| [Background]Colorectal cancer(CRC)accounts for up to 10%of all cancer deaths diagnosed worldwide and ranks second among all cancers,posing a threat to human development that cannot be ignored.Radical surgery is the most effective treatment for CRC,and lymph nodes(LNs)removed during surgery is extremely important.NCCN guidelines recommend removal of at least 12 LNs,and to date,various studies have investigated the optimal number of LNs to be removed in patients without distant metastases during radical CRC surgery.Most studies have focused on NO stage CRC,and Studies on the optimal number of LNs removed in patients with different N stages of CRC were limited.In addition,insufficient LNs removed is also closely related to stage migration.Therefore,this study investigated the optimal number of LNs removed in patients with stage NO to stage N2b CRC;The relationship between inadequate lymph nodes removal and stage migration in T3N0M0 pMMR CRC patients without other high-risk factors(Poorly differentiated tumor,preoperative obstruction or perforation of the tumor site,positive or unknown cut margins,inadequate safe distance to cut margins,combined with vascular or nerve invasion)except the number of LNs removed was analyzed.[Objective]To explore the optimal number of LNs removal in patients with N0,N1a,N1b,N2a,and N2b stages of CRC,respectively,and to construct nomogram prediction models for five-year OS.To redefine the cut-off value for the number of LNs removed in T3N0M0 pMMR CRC patients without other high-risk factors except the number of LNs removed and to analyze the relationship between inadequate LNs removal and staging migration.[Method]A retrospective study was conducted to follow up 4159 patients who underwent radical CRC surgery at Xijing Hospital of Digestive Diseases from October 2010 to October 2020.Clinical data were collected,including:gender,age,number of LNs removed,tumor location,tumor size,degree of differentiation,TNM stage,vascular invasion,nerve invasion,and preoperative tumor marker levels.The optimal number of LNs removed in patients with N0,N1a,N1b,N2a and N2b stages was calculated using X-tile software,and independent prognostic factors of patients after radical CRC surgery at each stage were examined using Cox proportional risk regression to create nomogram prediction models for five-year OS.The performance of the model was evaluated using the C-index,the area under the ROC curve,and the calibration curve with 1000 repetitions of the bootstrap method.The bootstrap method was also used to calculate the C-index and the area under the ROC curve for 1000 repetitions for internal validation.A retrospective study was conducted to follow up 950 patients who underwent radical CRC surgery at Xijing Hospital of Digestive Diseases from October 2010 to October 2020.Clinical data were collected,including:gender,age,number of LNs removed,tumor location,tumor size and degree of differentiation,TNM stage,vascular invasion,nerve invasion and mismatch repair protein assay results.Cut-off values for the number of LNs removed in T3N0M0 pMMR CRC patients without other high-risk factors except the number of LNs removed were calculated using X-tile software,and patients were screened for independent prognostic factors using Cox proportional risk regression.The relationship between insufficient LNs removed and staging migration was analyzed in T3N0M0 pMMR CRC patients without other high-risk factors except the number of LNs removed by grouping according to the cut-off value of the number of LNs removed.[Result]1.2560 patients with stage NO CRC were included in the analysis.According to X-tile software,the optimal number of LNs removed is 15.Univariate Cox regression analysis showed prognostic factors included age,number of LNs removed,preoperative CEA,CA125,CA19-9,vascular invasion and T stage.Multivariate Cox regression analysis showed that independent prognostic factors included age(HR:2.072;95%CI:1.636-2.624;P<0.001),number of LNs removed(HR:0.755;95%CI:0.577-0.988;P=0.041),preoperative CEA(HR:1.671;95%CI:1.301-2.145;P<0.001),CA125(HR:3.983;95%CI:2.899-5.472;P<0.001),CA19-9(HR:2.354;95%CI:1.772-3.128;P<0.001)and vascular invasion(HR:2.041;95%CI:1.554-2.681;P<0.001).Based on the above independent predictors,a nomogram prediction model of five-year OS was constructed.The model’s five-year OS had a C-index of 0.736 and an area under the ROC curve of 0.752,and the internal validation five-year OS had a C-index of 0.736 and an area under the ROC curve of 0.754,showing good predictive power,and the calibration curve for the bootstrap method repeated 1000 times for 5-years OS shows good agreement.2.497 patients with stage Nla CRC were included in the analysis.According to X-tile software,the optimal number of lymph nodes removed is 16.Univariate Cox regression analysis showed prognostic factors including:gender,age,number of LNs removed,preoperative CEA,CA125,CA19-9 and T stage.Multifactorial Cox regression analysis showed that independent prognostic factors included:number of LNs removed(HR:0.641;95%CI:0.416-0.987;P=0.043),preoperative CEA(HR:1.792;95%CI:1.166-2.752;P=0.008),CA125(HR:4.119;95%CI:2.444-6.943;P<0.001)and T stage(T4 vs T1/2,HR:4.873;95%CI:1.95-12.179;P=0.001).Based on the above independent prognostic factors,a nomogram prediction model of five-year OS was constructed.Gender was not significantly different,but being in a Multifactorial Cox regression equation,exclusion would have lost model predictive performance,so gender was also included in the model.The model’s five-year OS had a C-index of 0.740 and an area under the ROC curve of 0.773,and the internal validation five-year OS had a C-index of 0.738 and an area under the ROC curve of 0.785,showing good predictive power,and the calibration curve for the bootstrap method repeated 1000 times for 5-years OS shows good agreement.3.506 patients with stage N1b CRC were included in the analysis.According to X-tile software,the optimal number of LNs removed is 16.Univariate Cox regression analysis showed prognostic factors including:age,number of LNs removed,preoperative CEA,CA125,CA19-9,degree of tumor differentiation and T stage.Multivariate Cox regression analysis showed that independent prognostic factors included:number of LNs removed(HR:0.506;95%CI:0.361-0.708;P<0.001),preoperative CA125(HR:5.259;95%CI:3.447-8.023;P<0.001),CA19-9(HR:2.532;95%CI:1.704-3.761;P<0.001)and T stage(T4 vs T1/2,HR:2.375;95%CI:1.189-4.744;P=0.014).Based on the above independent prognostic factors,a nomogram prediction model of five-year OS was constructed.The model’s five-year OS had a C-index of 0.739 and an area under the ROC curve of 0.780,and the internal validation five-year OS had a C-index of 0.745 and an area under the ROC curve of 0.775,showing good predictive power,and the calibration curve for the bootstrap method repeated 1000 times for 5-years OS shows good agreement.4.323 patients with stage N2a CRC were included in the analysis.According to X-tile software,the optimal number of LNs removed is 17.Univariate Cox regression analysis showed prognostic factors including:age,number of LNs removed,preoperative CEA,CA125,CA19-9 and T stage.Multivariate Cox regression analysis showed that independent prognostic factors included:age(HR:2.090;95%CI:1.407-3.105;P<0.001),number of LNs removed(HR:0.658;95%CI:0.444-0.974;P=0.036),preoperative CA125(HR:2.885;95%CI:1.759-4.733;P<0.001),CA19-9(HR:2.107;95%CI:1.346-3.299;P=0.001)and T stage(T4 vs T1/2,HR:3.745;95%CI:1.423-9.851;P=0.007).Based on the above independent prognostic factors,a nomogram prediction model of five-year OS was constructed.The model’s five-year OS had a C-index of 0.713 and an area under the ROC curve of 0.763,and the internal validation five-year OS had a C-index of 0.711 and an area under the ROC curve of 0.757,showing good predictive power,and the calibration curve for the bootstrap method repeated 1000 times for 5-years OS shows good agreement.5.273 patients with stage N2b CRC were included in the analysis.According to X-tile software,the optimal number of LNs removed is 20.Univariate Cox regression analysis showed prognostic factors including:age,number of LNs removed,preoperative CEA,CA125,vascular invasion,neural invasion and T stage.Multivariate Cox regression analysis showed that independent prognostic factors included:number of LNs removed(HR:0.629;95%CI:0.417-0.949;P=0.027),preoperative CEA(HR:2.131;95%CI:1.512-3.003;P<0.001),CA125(HR:4.434;95%CI:2.620-7.503;P<0.001),T stage(T4 vs T2,HR:5.287;95%CI:1.634-17.107;P=0.005)and vascular invasion(HR:2.964;95%CI:1.794-4.894;P<0.001).Based on the above independent prognostic factors,a nomogram prediction model of five-year OS was constructed.The model’s five-year OS had a C-index of 0.705 and an area under the ROC curve of 0.744,and the internal validation five-year OS had a C-index of 0.710 and an area under the ROC curve of 0.743,showing good predictive power,and the calibration curve for the bootstrap method repeated 1000 times for 5-years OS shows good agreement.6.221 T3N0M0 pMMR CRC patients without other high-risk factors except the number of LNs removed were included in the analysis,and the cut-off values of LNs removed was 15 according to X-tile software.The results of univariate Cox regression analysis showed prognostic factors including:age and number of LNs removed.Multivariate Cox regression analysis showed that the number of LNs removed(HR:0.408;95%CI:0.203-0.819;P=0.012)was an independent prognostic factors.Grouping by the cut-off value for the number of LNs removed,compared with survival in T3N1M0 patients.In T3N0M0 pMMR CRC patients without other high-risk factors except the number of LNs removed,a significant difference was found between inadequate LNs removed group and the adequate LNs removed group,no significant difference was found between the inadequate LNs removed group and the T3N1M0 group.A significant difference was found in adequate lymph nodes removed group and the T3N1M0 group,and stage migration was found to be associated with inadequate lymph nodes removal.[Conclusion]1.The optimal number of LNs removed in patients with stage N0,N1a,N1b,N2a and N2b CRC is 15,16,16,17 and 20,respectively.The number of LNs removed is an independent prognostic factor of each N stage CRC patients.2.The nomogram prediction model for 5-year OS of patients with stage N0,N1a,N1b,N2a and N2b CRC has good prediction performance and has clinical application value.3.T3N0M0 pMMR stage Ⅱ CRC patients without other high-risk factors except the number of LNs removed have a cut-off value of 15 for the number of LNs removed,and those with fewer than 15 LNs removed should be identified as stage Ⅲ. |
PDF Full Text Request