| Objective:Analyzed the related factors of poor prognosis of gastrointestinal stromal tumors by collecting clinical and pathological data,and constructed a predictive nomograph model,and compared it with different prognosis models.Method:The clinicopathological data of 1190 patients with gastrointestinal stromal tumors who were surgically removed and confirmed by postoperative pathology and immunohistochemistry were retrospectively collected from January 2004 to August2022 in the First Hospital of Jilin University.Grouptest was used for comparison between measurement data groups,and~2 test was used for comparison between counting data groups.Kaplan-Meier method was used for survival analysis,and Log-rank test was used to compare the survival between groups.Survival analysis of postoperative imatinib use in patients with intermediate and high-risk GIST.A total of913 patients with complete follow-up were randomly divided into a modeling cohort(639 cases)and a validation cohort(274 cases)in a 7:3 ratio.The modeling cohort was analyzed by single and multiple factors,and the prognostic factors were selected to construct the nomograph model.ROC curve,C index,calibration curve and decision curve were used to evaluate the predictive ability,accuracy and clinical effectiveness of the nomogram model.And validated the model with the validation cohort.The Fletcher,AFIP and modified NIH risk assessment standards were used to evaluate the risk of 639 patients in the modeling cohort,and ROC curve was used to evaluate,respectively,to compare the ability of different models to predict recurrence riskResult:According to the modified NIH classification standard,459 patients were divided into middle and high risk patients for recurrence,and the time to take imatinib after operation and the relapse-free survival time were analyzed.Survival analysis was conducted on 133 patients in the middle-risk group and 326 patients in the high-risk group,respectively,and the duration of taking imatinib was divided into four groups within 12 months,13 to 36 months,37 to 60 months and above 60 months.The results showed that the RFS of 459 patients who took imatinib for less than 12 months had statistical significance compared with the other three groups(P<0.001).There was no statistical significance in the RFS of 133 middle-risk patients among all groups(P>0.05).The RFS of 326 high-risk patients who took imatinib for less than 12 months was statistically significant(P<0.001)compared with the other three groups.RFS of patients taking imatinib for 13 to 36 months were statistically significant(P<0.001 and P=0.033)compared with those taking imatinib for less than 12 months and those taking imatinib for more than 60 months,but were not statistically significant with those taking imatinib for 37 to 60 months(P=0.06).The RFS of patients taking imatinib for 37～60months and those taking imatinib for more than 60 months were not statistically significant(P=0.603).The univariate analysis of the modeling cohort showed that gender,tumor site,tumor size,mitotic count,Ki-67 proliferation index,CD34,DOG-1 and postoperative administration of imatinib were statistically significant(P<0.05).Multivariate analysis was performed for statistically significant factors.Finally,five independent correlated factors including gender,tumor site,size,mitotic image count and postoperative imatinib administration were included in the model.The AUC values of 1-year,3-year and 5-year RFS of the modeling cohort were0.871(95%CI:0.804～0.938),0.825(95%CI:0.771～0.878)and 0.838(95%CI:0.793～0.883),respectively.The AUC values of the validation cohort were0.829(95%CI:0.750-0.908),0.840(95%CI:0.755～0.925)and 0.790(95%CI:0.697～0.884),respectively.The C index of the modeling cohort was0.838(95%CI:0.801～0.875),and the C index of the validation cohort was 0.816(95%CI:0.755～0.877).The AUC values of the 1-year,3-year and 5-year RFS of the modeling cohort under the Fletcher classification criteria were 0.838(95%CI:0.799～0.878),0.788(95%CI:0.741～0.835)and 0.821(95%CI:0.779～0.862),respectively.The AUC values of the1-year,3-year and 5-year RFS of the modeling cohort under the AFIP classification standard were 0.785(95%CI:0.698～0.873),0.791(95%CI:0.741～0.841)and 0.824(95%CI:0.783～0.866),respectively.The AUC values of the 1-year,3-year and 5-year RFS of the modeling cohort under the improved NIH classification standard were 0.819(95%CI:0.799～0.840),0.804(95%CI:0.770～0.837)and 0.827(95%CI:0.791～0.862),respectively.The results showed that there was no statistical difference between the prediction of recurrence risk and the Fletcher grading standard at 1,3 and5 years after operation by the nomograph(P=0.386,P=0.192 and P=0.511),and there was no statistical difference between the prediction of recurrence risk and the AFIP grading standard at 1,3 and 5 years after operation by the nomograph(P=0.069,P=0.209 and P=0.591).There was no statistical difference between the prediction of recurrence risk at 1,3 and 5 years after operation by the nomograph and the modified NIH grading standard(P=0.141,P=0.396 and P=0.644).Conclusion:1.The use of adjuvant therapy drugs such as imatinib after surgery in patients with gastrointestinal stromal tumors can significantly improve the recurrence free survival of high-risk patients who take imatinib for less than 1 year and significantly reduce it.2.Gender,tumor location,tumor size,mitotic count,and postoperative use of adjuvant drugs are independent prognostic factors for patients with gastrointestinal stromal tumors3.The predictive column chart constructed in this study for patients with gastrointestinal stromal tumors has good predictive ability and provides valuable reference for evaluating prognosis.,Compared with the Fletcher,AFIP,and modified NIH risk assessment standards,personalized postoperative recurrence free survival probability can be accurately predicted. |
PDF Full Text Request