DWI多参数模型在FIGO2018宫颈癌新分期的应用价值

The application of multi-exponential models of diffusion-weighted MRI in the 2018 FIGO cervical cancer staging system

目的:探讨DWI多指数模型定量参数在FIGO 2018宫颈癌新分期中的应用价值。方法:回顾性分析51例经手术病理证实且术前未经任何治疗的宫颈癌患者MR-DWI指数模型定量参数。所有患者均行MR常规序列和DWI序列扫描;由两位放射科诊断医生(间隔2周重复测量)根据2018 FIGO分期测量早期宫颈癌与晚期宫颈癌的DWI单指数、双指数及拉伸指数模型所有参数值(测量参数包括ADCstand、D、D*,f、DDC、α),并进行组间比较,采用受试者工作特征(ROC)曲线下面积评价各参数的诊断效能及最佳诊断阈值。结果:同一测量者前后两次测量各参数值组内相关系数结果表明ADCstand、D、f、DDC及α值的一致性均较好,D*值一致性中等。早期宫颈癌组ADCstand、D、DDC值低于晚期宫颈癌组,组间比较有统计学差异(P<0.05),D*、f及α值组间比较无统计学差异;早期组中ⅠB1期、ⅠB2期与晚期组中的ⅠB3期各参数值进行比较。ADCstand、D、DDC值具有统计学差异(P<0.05),D*、f及α值无统计学差异。定量参数D及ADCstand具有较高诊断效能,曲线下面积(AUC)分别为0.957、0.904,诊断阈值为0.87×10-3mm2/s、1.19×10-3mm2/s。结论 :根据FIGO 2018宫颈癌新分期系统,DWI多指数模型有助于区分早期宫颈癌与晚期宫颈癌,其中参数D和ADCstand具有较高的诊断效能,在宫颈癌分期及精准诊疗中具有一定参考价值。

Objective: To investigate the value of diffusion-weighted imaging(DWI) with multi-exponential model parameters in the 2018 FIGO cervical cancer staging system. Methods: Fifty-one cases of patients with pathologically confirmed cervical cancer without any treatment underwent routine MR and IVIM-DWI. Mono-exponential, bi-exponential and stretched exponential models(including ADCstand, D, D*, f, DDC, α) were fitted. Differences in fitted parameters were assessed between early cervical cancer and advanced cervical cancer, and the receiver operating characteristic(ROC) curve was used to evaluate the diagnostic efficacy. Results: The results of correlation coefficient of each parameter measured by the same radiologist showed that the consistency of ADCstand, D, f, DDC and α values was high, and the consistency of D* values was poor.ADCstand, D and DDC of early cervical cancer group were lower than those of advanced cervical cancer, there was statistical difference between the two groups(P<0.05);D*, f and α had no statistical difference between the two groups. Besides, the results also showed ADCstand, D and DDC had statistical difference between stages of ⅠB1, ⅠB2. ⅠB1, ⅠB2 in the early group and IB3 in the late group(P<0.05), D*, f and α had no statistical difference. The largest areas under ROC curve was the cervical cancer D(0.957) and the second was the cervical cancer ADCstand(0.904). D and ADCstand have high diagnostic efficiency, and diagnostic threshold is 0.87×10-3 mm2/s and 1.19×10-3 mm2/s respectively. Conclusion: According to FIGO 2018 cervical cancer new staging system, multi-exponential model parameters of DWI can effectively distinguish early cervical can-cer from advanced cervical cancer, among which parameter D and ADCstand have high diagnostic efficiency.