Impact Of 3-Tesla Mr Spectroscopic Imaging In The Delineation Of High Grade Glioma Target Volumes For Radiation Therapy Planning

The MR spectroscopic imaging (MRSI) derived tumor metabolic information offers reliable measures of tumor activity. The purpose of this study is to investigate the impact of incorporating 3T MRSI data in the delineation of target volumes for RT planning of high-grade gliomas. We retrospectively reviewed 3D MRSI data from 15 patients. After excluding 6 data sets due to post-surgical susceptibility artifacts in the tumor bed, the remaining 9 patients with high-grade glioma (HGG) were included in the final analysis. All MR scans were performed using 3T (Siemens TRIO TIM) scanner in our department. A 5 mm thick T2 weighted 2D turbo spin echo images in three orthogonal planes were obtained with TE/TR: 78/4,360 ms for MRSI reference. The 3D MRSI scans using PRESS (point resolved spectroscopy) sequence were obtained with TE/TR: 135/1,510 ms, 2 NEX, 12 x 12 x 8 matrix, 16 x 16 x 8 FOV resulting in 10 x 10 x 10 mm3 interpolated resolution. The 3D post-contrast T1 MRI with 1.0 mm isotropic resolution was also obtained. The MR Spectra were analyzed using linear combination model. The Choline-to-N-Acetyl aspartate (Cho: NAA) ratio that define the presence and extent of tumor activity was assigned as Cho: NAA ≤1.0 normal, 1 ≤Cho: NAA <2 suspicious for tumor (Cho: NAA1), 2 ≤Cho: NAA <3 active tumor (Cho: NAA2), 3 ≤Cho: NAA <4 aggressive tumor (Cho: NAA3) and 4 ≤Cho: NAA most aggressive tumor (Cho: NAA4). Interpolated gray scale 2D Cho: NAA maps intrinsically registered to T2 MRI were generated and saved in DICOM format. The CT, MRI, and MRSI maps were transferred to Pinnacle treatment planning system for image registration and segmentation of the metabolically active tumor regions. The accuracy of this process was verified earlier using phantom in a separate study. Volumes of Cho: NAA2, Cho: NAA3, and Cho: NAA4 contours were computed and the extent of Cho: NAA spread within GTV and CTV were examined. The Cho: NAA2 volume was compared with radiation oncologists defined GTV and CTV. The MRSI derived Cho: NAA ratio maps revealed significant intra- and inter tumor variations in the degree of metabolic abnormality. The mean volumes of Cho: NAA2, Cho: NAA3 and Cho: NAA4 were 18.23, 7.89 and 4.35 cc, respectively. In contrast to the Cho: NAA2 volume, the radiation oncologists defined mean GTV and CTV volumes were 300% and 900%, respectively larger. Although Cho: NAA2 volume of 2 cases fall within CTV, other 7 cases had varying percentage of Cho: NAA2 volumes (5-10% n = 3, 20-40%, n = 3 and 88% n = 1) outside their CTV. The GTV does not include 3% (n = 1), 5-10% (n = 2), 30-60% (n = 3) and 90-100% (n = 3) of Cho: NAA2 volumes. Significant differences exist between physicians defined GTV, CTV, and the MRSI derived active tumor volumes. A prospective trial is needed to assess the value of MRSI-guided HGG target delineation in RT.