Gliomas are the most common primary neoplasms of brain and show heterogeneous histological spectrum. An optimal treatment strategy depends on accurately determining the tumor grades ¹. Gliomas are graded according to World Health Organization classification from grade I to grade IV ². Currently gliomas are graded based on histopathological assessment of excised tumor. Various perfusion MRI techniques namely dynamic contrast enhanced (DCE), dynamic susceptibility contrast (DSC) and arterial spin labelling MRI are used in assessment of brain neoplasm. This study was aimed to assess ability of DCE perfusion metrics to discriminate various grades of glioma.

Data acquisition: A total of 93 glioma (21 grade II, 29 grade III, 43 grade IV) with a post-surgical diagnosis were included in this study after the approval of institutional ethics committee. All patients underwent preoperative imaging with both conventional and DCE-MRI on 3 Tesla scanner (Philips Healthcare, The Netherland) using a 32 channel head coil. DCE-MRI was performed using a three dimensional Turbo field echo sequence [TR/TE=4.4/2.1ms, 10o flip angle, 6mm slice thickness, 240×240mm² FOV, 128×128 matrix size, 3.9s temporal resolution, 32 dynamic, 12 slices]. Gd-DTPA was administered intravenously after 3 dynamics with the help of a power injector at a rate of 3ml/sec, followed by a bolus injection of 30ml saline flush. Prior to perfusion scan, fast spin echo (FSE) T1-weighted (TR/TE=360/10ms), PD weighted (TR/TE=3500/23ms) and T2 weighted (TR/TE=3500/90ms) imaging were performed for the same slice position to quantify voxel wise pre-contrast tissue relaxation time T₁₀.

MRI data processing and quantitative analysis: Voxel wise tissue T10 was quantified and use to extract contrast time curve from signal intensity time curve. First-pass analysis was performed to quantify of cerebral blood volume (CBV) and cerebral blood flow (CBF). The pharmacokinetic model was implemented for permeability (K^trans & K_ep) fractional extra cellular extra vascular EES volume (V_e), plasma volume (V_p) and leakage calculations. Corrected CBV maps were generated by removing the leakage effect of the disrupted blood brain barrier ³. ROIs (10-20 mm² ) were drawn on the region of tumor with the highest value of each perfusion metrics as seen by respective map of that metrics. Multiple ROIs were drawn on each metrics map selecting regions showing the best values of respective perfusion metrics. Relative quantification of CBV (rCBV) and CBF (rCBF) were quantified by placing the ROI on normal contra-lateral grey/white matter of the brain. One-way ANOVA , multiple comparison using Bonferroni test and ROC analysis was performed using SPSS 16 to discriminate between grades.

Among all the parameter used for one-way ANOVA only corrected rCBV significantly (p<0.001) discriminated the glioma grades . Corrected rCBV value for Grade II (1.66±0.65), grade III (3.74±1.21) and grade IV (4.50±1.32)). Result of Bonferroni test and ROC analysis is shown in Table 1 and Table 2 respectively. In this study CBF, CBV and tracer kinetic parameters derived from DCE MRI were used to distinguish among various grades of gliomas. it was found that leakage corrected rCBV was useful for differentiating grade II, grade III and grade IV gliomas. Previous perfusion imaging studies have shown that rCBV is a useful parameter for distinguishing between low and high grade gliomas which has been correlated with increased neoangiogenesis in these tumors ⁴. However majority of previous studies have used DSC perfusion MRI and have not been able to discriminate between grade III and grade IV tumors ⁵. In conclusion DCE derived leakage corrected rCBV is useful for discriminating various grades of glioma.