Comparison of DWI based on monoexponential, biexponential and stretched-exponential models in differentiating tumor recurrence and pseudoprogression of glioblastoma
10.13929/j.1003-3289.201903059
- Author:
Dan LIAO
1
Author Information
1. Department of Radiology, Guizhou Provincial People's Hospital
- Publication Type:Journal Article
- Keywords:
Glioblastoma;
Intravoxel incoherent motion;
local;
Magnetic resonance imaging;
Neoplasm recurrence;
Pseudoprogression
- From:
Chinese Journal of Medical Imaging Technology
2019;35(10):1450-1455
- CountryChina
- Language:Chinese
-
Abstract:
Objective: To investigate the value of parameters derived from monoexponential, biexponential and stretched exponential intravoxel incoherent motion (IVIM) DWI models in differentiating tumor recurrence and pseudoprogression in patients with glioblastoma. Methods: Totally 38 patients with pathologically confirmed glioblastoma after surgery, chemotherapy and radiation therapy were enrolled and divided into tumor recurrence group (n=20) and pseudoprogression group (n=18) according to follow-up or reoperation results. IVIM DWI was performed, and the parameters of monoexponential model (ADC), biexponential model (true molecular diffusion coefficient [D], pseudo-diffusion coefficient [D*], perfusion fraction [f]) and stretched exponential model (distributed diffusion coefficient [DDC], stretched index [α]) were measured and compared between the 2 groups. ROC curve was used to analyze the efficacy of these parameters in differentiating tumor progression and pseudoprogression. Results: ADC, D, DDC and α values were significantly lower in tumor recurrence group than those in pseudoprogression group, while D* and f were significantly higher than those in pseudoprogression group (all P<0.05). ROC curve analysis showed AUC of ADC, D, D*, f, DDC and α value for differentiating tumor recurrence and pseudoprogression was 0.908, 0.925, 0.804, 0.743, 0.901 and 0.961, respectively. Conclusion: Parameters derived from monoexponential, biexponential and stretched exponential IVIM models have great value in differentiating tumor recurrence from pseudoprogression in patients with glioblastoma, and α from stretched exponential model has higher efficacy.
