基于多模态MRI影像组学诊断子宫内膜癌病理分型的研究进展

Research Progress in Pathological Classification of Endometrial Carcinoma Based on Multimodal MRI Imageomics

Ⅰ、Ⅱ型子宫内膜癌(EC)患者的临床特征、病理特征、治疗方案及预后等方面存在明显差异,临床需对此进行准确诊断。多模态磁共振成像(MRI)影像组学可利用MRI平扫+动态增强扫描(DCE)模型、MRI平扫+弥散加权成像(DWI)模型、MRI平扫+DWI模型、MRI平扫+DCE+DWI模型等多种预测模型深入挖掘病灶的肌层浸润、恶性程度等方面信息,并将其转化为定量数据,利于临床对EC病理分型做出准确诊断。但检查设备型号及影像组学分析、建模、图像分割方法的不同会在一定程度上影响诊断结果的准确性,今后还需形成统一的操作规范。

There are obvious differences in clinical features,pathological features,treatment schemes and prognosis between typeⅠand typeⅡendometrial carcinoma(EC),which need to be accurately diagnosed in clinic.Multimodal magnetic resonance imaging(MRI)imageomics can use MRI plain scan+dynamic contrast-enhanced scan(DCE)model,MRI plain scan+diffusion weighted imaging(DWI)model,MRI plain scan+DCE+DWI model and other prediction models to deeply excavate the information about muscle invasion,malignant degree and other aspects of lesions,and convert them into quantitative data,which is helpful to making accurate diagnosis of EC pathological classification.However,different inspection equipment models and imageomics analysis,modeling and image segmentation methods will also affect the accuracy of diagnosis results to a certain extent.In the future,unified operation specifications still need to be formed.