Objective:

To explore the value of magnetic resonance imaging (MRI) based texture analysis in predicting the expression of vascular endothelial growth factor (VEGF) in extrahepatic cholangiocarcinoma (ECC).

Methods:

Seventy patients with ECC confirmed by pathology were retrospectively analyzed. The texture features w

ere extracted from MRI images using Mazda software. After texture feature selection

binary logistic regression analysis was performed. Receiver operating characteristic (ROC) curve was established to evaluate the prediction efficiency. The correlation between apparent diffusion coefficient (ADC) and VEGF expression was analyzed.

Results:

Four texture features met the standard

and logistic regression analysis was performed to obtain the prediction probability of VEGF expression in each patient. The area under curve (AUC) of prediction probability was 0.708 (

P

=0.003). The sensitivity

specificity and accuracy were 65.6%

78.9% and 72.8%

respectively. The ADC values of VEGF negative group and VEGF positive group were 1 368.210

- 6

mm

2

/s and 1 292.210

- 6

mm

2

/s

respectively. There was no significant difference between the two groups (

P

=0.398).

Conclusion:

MRI based texture analysis is helpful to predict the expression of VEGF in ECC.