DWI定量参数在直肠癌患者术后复发转移预测中的增益价值OA

Objective:Combine diffusion-weighted imaging(DWI)technology to construct a quantitative model for predicting postoperative recurrence and metastasis in rectal cancer patients,thereby improving the accuracy of risk prediction for postoperative recurrence and metastasis.Methods:A retrospective analysis was conducted on the clinical data of 205 rectal cancer patients who underwent radical total mesorectal excision at Nanchong Hospital Affiliated to Beijing Anzhen Hospital,Capital Medical University(Nanchong Central Hospital)from January 2021 to June 2022.All patients underwent diffusion-weighted imaging(DWI)with b-values of 50 s/mm2 and 800 s/mm2 to obtain the apparent diffusion coefficient(ADC)values.Based on the follow-up results within 3 years postoperatively,patients were divided into recurrence/metastasis group and non-recurrence/non-metastasis group.Logistic regression analysis was performed to identify influencing factors of postoperative recurrence and metastasis in rectal cancer patients.Two risk prediction models were established accordingly:a conventional model(excluding ADC values)and an improved model(including ADC values).The predictive performance of the models was evaluated using receiver operating characteristic(ROC)curves,calibration curves(combined with Hosmer-Lemeshow test),and decision curve analysis(DCA)to assess predictive efficacy,accuracy,and clinical benefits,respectively.Results:Among the 205 patients,71 cases(34.63%)experienced recurrence/metastasis within 3 years postoperatively.Logistic regression analysis identified clinical stage,differentiation degree,neural invasion,vascular invasion,serum CEA,and the DWI quantitative parameter ADC value as independent influencing factors of postoperative recurrence/metastasis in rectal cancer patients(P<0.05).The ROC curve demonstrated that the improved risk prediction model achieved an AUC of 0.921(95%CI:0.886～0.956),significantly higher than that of the conventional model(0.887,95%CI:0.843～0.932)(P<0.05).The calibration curve showed good agreement between observed and predicted values for the improved model,with a Hosmer-Lemeshow test P>0.05,indicating excellent model fit.Decision curve analysis(DCA)revealed that the improved model provided clinical net benefits in predicting postoperative recurrence/metastasis within the threshold probability range of 0.1～0.9.Conclusion:The preoperative DWI quantitative parameter ADC value was identified as an independent predictor of postoperative recurrence/metastasis in rectal cancer patients.The Nomogram model constructed based on ADC value and clinicopathological indicators demonstrated excellent predictive performance.Moreover,the incorporation of ADC value increased the AUC of the conventional clinical model from 0.887 to 0.921,delivering significant gain in predictive efficacy for risk assessment.