿瘤位置(降结肠-乙状结肠、直肠乙状结肠交界-直肠),R0切除,肿瘤分化程度(中分化、高分化),术后放化疗是影响结直肠癌手术患者预后的独立保护因素。以此构建的列线图预测模型可用于预测结直肠癌手术患者生存状况。

Objective To investigate the predictive value of Charlson comorbidity index(CCI)in the operative prognosis of colorectal cancer(CRC).Methods The retrospective cohort study was conducted.The clinicopathological data of 1337 CRC patients who underwent surgery in the First Affiliated Hospital of Xi'an Jiaotong University from January 2013 to February 2019 were collected.There were 774 males and 563 females,aged 62(range,22‒80)years.All patients were evaluated by CCI.Observation indicators:(1)clinicopathological characteristics of CRC patients undergoing operation;(2)follow-up and survival;(3)prognostic factors analysis of CRC patients undergoing operation;(4)establishment and evaluation of a nomogram prediction model based on CCI.Follow-up was conducted using the telephone interview or outpatient examination to detect the survival of patients up to March 2020.Measurement data with normal distribution were represented as Mean±SD,and comparison between groups was analyzed using the t test.Measurement data with skewed distribution were described as M(range)or M(Q1,Q3),and comparison between groups was analyzed using the Mann-Whitney U test.Count data were described as absolute numbers or percentages,and comparison between groups was analyzed using the chi-square test.Non-para-meter rank sum test was used for comparison of ordinal data.The Kaplan‐Meier method was used to calculate survival rates and draw survival curves,and Log-Rank test was used for survival analysis.Univariate and multivariate analyses were performed using the COX proportional hazard regression model.The independent risk factors were included into R4.0.4 software to construct a nomogram prediction model.The receiver operating characteristic(ROC)curve was drawn,and the area under curve(AUC)was used to evaluate discrimination of the nomogram prediction model.The C-index and calibration chart were used to evaluate consistency of the nomogram prediction model.Results(1)Clinicopathological characteristics of CRC patients undergoing operation.Of the 1337 patients,there were 1041 cases with CCI≤3 and 296 cases with CCI≥4.Age,cases with non-smoking history,smoking cessation or smoking history,cases without or with R0 resection,cases with low,moderate,well differentiated tumor,cases in stageⅠ‒ⅡorⅢ‒Ⅳof clinical TNM staging,preoperative carcinoembryonic antigen(CEA)were 61(53,68)years,717,43,281,12,1029,123,859,59,666,375,3.22(1.84,7.75)μg/L for the 1041 patients with CCI≤3,versus 70(61,75)years,217,19,60,43,253,48,237,11,102,194,5.55(2.43,17.64)μg/L for the 296 patients with CCI≥4,showing significant differences in the above indicators between them(Z=‒10.50,χ^(2)=7.34,104.51,Z=‒2.31,χ^(2)=82.14,Z=‒5.78,P<0.05).(2)Follow-up and survival.All the 1337 patients were followed up for 31(range,1‒84)months.Of the 1337 patients,1024 cases survived and 313 cases died.The 1-,3-,5-year survival rates were 94.8%,85.5%,80.1%for the 1041 patients with CCI≤3,versus 73.6%,46.9%,34.0%for the 296 patients with CCI≥4,showing significant differences between them(χ^(2)=181.93,P<0.05).(3)Prognostic factors analysis of CRC patients undergoing operation.Results of univariate analysis showed that age,smoking history(having a history of smoking),tumor location(decending colon-sigmoid colon,recto-sigmoid junction-rectum),R0 resection,tumor differentiation degree(moderate differentiation,well differentiation),clinical TNM staging,postoperative radio-therapy and chemotherapy,preoperational CEA and CCI were related factors for operative prognosis of CRC patients(odds ratios=1.76,0.71,0.72,0.61,0.08,0.39,0.13,3.02,0.60,2.41,4.96,95%confidence intervals as 1.39‒2.23,0.53‒0.93,0.52‒0.99,0.47‒0.78,0.06‒0.11,0.30‒0.50,0.05‒0.31,2.39‒3.81,0.48‒0.76,1.92‒3.01,3.97‒6.20,P<0.05).Results of multivariate analysis showed that age>60 years,clinical TNM staging as stageⅢ‒Ⅳ,preoperational CEA>5μg/L and CCI≥4 were independent risk factors for operative prognosis of CRC patients(odds ratios=1.29,1.88,1.77,2.84,95%confidence intervals as 1.00‒1.65,1.45‒2.44,1.40‒2.23,2.20‒3.67,P<0.05);tumor located in descending colon to sigmoid colon and recto-sigmoid junction to rectum,R0 resection,tumor differen-tiation degree as moderate and well differentiation,postoperative radiotherapy and chemotherapy were independent protect factors for operative prognosis of CRC patients(odds ratios=0.71,0.72,0.27,0.50,0.25,0.56,95%confidence intervals as 0.51‒0.98,0.56‒0.93,0.19‒0.37,0.38‒0.65,0.10‒0.62,0.44‒0.70,P<0.05)(4)Establishment and evaluation of a nomogram prediction model based on CCI.Based on age,tumor location,R0 resection,tumor differentiation degree,clinical TNM staging,postoperative radiotherapy and chemotherapy,preoperational CEA and CCI of multivariate analysis results,a nomogram prediction model for operative prognosis of CRC patients was established.The nomogram score was 1.0 for age>60 years,18.0 for tumor located in proximal colon,9.0 for tumor located in distal colon,53.0 for non-R0 resection,62.0 for low differentiated tumor,31.0 for morderate differentiated tumor,32.0 for stageⅢ‒Ⅳof clinical TNM staging,26.0 for no postoperative radiotherapy and chemotherapy,4.6 for each increase of 100μg/L in preoperative CEA and 12.6 for each increase of 1 score in CCI respectively.The total of different scores for risk factors was used to evaluate total 1,3,5-year survival rates.The ROC curve was drawn to evaluate the predictive ability for prognosis of nomogram model,with the AUC as 0.75(95%confidence interval as 0.71‒0.79,P<0.05).The C-index was 0.80(95%confidence interval as 0.77‒0.72).The calibration chart showed a good consistency between the probability of survival predicted by nomogram and the actual probability of survival.Conclusions Age>60 years,stageⅢ‒Ⅳof clinical TNM staging,preoperational CEA>5μg/L and CCI≥4 are independent risk factors for operative prognosis of CRC patients.Tumor located in descending colon to sigmoid colon and recto-sigmoid junction to rectum,R0 resection,tumor differentiation degree as moderate and well differentiation,postoperative radiotherapy and chemotherapy are independent protective factors for operative prognosis of CRC patients.The nomogram prediction model contributes to prediction of the survival of CRC patients.