目的 探讨构建基于肝切除术后复发性肝细胞癌(recurrent hepatocellular carcinoma,rHCC)瘤周区域表观扩散系数(apparent diffusion coefficient,ADC)列线图模型预测局部进展的可行性.材料与方法 采用回顾性队列研究方法,收集川北医学院附属医院2021年1月至2022年12月诊断为肝切除术后rHCC患者的MRI资料及临床特征.基于Firevoxel软件测量rHCC瘤周肝组织ADC平均值(peritumor ADCmean,pADCmean)、最小值(peritumor ADCmin,pADCmin)、最大值(peritumor ADCmax,pADCmax)和肿瘤ADC平均值(tumor ADCmean,tADCmean)、最小值(tumor ADCmin,tADCmin)、最大值(tumor ADCmax,tADCmax),以及背景肝组织ADC平均值(background liver tissue ADCmean,bADCmean)、最小值(background liver tissue ADCmin,bADCmin)、最大值(background liver tissue ADCmax,bADCmax).计算瘤周与背景肝组织 ADCmean比值(ratio of pADCmean to bADCmean,RPB-ADCmean)、ADCmin 比 值(ratio of pADCmin to bADCmin,RPB-ADCmin)、ADCmax 比 值(ratio of pADCmax to bADCmax,RPB-ADCmax),以及肿瘤与背景肝组织ADCmean比值(ratio of tADCmean to bADCmean,RTB-ADCmean)、ADCmin比值(ratio of tADCmin to bADCmin,RTB-ADCmin)、ADCmax比值(ratio of tADCmax to bADCmax,RTB-ADCmax).采用Cox回归分析筛选独立危险因素,并构建列线图预测模型,绘制受试者工作特征(receiver operating characteristic,ROC)曲线和决策曲线分析(decision curve analysis,DCA),评价模型预测rHCC局部进展的价值.结果 本研究共纳入70例肝切除术后rHCC患者,随访证实肿瘤局部进展率为65.7%(46/70).多因素Cox回归分析显示,pADCmin、RPB-ADCmean和异常凝血酶原是肝切除术后rHCC局部进展的独立危险因素(P均<0.05).列线图模型预测肝切除术后rHCC在3个月内、6个月内局部进展的ROC曲线下面积分别为0.834、0.841,DCA显示模型有较好临床净收益.结论 pADCmin、RPB-ADCmean和异常凝血酶原是预测肝切除术后rHCC局部进展的独立危险因素,构建的列线图模型可直观地预测rHCC局部进展,且具有较好的效能和临床价值.
Development of a nomogram based on diffusion weighted imaging of peritumoral liver tissue to predict local progression of recurrent hepatocellular carcinoma after hepatectomy
Objective:To investigate feasibility of a nomogram model developed with apparent diffusion coefficient(ADC)of peritumoral liver tissue to predict local progression of recurrent hepatocellular carcinoma(rHCC)after hepatectomy.Materials and Methods:A retrospective cohort study was conducted by collecting MRI and clinical data of patients with diagnosed rHCC after hepatectomy at the Affiliated Hospital of North Sichuan Medical College from January 2021 to December 2022.Using Firevoxel software,the peritumor mean ADC(pADCmean),minimum ADC(pADCmin),and maximum ADC(pADCmax)values,as well as the tumor mean ADC(tADCmean),minimum ADC(tADCmin),and maximum(tADCmax)values were measured.The background liver tissue mean ADC(bADCmean),minimum ADC(bADCmin),and maximum ADC(bADCmax)values were also obtained.The ratios of pADCmean to bADCmean(RPB-ADCmean),ADCmin(RPB-ADCmin),and ADCmax(RPB-ADCmax)along with the ratios of tADCmean to bADCmean(RTB-ADCmean),ADCmin(RTB-ADCmin),and ADCmax(RTB-ADCmax)were calculated.Cox regression analysis was used to identify independent risk factors,and then a nomogram model was constructed to predict local progression of rHCC after hepatectomy.Receiver operating characteristic(ROC)curve and decision curve analysis(DCA)were employed to evaluate the predictive value of the model for prediction of local progression of rHCC.Results:A total of 70 patients with rHCC after hepatectomy were enrolled,and the local progression rate of rHCC was 65.7％(46/70)confirmed by follow-up.Multivariate Cox regression analysis revealed that RPB-ADCmean,pADCmin and vitamin K absence antagonist-Ⅱ were independent risk factors for local progression of rHCC after hepatectomy(all P<0.05).The area under the ROC curve of the nomogram model to predict local progression of rHCC within 3 months and within 6 months after hepatectomy was 0.834 and 0.841,respectively.DCA demonstrated a favorable clinical net benefit of the model.Conclusions:The pADCmin,RPB-ADCmean and vitamin K absence antagonist-Ⅱ can be independent risk factors associated with local progression of rHCC after hepatectomy,and the developed nomogram model can intuitively predict local progression of rHCC with good performance and net clinical benefit.
万方数据
维普
