基于T2加权和弥散加权成像的影像组学定量评估膀胱癌肌层浸润

目的探讨联合T2加权成像(T2WI)和弥散加权成像(DWI)的影像组学研究在评估膀胱癌肌层浸润的实用性。方法收集2016年8月至2020年8月符合纳入及排除标准的病例共165例,非肌层浸润性膀胱癌组94例,肌层浸润性膀胱癌组71例。手动对肿瘤区域进行勾画,在组学模块界面上进行特征提取,运用LASSO方法筛选有意义特征,并计算影像组学分数(Radscore),构建比较单个序列和两序列联合模型,后选择最优的Radscore和有统计学差异的其他特征,运用逻辑回归的方法构建诺莫图(Nomogram)模型,通过曲线下面积(AUC)值判断各自模型的诊断效能。结果最后选用两序列联合的Radscore联合平均表观弥散系数(ADC)值、有无膀胱刺激症共同构建Nomogram模型,发现该模型较单独的Radscore模型和其他特征模型有更好的性能。模型在训练集的AUC、特异度及敏感度分别是0.877、0.864、0.806;在测试集中的AUC、特异度及敏感度分别是0.87、0.811、0.761。结论基于序列联合的Radscore和其他特征构建的Nomogram模型,对膀胱癌肌层浸润具有较好的定量诊断价值,进一步指导临床。

Objective To explore the applicability of Radiomics study combined with T2-weighted imaging(T2WI)and diffusion weighted imaging(DWI)in evaluating bladder cancer myometrial invasion.Methods A total of 165 cases that met the inclusion and exclusion criteria were collected from August 2016 to August 2020,including 94 cases in the non-muscle invasive bladder cancer group and 71 cases in the muscle invasive bladder cancer group.The tumor area was manually delineated,feature extraction was performed on the radiomic interface,the LASSO method was used to screen meaningful features,and the Radscore was calculated,and a single sequence and a two-sequence combined model were constructed.Finally,selecting the best Radscore and other features with statistical differences,the Nomogram model was constructed using the logistic regression method,and the diagnostic efficacy of each model was judged by the AUC value.Results The Radscore combined with the average ADC value and bladder irritation sign were ultimately selected to construct the Nomogram model.It was found that the model has better performance than the Radscore or other feature models alone.The AUC,specificity,and sensitivity of the model in the training set are 0.877,0.864,and 0.806,respectively.The AUC,specificity,and sensitivity of the model in the testing set are 0.87,0.811,and 0.761,respectively.Conclusion The Nomogram model constructed based on the combined Radscore and other features has a good quantitative diagnostic value for bladder cancer myometrial invasion,and further guides the clinic.