DCE-MRI参数及纹理特征在宫颈癌新辅助化疗疗效中的预测价值

The value of DCE-MRI parameters and texture features for predicting the efficacy of neoadjuvant chemotherapy in cervical cancer patients

目的:探讨动态增强磁共振成像(DCE-MRI)参数及纹理特征在宫颈癌新辅助化疗(NACT)疗效中的预测价值。方法:回顾性收集63例NACT的宫颈癌患者,根据实体瘤疗效评价标准将患者分为有效组(完全缓解和部分缓解)及无效组(疾病稳定和疾病进展),所有患者NACT治疗前一周行DCE-MRI扫描,并获得定量参数[最大斜率(Maxslope)、对比增强率(CER)、对比剂浓度时间曲线下面积(IAUGC)、转运常数(Ktrans)、速率常数(Kep)、血管外细胞间隙体积百分数(Ve值)]及纹理参数,比较上述参数在两组间的差异。采用受试者操作特征(ROC)曲线分析参数的诊断效能。结果:有效组的CER、Ktrans值均显著低于无效组(P=0.004、0.012),有效组Mean Value、Grey Level Nonuniformity显著低于无效组(P=0.003、0.002),Energy显著高于无效组(P=0.024)。CER、Ktrans评估NACT疗效的AUC值分别为0.712、0.695。纹理参数评估NACT疗效的AUC值在0.647~0.716之间。将有组间差异的DCE-MRI参数、纹理参数、临床参数联合后,预测NACT疗效的AUC为0.922。结论:DCE-MRI参数及纹理参数可作为潜在的预测宫颈癌NACT疗效的影像学标志物。

Objective:To explore the value of dynamic contrast-enhanced magnetic resonance imaging(DCE-MRI) parameters and texture features for predicting the efficacy of neoadjuvant chemotherapy(NACT) in cervical cancer patients.Methods:Sixty-three cervical cancer patients who underwent NACT were retrospectively included and divided into effective group(complete remission and partial remission) and ineffective group(disease stable and disease progression) according to the response evaluation criteria in solid tumors.All patients underwent DCE-MRI a week before NACT.Quantitative parameters(Maxslope,CER,IAUGC,Ktrans,Kep,Ve) and texture parameters were obtained,and compared between effective and ineffective group.Receiver operating characteristic(ROC) curve was used to deterermine the diagnostic performance.Results:The CER and Ktrans in effective group were significantly lower than those in ineffective group(P=0.004,0.012).The mean value and grey level nonuniformity of effective group were significantly lower than those of ineffective group(P=0.003,0.002).Energy of effective group was significantly higher than that in ineffective group(P=0.024).The AUC of CER and Ktrans was 0.712 and 0.695,respectively.The AUC of texture parameters varied between 0.647 to 0.716.The AUC of combining DCE-MRI,texture and clinical parameters to predict NACT efficacy was 0.922.Conclusion:DCE-MRI parameters and texture features may be used as potential imaging biomarkers for predicting NACT efficacy in cervical cancer patients.