Abstract:
Objective: Low grade and atypical/anaplastic (high grade) meningiomas may look similar in conventional MRI. Since the surgery or treatment plan, tumor recurrence and treatment results of these two tumors are different from each other, it is important to differentiate between them. The objective of this study is to determine whether the combination of MR diffusion [Diffusion tensor imaging (DTI)– fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD) and diffusion weighted imaging (DWI) –apparent diffusion coefficient (ADC)] parameters are helpful in differentiating low grade meningiomas from high grade meningiomas.
Method: 45 pathologically proven meningioma patients [32 low grade, 13 high grade (11 atypical, 2 anaplastic)] who had pre-surgery conventional MRI, DWI and DTI were assessed retrospectively. ADCmin, ADCmean, ADCmax, FA, ADmean and RDmean were measured from contrasted tumor. Receiver operating characteristic (ROC) analysis was used in finding out the optimum cutoff values of all parameters to differentiate between low grade and high grade meningiomas. Combined ROC curves were measured for the combination of parameters. Combined ROC curves and area under curve (AUC) for each parameter were calculated and the best combination was determined in differentiation.
Results: While the ADCmin, ADCmax, ADCmean, ADmean and RDmean values of high grade meningiomas (p=0.007,p=0.045,p= 0.035, p= 0.045, p= 0.003, respectively) were significantly lower when compared with low grade meningioma values, FA values were significantly higher (p=0.007). According to ROC curve analysis, in differentiating these two groups, for ADCmin, ADCmax, ADCmean, RDmean and FA, cut off values were 0.782, 0.914, 0.822, 0.745, 0.239, respectively; while sensitivity, specificity and AUC values were 69%, 72%, 0.732; 69%, 62%, 0.696; 69%, 66%, 0.702; 69%, 78%, 0.770; 69%, 69%, 0.778. The best combination in differentiating low grade meningiomas from high grade meningiomas was formed by ADCmin, RDmean and FA parameters and sensitivity, specificity and AUC values were 92.3%, 100%, 0.962, respectively.
Conclusion: This study showed that combined MR diffusion parameters of ADCmin, RDmean and FA could differentiate high grade meningiomas from low grade meningiomas with a diagnostic accuracy of 96.2%. Thus, combination of MR diffusion parameters can provide more accurate diagnostic information in differentiating these two tumors.
