科研之友
微信
新浪微博
Facebook
分享链接
摘要
Purpose: To evaluate the utility of histogram analysis of monoexponential, biexponential, and stretched-exponential models to a dualistic model of epithelial ovarian cancer (EOC). @@@ Materials and Methods: Fifty-two patients with histopathologically proven EOC underwent preoperative magnetic resonance imaging (MRI) (including diffusion-weighted imaging [DWI] with 11 b-values) using a 3.0T system and were divided into two groups: types I and II. Apparent diffusion coefficient (ADC), true diffusion coefficient (D), pseudodiffusion coefficient (D*), perfusion fraction (f), distributed diffusion coefficient (DDC), and intravoxel water diffusion heterogeneity (alpha) histograms were obtained based on solid components of the entire tumor. The following metrics of each histogram were compared between two types: 1) mean; 2) median; 3) 10th percentile and 90th percentile. Conventional MRI morphological features were also recorded. @@@ Results: Significant morphological features for predicting EOC type were maximum diameter (P=0.007), texture of lesion (P=0.001), and peritoneal implants (P=0.001). For ADC, D, f, DDC, and alpha , all metrics were significantly lower in type II than type I (P < 0.05). Mean, median, 10th, and 90th percentile of D* were not significantly different (P=0.336, 0.154, 0.779, and 0.203, respectively). Most histogram metrics of ADC, D, and DDC had significantly higher area under the receiver operating characteristic curve values than those of f and alpha (P < 0.05) @@@ Conclusion: It is feasible to grade EOC by morphological features and three models with histogram analysis. ADC, D, and DDC have better performance than f and alpha ; f and alpha may provide additional information.
- 出版日期2017-12
- 单位北京大学
