Clinical Investigative Study Radiation Injury to the Normal Brain Measured by 3D-Echo-Planar Spectroscopic Imaging and Diffusion Tensor Imaging: Initial Experience Sanjeev Chawla, PhD, Sumei Wang, PhD, Sungheon Kim, PhD, Sulaiman Sheriff, MS, Peter Lee, MSE, Ramesh Rengan, MD, Alexander Lin, MD, Elias Melhem, MD, PhD, Andrew Maudsley, PhD, Harish Poptani, PhD From the Department of Radiology, University of Pennsylvania, Philadelphia, PA (SC, SW, PL, EM, HP); Department of Radiology, New York University, New York, NY (SK); University of Miami, Miami, FL (SS, AM); Department of Radiation Oncology, University of Pennsylvania, Philadelphia, PA (RR, AL)
ABSTRACT BACKGROUND AND PURPOSE
Whole brain radiation therapy (WBRT) may cause cognitive and neuropsychological impairment and hence objective assessment of adverse effects of radiation may be valuable to plan therapy. The purpose of our study was to determine the potential of echo planar spectroscopic imaging (EPSI) and diffusion tensor imaging (DTI) in detecting subacute radiation induced injury to the normal brain. MATERIALS AND METHODS
Four patients with brain metastases and three patients with lung cancer underwent cranial irradiation. These patients were subjected to 3D-EPSI and DTI at two time points (preradiation, and 1 month post-irradiation). Parametric maps of N-acetyl aspartate (NAA), creatine (Cr), choline (Cho), mean diffusivity (MD), and fractional anisotropy (FA) were generated and co-registered to post-contrast T1-weighted images. Normal appearing graymatter and white-matter regions were compared between the two time points to assess sub-acute effects of radiation using independent sample t-tests.
Keywords: Radiation injury, whole brain radiation therapy, prophylactic cranial irradiation, echo-planar spectroscopic imaging, diffusion tensor imaging. Acceptance: Received May 3, 2013, and in revised form July 8, 2013. Accepted for publication August 3, 2013. Correspondence: Address correspondence to Harish Poptani, PhD, Department of Radiology, University of Pennsylvania. E-mail: Harish.Poptani@ uphs.upenn.edu. J Neuroimaging 2013;00:1-8. DOI: 10.1111/jon.12070
RESULTS
Significantly increased MD (P = .02), Cho/Cr (P = .02) and a trend towards a decrease in NAA/Cr (P = .06) was observed from the hippocampus. Significant decrease in FA (P = .02) from the centrum-semiovale and a significant increase in MD (P = .04) and Cho/Cr (P = .02) from genu of corpus-callosum was also observed. CONCLUSIONS
Our preliminary findings suggest that 3D-EPSI and DTI may provide quantitative measures of radiation induced injury to the normal brain.
Introduction Whole brain radiation therapy (WBRT) is a standard palliative treatment option for patients with brain metastases and has been shown to improve local control.1 In patients with small cell lung cancer exhibiting complete clinical remission, the addition of prophylactic cranial irradiation (PCI) has been shown to improve overall survival.2 While WBRT and PCI have a therapeutic benefit, care is needed in delivering the radiation dose as there is a strong body of evidence suggesting that WBRT, may result in severe neuropsychological and cognitive impairment.3 The dentate subgranular zone of the hippocampus is involved in neurogenesis and multipotent progenitor cells located in this region are sensitive to radiation induced injury.4 Damage to the supratentorial white matter such as centrum semiovale (CS) and corpus callosum is also associated with impaired neurocognition.5 However, clinical symptoms usually occur several months after radiation therapy and till date, there
is no reliable neurobehavioral assay to measure the subacute effects of radiation induced injury.6 Therefore, there is a critical need for objective and quantitative methods to assess early treatment-induced damage due to WBRT to ascertain the riskto-benefit ratio and dose determination so that the radiation beam can be specifically tailored to limit damage to the normal brain. 1 H MRS has been used to demonstrate abnormal metabolite pattern from the normal brain, secondary to irradiation.7–12 However, these studies suffered from limited volumetric coverage of the brain since a majority of these studies either used a single voxel or a single slice two-dimensional chemical shift imaging method. The only study that used a 3D 1 H MRS method, reported increased averaged Cho from normal brain areas in patients with gliomas treated with radiation therapy,13 however, regional differences in the Cho levels were not reported in this study. Another study using a non-localized whole
Copyright
◦ 2013 by the American Society of Neuroimaging C
1
brain 1 H MRS approach14 reported a significant reduction in global N-acetyl aspartate (NAA) levels in patients receiving PCI. However, this method does not provide spatial information and hence is not suitable for assessing regional damage to the tissue. Recently, a 3D whole brain echo planar spectroscopic imaging (EPSI) method has been described which provides metabolite maps throughout the brain with excellent spatial resolution.15, 16 The 3D-EPSI technique has been used to ascertain neuronal damage in diffuse neurological disorders17–19 and in brain tumors.20 However, to the best of our knowledge, EPSI has not been used to evaluate radiation induced injury in the normal brain. Diffusion tensor imaging (DTI) is commonly used in the evaluation of brain tumors21 and to evaluate radiation injury to the normal brain in pre-clinical models22, 23 as well as in clinical studies.24–27 These studies reported alterations in mean diffusivity (MD) and fractional anisotropy (FA) from various regions of the normal brain. However, DTI has not been reported in evaluating radiation induced injury in the normal brain of patients with brain metastases. The adverse reactions of radiation can be described in three stages, acute, subacute, and late.28 While delayed radiation injury can be accurately assessed by conventional MR imaging, these methods are not sensitive in demonstrating early injury after irradiation.29 Therefore, we sought to investigate the utility of 3D-EPSI and DTI in detecting subacute radiation induced damage to the normal brain parenchyma in patients with brain metastases and in lung cancer patients receiving PCI.
Materials and Methods Subjects The study was approved by the Institutional Review Board and was compliant with the Health Insurance Portability and Accountability Act. Adult patients with brain metastases harboring at least one lesion greater than 1 cm3 were included in the study. Additionally, patients with lung cancer who were scheduled for PCI and had no prior history of cranial irradiation or cranial surgery were included. Patients previously treated with cranial irradiation were excluded from the study. The study comprised a cohort of 15 patients, 10 patients (mean age ± standard deviation = 65 ± 4.24 years, 5 males and 5 females) with brain metastases were irradiated with fractionated WBRT [total dose = 30-40 gray (Gy) in 10-20 fractions] and five patients (65 ± 15.71 years, 2 males and 3 females) with lung cancer underwent PCI (total dose = 25Gy in 10 fractions). All patients underwent a baseline (prior to irradiation) conventional MR imaging, 3D-EPSI and DTI and were scheduled for a followup study (1 month post-irradiation). However, eight of these patients were lost on follow-up. Of the remaining 7 patients, 4 had metastatic brain lesions [confirmed by histopathological analysis of resected brain specimens (n = 2) or by consistent features on multiple conventional MR imaging (n = 2)] and 3 patients with lung cancer did not harbor any brain lesions. The mean duration between the end of radiation therapy and follow-up imaging study was 30.43 ± 9.02 days for these 7 patients. The clinical characteristics and therapeutic regimens for these patients are presented in Table 1.
2
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
MR Imaging MR imaging was performed on a 3T Tim Trio MR scanner (Siemens Medical System, Erlangen, Germany) using a 12-channel phased array head coil. The imaging protocol included three-plane scout localizer, 3D-magnetization prepared rapid gradient echo (MPRAGE) [Repetition time (TR)/echo time (TE)/inversion time (TI) = 1620/3.9/950 ms)], 192 × 256 matrix, slice thickness = 1mm, 192 slices per slab, flip angle = 15◦ , number of excitations (NEX) = 1, bandwidth (BW) = 150 Hz/pixel); axial proton density weighted images (TR/TE = 4100/13 ms, NEX = 1, slices = 40, slice thickness = 3.0 mm, BW = 130 Hz, flip angle = 120◦ ); axial T2 weighted FLAIR image (TR/TE/TI = 9420/141/2500 ms), slice thickness = 3 mm, slices = 60, flip angle = 170◦ , NEX = 1, BW = 287 Hz/pixel; and contrast enhanced T1 weighted MPRAGE images after administration of standard dose (.1 mmol/kg) of contrast agent (Magnevist, gadopentetate dimeglumine, Bayer Healthcare Pharmaceuticals Inc, Wayne, NJ).
3D Echo-Planar Spectroscopic Imaging Scan parameters for the 3D spin-echo EPSI sequence were: TR/TE = 1710/70 ms, spatial points = 50 × 50 × 18, field of view (FOV) = 280 × 280 × 180 mm3 , voxel size = 5.6 × 5.6 × 10 mm3 (.31 cm3 ), excitation angle = 73◦ , 512 complex points, spectral BW = 2,500 Hz with radiofrequency excitation pulse centered at water resonance, NEX = 1. Water suppression using frequency-selective saturation pulses and inversion-recovery nulling of lipid signal was performed with TI of 198 ms. The acquisition time was 26 minutes including an interleaved water reference acquisition scan obtained using a gradient-echo acquisition with 20◦ excitation angle and a 6.3 ms TE. This water-unsuppressed image was used to perform signal normalization, eddy current correction, and image coregistration.
Diffusion Tensor Imaging DTI data were acquired using 30 non-collinear/non-coplanar directions with a single-shot spin-echo, echo-planar read-out sequence with parallel imaging by using generalized autocalibrating partially parallel acquisition (GRAPPA) and acceleration factor of 2. Sequence parameters were as follows: TR/TE = 5,000/86 ms, number of acquisitions = 3, field of view (FOV) = 22 cm × 22 cm, matrix size = 128 × 128, slice thickness = 3 mm, b = 0, 1,000 s/mm2 , slices = 40 covering the whole brain.
Image Processing 3D-EPSI data were processed offline using the metabolic imaging and data analysis system (MIDAS) package developed by Maudsley et al.16 The data were corrected for inhomogeneity in the B0 field, followed by eddy current correction and interpolation to a spatial resolution of 64 × 64 × 32 voxels. The processing also included steps to reduce ringing artifacts from subcutaneous lipids using a mask of the scalp region derived from the T1-weighted images. Signal intensity normalization of metabolite maps was performed using tissue water as an
Table 1. Clinical Characteristics and Therapeutic Regimens Patient ID
1 2 3 4 5 6 7
Gender/Age (Years)
Primary Site of Cancer
Radiation Type
Radiation Dose (Gy)/Fraction
Number of Fractions
Total Radiation Dose (Gy)
M/68 M/65 F/59 F/76 M/68 F/72 F/47
Laryngeal carcinoma and non-small cell lung cancer Thyroid carcinoma and non-small cell lung cancer Non-small cell lung cancer Small cell lung cancer Non-small cell lung cancer Non-small cell lung cancer Small cell lung cancer
WBRT WBRT WBRT PCI WBRT PCI PCI
3 2 2.5 2.5 2.5 2.5 2.5
10 20 14 10 14 10 10
30 40 35 25 35 25 25
WBRT = whole brain radiation therapy; PCI = prophylactic cranial irradiation; Gy = gray.
Fig 1. Gray and white matter regions analyzed to evaluate the effects of radiation therapy are shown as circular ROIs (20 pixels). Motor cortex (green), CS (black ring), genu (orange), and splenium (yellow) of corpus callosum, periventricular frontal white matter (black), periventricular occipital white matter (brown), thalamus (blue), basal ganglia (pink), cingulate gyrus (red), and hippocampus (purple). internal reference. The automated MIDAS tool was used to compute parametric maps of NAA, Cr, and Cho. The diffusion-weighted images were co-registered to the non-diffusion weighted (b = 0 s/mm2 ) images to minimize eddy-current and/or subject motion induced artifacts using a 3D affine transformation estimated by maximizing the mutual information between the images as described previously.21 The corrected raw images were combined to compute MD, and FA maps. Unsuppressed water images, corresponding 3D-EPSI maps (NAA, Cr, Cho), DTI maps (MD, FA) and FLAIR images were co-registered to contrast enhanced T1-weighted images using a 3D non-rigid transformation and mutual information by combining affine transformation and discrete sine bases. All image processing procedures were performed using in-house developed interactive data language (IDL) routines (ITT Visual Information Solutions, Boulder, CO).
Data Analysis To evaluate the effect of radiation therapy on normal brain parenchyma, circular bilateral regions of interests (ROIs), (20
pixels each) were drawn on normal appearing gray matter (hippocampus, thalamus, motor cortex, cingulate gyrus, and basal ganglia) and white matter [genu and splenium of corpus callosum, periventricular frontal and occipital white matter, and CS] regions (Fig 1). While placing ROIs, care was taken to avoid the metastatic lesions as well as peri-lesional edema. The ROIs were chosen from these areas as they are sensitive to radiation induced injury and are associated with cognitive, behavior, and motor activities.4, 5 The mean NAA, Cr, Cho, MD, and FA values were computed from each of the selected ROIs at baseline and post-irradiation periods. The NAA and Cho values were normalized with respect to Cr for each ROI to obtain the NAA/Cr and Cho/Cr ratios. To evaluate the subacute effects of radiation, average NAA/Cr, Cho/Cr, MD, and FA from all patients were compared between baseline and 1-month post-irradiation periods. The baseline 3D-EPSI data from one of the brain metastasis patients (patient no. 1, Table 1) was corrupted due to severe motion artifacts. As a result, DTI data were incorporated from 7 patients and 3D-EPSI from 6 patients in the final data analysis. Independent sample t-tests were performed with an assumption that significant differences in 3D-EPSI (NAA/Cr and
Chawla et al: Assessment of radiation injury to normal brain
3
Fig 2. Representative brain regions from a 47-year-old female patient with lung cancer receiving PCI are shown to illustrate the feasibility of high spatial resolution co-registered 3D-EPSI and DTI data. Co-registered post-contrast T1 weighted image (A and F), MD (B and G), FA (C and H), NAA (D and I), and Cho (E and J) maps through the supratentorial (upper-panel) and infratentorial (lower-panel) region. CS (open black ROI and white arrow) and hippocampus (open black ROI and black arrow) are shown as representative regions from supratentorial and infratentorial brain parenchyma, respectively. Cho/Cr) and DTI (MD and FA) values from baseline to postirradiation would reflect radiation induced injury. A P value < .05 was considered significant. To investigate the relationship between EPSI (NAA/Cr and Cho/Cr) and DTI parameters (MD and FA) within each anatomical region, Pearson’s correlation coefficient analyses were performed. Patients receiving PCI did not have any detectable brain lesions on diagnostic MR imaging, while patients with brain metastases had up to three lesions with surrounding perifocal edema encompassing the eloquent locations of the brain that are connected to different regions of the brain through a network of fiber tracts. Thus the lesion load and peri-tumoral edema may affect the metabolic and diffusion status of the normal appearing brain regions as well. Therefore, a sub-analysis was also performed to compare the baseline 3D-EPSI and DTI parameters between patients with brain metastases (n = 10) and patients who did not have any brain lesions (n = 5).
Results Figure 2 demonstrates representative images from a patient with small-cell lung cancer (who received PCI) displaying coregistered DTI (MD and FA) and 3D-EPSI (NAA and Cho) maps through the supra-tentorial and infra-tentorial regions. The baseline and 1-month post-irradiation MD, FA, NAA/Cr, and Cho/Cr values from the hippocampus, CS, thalamus, genu, and splenium of corpus callosum are presented as scatter plots in Figure 3. Significantly increased MD [(.98 ± .01) × 10−3 mm2 /s) vs [(1.16 ± .16) × 10−3 mm2 /s], P = .021), Cho/Cr
4
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
(.24 ± .02 vs .27 ± .02, P = .02) and a trend toward decreased NAA/Cr (1.48 ± .07 vs 1.27 ± .20, P = .06) were observed from the hippocampus post-irradiation. Compared to baseline values, a significant reduction in FA (.35 ± .02 vs .31 ± .03, P = .02) was also observed from the CS after radiation therapy. Additionally, significant increases in MD (.82 × ± .07 × 10−3 mm2 /s vs .95 × ± .13 × 10−3 mm2 /s, P = .04) and Cho/Cr (.26 ± .02 vs .29 ± .02, P = .02) were also observed from genu of the corpus callosum. Increase in MD and Cho/Cr and reduction in FA and NAA/Cr were also observed from the splenium of corpus callosum and thalamus, however, these changes were not significant. No significant differences in the 3D-EPSI and DTI parameters were observed from any of the other regions studied. There was no significant association between the EPSI and DTI parameters from different regions using Pearson’s correlation analysis. When the differences in the baseline 3D-EPSI and DTI parameters between the two groups of patients [patients with (n = 10) and without brain lesion (n = 5)] were evaluated, no significant differences in NAA/Cr, Cho/Cr, and MD values were observed. However, FA values were significantly different (P < .05) from the hippocampus, CS, genu, splenium, thalamus, basal ganglia, and cingulate gyrus regions between the two groups of patients. To account for differences in the baseline FA values, FA values at post-irradiated period were normalized with respect to the baseline values so that each patient acts as its own control. Differences in the baseline FA values from the hippocampus and CS along with variation in the normalized (from baseline) FA values from the 7 patients are shown in Figure 4.
Fig 3. Scatter Plots demonstrate variations in individual patient values of MD (A), FA (B), NAA/Cr (C), and Cho/Cr (D) between baseline (solid circle) and post-irradiation period (open circle) from hippocampus, CS, thalamus, genu and splenium of the corpus callosum. *P < .05.
Predominantly, a decline in FA values from these regions was observed in most patients. However, other regions, such as the cingulate gyrus, motor cortex and frontal and occipital white matter did not demonstrate this trend in FA.
Discussion In this preliminary study, we demonstrated the feasibility of using high spatial resolution 3D-EPSI and DTI data in determining injury to normal appearing regions of the brain subsequent to radiation therapy in the subacute phase. Significant alterations in the EPSI and DTI parameters were observed from these regions due to radiation therapy. Although clinical trials have demonstrated that radiation therapy leads to local control and improved survival,2 concerns about potential damage to the normal brain remain.3 This is further complicated by clinical data which suggests that higher doses of radiation may in fact be more effective for treating the neoplasm. Therefore, early determination of radiation induced injury to the normal brain regions would provide support for safer and more efficacious alternative therapeutic interventions tailored for the individual patient.
While the underlying mechanism of radiation-induced injury is not clear, it has been proposed that demyelination and necrosis in white matter is related to a gradual loss of oligodendrocytes or their precursors after irradiation.30 Stem and precursor cells located in hippocampus are believed to be crucial for normal memory function.31 Recent evidence indicates that the presence of free radicals, impaired antioxidant pool, high metabolic rate, and presence of actively dividing cells in the hippocampal dentate gyrus and subventricular zone increases the vulnerability of these regions to even lower doses of radiation.32 Because the hippocampus is rarely involved in intracranial metastatic disease,33, 34 recent studies have evaluated the feasibility of sparing the hippocampus while simultaneously treating the rest of the brain with radiotherapy.35, 36 In this endeavor, helical tomotherapy, linear accelerator based intensity modulated radiotherapy and proton therapy are currently being explored to reduce radiation induced injury.37, 38 Availability of high-resolution quantitative 1 H MRS and DTI parameters that can assess brain damage, as demonstrated in the present study, can aid in a more accurate dose-painting with these aforementioned techniques.
Chawla et al: Assessment of radiation injury to normal brain
5
Fig 4. Bar diagrams demonstrating the variations in baseline mean values of FA from hippocampus and CS between PCI (white bars, n = 5) and brain metastases (gray bars, n = 10) patients. Error bars indicate ± 1SD, *indicates P < .05. Variations in percentage normalized FA values (from baseline) of individual subjects from the hippocampus and CS region are shown as scatter plots. PCI patients are denoted in black squares and brain metastatses patients by black triangles.
Several studies have reported the potential of 1 H MRS in evaluating both short and long-term effects of radiation therapy to normal brain parenchyma in patients with brain tumors.7–12 All of these studies have reported significant reduction in the NAA and few of these studies8, 9 have reported elevation in the Cho levels from several different regions including the periventricular white matter, thalamus or the occipital lobe. In the present study, we observed a significant increase in Cho/Cr from the hippocampus and genu of corpus callosum. Elevated Cho from the normal brain after irradiation has also been reported earlier.8, 9 Cho containing phospholipids are abundant in myelin and the cell membranes39 and neurodegeneration and demyelination have been described in irradiated brains.40 It is postulated that irradiation induced damage of oligodendrocytes may cause breakdown of myelin sheath and the cell membranes, resulting in rapid membrane turnover and increased Cho levels following irradiation.40 However, Sundgren et al10 reported decreased Cho from normal brain following irradiation. We believe that this discrepancy might result from difference in the clinical follow-up time or larger size of the voxel used in that study. Nevertheless, the alterations in the Cho levels support the notion that radiation therapy disrupts the metabolism of the normal brain. Additionally, we also observed a trend toward a decrease in NAA/Cr from the hippocampus. While the deleterious effects of radiation in the
6
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
human hippocampus have not been reported, a previous study has reported abnormal metabolite levels in the rat hippocampus after proton therapy.41 The complex pathological events after radiation have been confirmed by histopathological studies in animal models42–44 and postmortem human brain specimens.45 Previous 1 H MRS studies7–12 used single voxel or single slice multi-voxel methods and thus were limited to a small region of the brain. By contrast, the 3D-EPSI method allows measurements of metabolite alterations from the entire brain with excellent spatial resolution and signal-to-noise ratio in a relatively short period. A major strength of the 3D-EPSI method is the availability of high-quality spectra from both the supratentorial and infratentorial brain regions which facilitated comprehensive assessment of adverse effects of radiation injury from different regions of the brain in the same session. Significant dose-dependent changes in diffusivity from the corpus callosum have been reported as early as 4 weeks after completion of radiation therapy.26 In another study, Kitahara et al24 reported increased MD and reduced FA from the supraventricular white matter region of brain tumor patients treated with a combination of surgery and chemo-radiation therapy. In a related study, Chapman et al25 reported significant alterations in the diffusivity from the parahippocampal cingulum in patients undergoing conformal fractionated radiation
therapy. Similarly decreased FA was observed from the normal brain regions in pediatric medulloblastoma patients treated with radiotherapy5 and from the supratentorial and infratentorial regions in adult small cell lung cancer patients receiving PCI.46 Our DTI results are in agreement with these studies and imply that DTI can detect changes in the normal brain tissues induced by radiation. Our findings and published reports support the hypothesis that corpus callosum and central semiovale are radiosensitive white matter regions and that radiation causes significant loss to structural integrity characterized by predominant demyelination, axonal degeneration, and incoherent orientation of the fiber tracts. We did not observe any significant correlation between EPSI and DTI parameters in this study. When considering relationships between various parameters, it is important to consider how different pathophysiological phenomena can influence their values. While NAA and Cho levels are influenced by neuronal integrity and cellular density/membrane turnover, several factors, such as viscosity of the medium, cell density, nuclear-cytoplasmic ratio, barriers to diffusion by membrane bound organelles, molecular crowding and presence of active transport affect MD and FA. The absence of a significant correlation between MRS and DTI parameters may be due to the fact that these parameters may be independent of each other since they do not measure the same physiological phenomena. The relatively small sample size in our study may also contribute to the lack of correlation between EPSI and DTI parameters. Multi-parametric data analysis, as performed in the present study allows us to use the unique strengths of different imaging techniques. Given the inherently different metabolic and physiological information that 3D-EPSI and DTI parameters provide, we believe that when used in conjunction, these parameters can play a complementary role and help in broader assessment of the extent of radiation induced injury. Though promising, the results of our initial experience should be treated with caution as these findings are from a relatively small number of patients as we lost a good number of patients on follow-up. While the sample size was small, our findings demonstrate the potential of EPSI and DTI in assessing the adverse effect of radiation therapy. Future studies comprising a larger cohort of patients are warranted to validate our findings. Another limitation of the present study was the variation in radiation doses between patients, as treatment intent differed between patients receiving prophylactic PCI versus curative WBRT for brain metastasis. While it would be of interest to correlate alterations in 1 H MRS and DTI parameters with changes in cognitive function, the present study was not designed to perform the neurocognitive assessment in these patients. Another potential limitation of the current study is the use of metabolite ratios rather than absolute concentrations for the EPSI data. This study was performed using a phased array coil and these coils generally suffer from B1 inhomogeneity, which leads to variable signal intensities across the FOV. Therefore, we normalized the NAA and Cho concentration with respect to that of Cr from the same ROI. We believe that normalization to Cr does not alter the true metabolic alterations as previous studies8, 11 have reported non-significant variation in the Cr concentration from normal brain after irradiation. No significant changes in the Cr concentrations were noted in our study as well (data not shown).
In conclusion, preliminary findings suggest 3D-EPSI and DTI may be used to detect the subacute radiation injury from different regions of the normal brain.
The support of MR imaging coordinators Lisa Desiderio, Kelly Sexton, Mary Giancoli, and Matthew Voluck is gratefully acknowledged. The authors greatly acknowledge the financial support by Bayer Healthcare Pharmaceuticals Inc, Wayne, NJ, which partially funded this work.
References 1. Patchell RA, Tibbs PA, Regine WF, et al. Postoperative radiotherapy in the treatment of single metastases to the brain: a randomized trial. JAMA 1998;280(17):1485-1489. 2. Auperin A, Arriagada R, Pignon JP, et al. Prophylactic cranial irradiation for patients with small-cell lung cancer in complete remission. Prophylactic Cranial Irradiation Overview Collaborative Group. N Engl J Med 1999;341(7):476-484. 3. Davey P. Brain metastases: treatment options to improve outcomes. CNS Drugs 2002;16(5):325-338. 4. Hellstrom NA, Bjork-Eriksson T, Blomgren K, et al. Differential recovery of neural stem cells in the subventricular zone and dentate gyrus after ionizing radiation. Stem Cells (Dayton, Ohio) 2009;27(3):634-641. 5. Khong PL, Kwong DL, Chan GC, et al. Diffusion-tensor imaging for the detection and quantification of treatment-induced white matter injury in children with medulloblastoma: a pilot study. Am J Neuroradiol 2003;24(4):734-740. 6. Ricard D, Taillia H, Renard JL. Brain damage from anticancer treatments in adults. Curr Opin Oncol 2009;21(6):559-565. 7. Virta A, Patronas N, Raman R, et al. Spectroscopic imaging of radiation-induced effects in the white matter of glioma patients. Magn Reson Imaging 2000;18(7):851-857. 8. Kaminaga T, Shirai K. Radiation-induced brain metabolic changes in the acute and early delayed phase detected with quantitative proton magnetic resonance spectroscopy. J Comput Assist Tomogr 2005;29(3):293-297. 9. Esteve F, Rubin C, Grand S, et al. Transient metabolic changes observed with proton MR spectroscopy in normal human brain after radiation therapy. Int J Radiat Oncol Biol Phys 1998;40(2):279286. 10. Sundgren PC, Nagesh V, Elias A, et al. Metabolic alterations: a biomarker for radiation-induced normal brain injury-an MR spectroscopy study. J Magn Reson Imaging 2009;29(2):291-297. 11. Usenius T, Usenius JP, Tenhunen M, et al. Radiation-induced changes in human brain metabolites as studied by 1H nuclear magnetic resonance spectroscopy in vivo. Int J Radiat Oncol Biol Phys 1995;33(3):719-724. 12. Szigety SK, Allen PS, Huyser-Wierenga D, et al. The effect of radiation on normal human CNS as detected by NMR spectroscopy. Int J Radiat Oncol Biol Phys 1993;25(4):695-701. 13. Lee MC, Pirzkall A, McKnight TR, et al. 1H-MRSI of radiation effects in normal-appearing white matter: dose-dependence and impact on automated spectral classification. J Magn Reson Imaging 2004;19(4):379-388. 14. Movsas B, Li BS, Babb JS, et al. Quantifying radiation therapyinduced brain injury with whole-brain proton MR spectroscopy: initial observations. Radiology 2001;221(2):327-331. 15. Ebel A, Soher BJ, Maudsley AA. Assessment of 3D proton MR echo-planar spectroscopic imaging using automated spectral analysis. Magn Reson Med 2001;46(6):1072-1078. 16. Maudsley AA, Darkazanli A, Alger JR, et al. Comprehensive processing, display and analysis for in vivo MR spectroscopic imaging. NMR Biomed 2006;19(4):492-503.
Chawla et al: Assessment of radiation injury to normal brain
7
17. Govind V, Sharma KR, Maudsley AA, et al. Comprehensive evaluation of corticospinal tract metabolites in amyotrophic lateral sclerosis using whole-brain 1H MR spectroscopy. PLoS One 2012;7(4):e35607. doi: 10.1371/journal.pone.0035607. 18. Mathiesen HK, Tscherning T, Sorensen PS, et al. Multi-slice echoplanar spectroscopic MR R2.9 imaging provides both global and local metabolite measures in multiple sclerosis. Magn Reson Med 2005;53(4):750-759. 19. Verma G, Woo J, Chawla S, et al. Whole-brain analysis of amyotrophic lateral sclerosis by using echo-planar spectroscopic imaging. Radiology 2013;267:851-857. 20. Roy B, Gupta RK, Maudsley AA, et al. Utility of multiparametric 3-T MRI for glioma characterization. Neuroradiology 2013;55:603613. 21. Wang S, Kim S, Chawla S, et al. Differentiation between glioblastomas and solitary brain metastases using diffusion tensor imaging. Neuroimage 2009;44:653-660. 22. Wang S, Wu EX, Qiu D, et al. Longitudinal diffusion tensor magnetic resonance imaging study of radiation-induced white matter damage in a rat model. Cancer Res 2009;69(3):1190-1198. 23. Trivedi R, Khan AR, Rana P, et al. Radiation-induced early changes in the brain and behavior: serial diffusion tensor imaging and behavioral evaluation after graded doses of radiation. J Neurosci Res 2012;90(10):2009-2019. 24. Kitahara S, Nakasu S, Murata K, et al. Evaluation of treatmentinduced cerebral white matter injury by using diffusion-tensor MR imaging: initial experience. Am J Neuroradiol 2005;26(9):2200-2206. 25. Chapman CH, Nagesh V, Sundgren PC, et al. Diffusion tensor imaging of normal-appearing white matter as biomarker for radiation-induced late delayed cognitive decline. Int J Radiat Oncol Biol Phys 2012;82(5):2033-2040. 26. Nagesh V, Tsien CI, Chenevert TL, et al. Radiation-induced changes in normal-appearing white matter in patients with cerebral tumors: a diffusion tensor imaging study. Int J Radiat Oncol Biol Phys 2008;70(4):1002-1010. 27. Khong PL, Leung LH, Chan GC, et al. White matter anisotropy in childhood medulloblastoma survivors: association with neurotoxicity risk factors. Radiology 2005;236(2):647-652. 28. Schultheiss TE, Kun LE, Ang KK, et al. Radiation response of the central nervous system. Int J Radiat Oncol Biol Phys 1995;31(5):10931112. 29. Constine LS, Konski A, Ekholm S, et al. Adverse effects of brain irradiation correlated with MR and CT imaging. Int J Radiat Oncol Biol Phys 1988;15(2):319-330. 30. Kurita H, Kawahara N, Asai A, et al. Radiation-induced apoptosis of oligodendrocytes in the adult rat brain. Neurol Res 2001;23(8):869-874. 31. Koehl M, Abrous DN. A new chapter in the field of memory: adult hippocampal neurogenesis. Eur J Neurosci 2011;33(6):1101-1114. 32. Manda K, Reiter RJ. Melatonin maintains adult hippocampal neurogenesis and cognitive functions after irradiation. Prog Neurobiol 2010;90(1):60-68.
8
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
33. Marsh JC, Herskovic AM, Gielda BT, et al. Intracranial metastatic disease spares the limbic circuit: a review of 697 metastatic lesions in 107 patients. Int J Radiat Oncol Biol Phys 2010;76(2):504512. 34. Regine WF, Schmitt FA, Scott CB, et al. Feasibility of neurocognitive outcome evaluations in patients with brain metastases in a multi-institutional cooperative group setting: results of Radiation Therapy Oncology Group trial BR-0018. Int J Radiat Oncol Biol Phys 2004;58(5):1346-1352. 35. Ghia A, Tome WA, Thomas S, et al. Distribution of brain metastases in relation to the hippocampus: implications for neurocognitive functional preservation. Int J Radiat Oncol Biol Phys 2007;68(4):971-977. 36. Gutierrez AN, Westerly DC, Tome WA, et al. Whole brain radiotherapy with hippocampal avoidance and simultaneously integrated brain metastases boost: a planning study. Int J Radiat Oncol Biol Phys 2007;69(2):589-597. 37. Gondi V, Tolakanahalli R, Mehta MP, et al. Hippocampalsparing whole-brain radiotherapy: a “how-to” technique using helical tomotherapy and linear accelerator-based intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys 2010;78(4):12441252. 38. Yuh GE, Loredo LN, Yonemoto LT, et al. Reducing toxicity from craniospinal irradiation: using proton beams to treat medulloblastoma in young children. Cancer J 2004;10(6):386-390. 39. Lampert PW, Davis RL. Delayed effects of radiation on the human central nervous system. ‘‘Early’’ and ‘‘late’’ delayed reaction. Neurology 1964;14:912-917. 40. Belka C, Budach W, Kortmann RD, et al. Radiation induced CNS toxicity—molecular and cellular mechanisms. Br J Cancer 2001;85:1233-1239. 41. Rabinov JD, Cheng LL, Lee PL, et al. MR spectroscopic changes in the rat hippocampus following proton radiosurgery. Stereotact Funct Neurosurg 2006;84(4):147-154. 42. Price RE, Langford LA, Jackson EF, et al. Radiation-induced morphologic changes in the rhesus monkey (Macaca mulatta) brain. J Med Primatol 2001;30(2):81-87. 43. Chiang CS, McBride WH, Withers HR. Radiation-induced astrocytic and microglial responses in mouse brain. Radiother Oncol 1993;29(1):60-68. 44. Benczik J, Tenhunen M, Snellman M, et al. Late radiation effects in the dog brain: correlation of MRI and histological changes. Radiother Oncol 2002;63(1):107-120. 45. Burger PC, Mahley MS, Jr., Dudka L, et al. The morphologic effects of radiation administered therapeutically for intracranial gliomas: a postmortem study of 25 cases. Cancer 1979;44(4):12561272. 46. Welzel T, Niethammer A, Mende U, et al. Diffusion tensor imaging screening of radiation-induced changes in the white matter after prophylactic cranial irradiation of patients with small cell lung cancer: first results of a prospective study. Am J Neuroradiol 2008;29(2):379-383.
ABSTRACT BACKGROUND AND PURPOSE
Whole brain radiation therapy (WBRT) may cause cognitive and neuropsychological impairment and hence objective assessment of adverse effects of radiation may be valuable to plan therapy. The purpose of our study was to determine the potential of echo planar spectroscopic imaging (EPSI) and diffusion tensor imaging (DTI) in detecting subacute radiation induced injury to the normal brain. MATERIALS AND METHODS
Four patients with brain metastases and three patients with lung cancer underwent cranial irradiation. These patients were subjected to 3D-EPSI and DTI at two time points (preradiation, and 1 month post-irradiation). Parametric maps of N-acetyl aspartate (NAA), creatine (Cr), choline (Cho), mean diffusivity (MD), and fractional anisotropy (FA) were generated and co-registered to post-contrast T1-weighted images. Normal appearing graymatter and white-matter regions were compared between the two time points to assess sub-acute effects of radiation using independent sample t-tests.
Keywords: Radiation injury, whole brain radiation therapy, prophylactic cranial irradiation, echo-planar spectroscopic imaging, diffusion tensor imaging. Acceptance: Received May 3, 2013, and in revised form July 8, 2013. Accepted for publication August 3, 2013. Correspondence: Address correspondence to Harish Poptani, PhD, Department of Radiology, University of Pennsylvania. E-mail: Harish.Poptani@ uphs.upenn.edu. J Neuroimaging 2013;00:1-8. DOI: 10.1111/jon.12070
RESULTS
Significantly increased MD (P = .02), Cho/Cr (P = .02) and a trend towards a decrease in NAA/Cr (P = .06) was observed from the hippocampus. Significant decrease in FA (P = .02) from the centrum-semiovale and a significant increase in MD (P = .04) and Cho/Cr (P = .02) from genu of corpus-callosum was also observed. CONCLUSIONS
Our preliminary findings suggest that 3D-EPSI and DTI may provide quantitative measures of radiation induced injury to the normal brain.
Introduction Whole brain radiation therapy (WBRT) is a standard palliative treatment option for patients with brain metastases and has been shown to improve local control.1 In patients with small cell lung cancer exhibiting complete clinical remission, the addition of prophylactic cranial irradiation (PCI) has been shown to improve overall survival.2 While WBRT and PCI have a therapeutic benefit, care is needed in delivering the radiation dose as there is a strong body of evidence suggesting that WBRT, may result in severe neuropsychological and cognitive impairment.3 The dentate subgranular zone of the hippocampus is involved in neurogenesis and multipotent progenitor cells located in this region are sensitive to radiation induced injury.4 Damage to the supratentorial white matter such as centrum semiovale (CS) and corpus callosum is also associated with impaired neurocognition.5 However, clinical symptoms usually occur several months after radiation therapy and till date, there
is no reliable neurobehavioral assay to measure the subacute effects of radiation induced injury.6 Therefore, there is a critical need for objective and quantitative methods to assess early treatment-induced damage due to WBRT to ascertain the riskto-benefit ratio and dose determination so that the radiation beam can be specifically tailored to limit damage to the normal brain. 1 H MRS has been used to demonstrate abnormal metabolite pattern from the normal brain, secondary to irradiation.7–12 However, these studies suffered from limited volumetric coverage of the brain since a majority of these studies either used a single voxel or a single slice two-dimensional chemical shift imaging method. The only study that used a 3D 1 H MRS method, reported increased averaged Cho from normal brain areas in patients with gliomas treated with radiation therapy,13 however, regional differences in the Cho levels were not reported in this study. Another study using a non-localized whole
Copyright
◦ 2013 by the American Society of Neuroimaging C
1
brain 1 H MRS approach14 reported a significant reduction in global N-acetyl aspartate (NAA) levels in patients receiving PCI. However, this method does not provide spatial information and hence is not suitable for assessing regional damage to the tissue. Recently, a 3D whole brain echo planar spectroscopic imaging (EPSI) method has been described which provides metabolite maps throughout the brain with excellent spatial resolution.15, 16 The 3D-EPSI technique has been used to ascertain neuronal damage in diffuse neurological disorders17–19 and in brain tumors.20 However, to the best of our knowledge, EPSI has not been used to evaluate radiation induced injury in the normal brain. Diffusion tensor imaging (DTI) is commonly used in the evaluation of brain tumors21 and to evaluate radiation injury to the normal brain in pre-clinical models22, 23 as well as in clinical studies.24–27 These studies reported alterations in mean diffusivity (MD) and fractional anisotropy (FA) from various regions of the normal brain. However, DTI has not been reported in evaluating radiation induced injury in the normal brain of patients with brain metastases. The adverse reactions of radiation can be described in three stages, acute, subacute, and late.28 While delayed radiation injury can be accurately assessed by conventional MR imaging, these methods are not sensitive in demonstrating early injury after irradiation.29 Therefore, we sought to investigate the utility of 3D-EPSI and DTI in detecting subacute radiation induced damage to the normal brain parenchyma in patients with brain metastases and in lung cancer patients receiving PCI.
Materials and Methods Subjects The study was approved by the Institutional Review Board and was compliant with the Health Insurance Portability and Accountability Act. Adult patients with brain metastases harboring at least one lesion greater than 1 cm3 were included in the study. Additionally, patients with lung cancer who were scheduled for PCI and had no prior history of cranial irradiation or cranial surgery were included. Patients previously treated with cranial irradiation were excluded from the study. The study comprised a cohort of 15 patients, 10 patients (mean age ± standard deviation = 65 ± 4.24 years, 5 males and 5 females) with brain metastases were irradiated with fractionated WBRT [total dose = 30-40 gray (Gy) in 10-20 fractions] and five patients (65 ± 15.71 years, 2 males and 3 females) with lung cancer underwent PCI (total dose = 25Gy in 10 fractions). All patients underwent a baseline (prior to irradiation) conventional MR imaging, 3D-EPSI and DTI and were scheduled for a followup study (1 month post-irradiation). However, eight of these patients were lost on follow-up. Of the remaining 7 patients, 4 had metastatic brain lesions [confirmed by histopathological analysis of resected brain specimens (n = 2) or by consistent features on multiple conventional MR imaging (n = 2)] and 3 patients with lung cancer did not harbor any brain lesions. The mean duration between the end of radiation therapy and follow-up imaging study was 30.43 ± 9.02 days for these 7 patients. The clinical characteristics and therapeutic regimens for these patients are presented in Table 1.
2
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
MR Imaging MR imaging was performed on a 3T Tim Trio MR scanner (Siemens Medical System, Erlangen, Germany) using a 12-channel phased array head coil. The imaging protocol included three-plane scout localizer, 3D-magnetization prepared rapid gradient echo (MPRAGE) [Repetition time (TR)/echo time (TE)/inversion time (TI) = 1620/3.9/950 ms)], 192 × 256 matrix, slice thickness = 1mm, 192 slices per slab, flip angle = 15◦ , number of excitations (NEX) = 1, bandwidth (BW) = 150 Hz/pixel); axial proton density weighted images (TR/TE = 4100/13 ms, NEX = 1, slices = 40, slice thickness = 3.0 mm, BW = 130 Hz, flip angle = 120◦ ); axial T2 weighted FLAIR image (TR/TE/TI = 9420/141/2500 ms), slice thickness = 3 mm, slices = 60, flip angle = 170◦ , NEX = 1, BW = 287 Hz/pixel; and contrast enhanced T1 weighted MPRAGE images after administration of standard dose (.1 mmol/kg) of contrast agent (Magnevist, gadopentetate dimeglumine, Bayer Healthcare Pharmaceuticals Inc, Wayne, NJ).
3D Echo-Planar Spectroscopic Imaging Scan parameters for the 3D spin-echo EPSI sequence were: TR/TE = 1710/70 ms, spatial points = 50 × 50 × 18, field of view (FOV) = 280 × 280 × 180 mm3 , voxel size = 5.6 × 5.6 × 10 mm3 (.31 cm3 ), excitation angle = 73◦ , 512 complex points, spectral BW = 2,500 Hz with radiofrequency excitation pulse centered at water resonance, NEX = 1. Water suppression using frequency-selective saturation pulses and inversion-recovery nulling of lipid signal was performed with TI of 198 ms. The acquisition time was 26 minutes including an interleaved water reference acquisition scan obtained using a gradient-echo acquisition with 20◦ excitation angle and a 6.3 ms TE. This water-unsuppressed image was used to perform signal normalization, eddy current correction, and image coregistration.
Diffusion Tensor Imaging DTI data were acquired using 30 non-collinear/non-coplanar directions with a single-shot spin-echo, echo-planar read-out sequence with parallel imaging by using generalized autocalibrating partially parallel acquisition (GRAPPA) and acceleration factor of 2. Sequence parameters were as follows: TR/TE = 5,000/86 ms, number of acquisitions = 3, field of view (FOV) = 22 cm × 22 cm, matrix size = 128 × 128, slice thickness = 3 mm, b = 0, 1,000 s/mm2 , slices = 40 covering the whole brain.
Image Processing 3D-EPSI data were processed offline using the metabolic imaging and data analysis system (MIDAS) package developed by Maudsley et al.16 The data were corrected for inhomogeneity in the B0 field, followed by eddy current correction and interpolation to a spatial resolution of 64 × 64 × 32 voxels. The processing also included steps to reduce ringing artifacts from subcutaneous lipids using a mask of the scalp region derived from the T1-weighted images. Signal intensity normalization of metabolite maps was performed using tissue water as an
Table 1. Clinical Characteristics and Therapeutic Regimens Patient ID
1 2 3 4 5 6 7
Gender/Age (Years)
Primary Site of Cancer
Radiation Type
Radiation Dose (Gy)/Fraction
Number of Fractions
Total Radiation Dose (Gy)
M/68 M/65 F/59 F/76 M/68 F/72 F/47
Laryngeal carcinoma and non-small cell lung cancer Thyroid carcinoma and non-small cell lung cancer Non-small cell lung cancer Small cell lung cancer Non-small cell lung cancer Non-small cell lung cancer Small cell lung cancer
WBRT WBRT WBRT PCI WBRT PCI PCI
3 2 2.5 2.5 2.5 2.5 2.5
10 20 14 10 14 10 10
30 40 35 25 35 25 25
WBRT = whole brain radiation therapy; PCI = prophylactic cranial irradiation; Gy = gray.
Fig 1. Gray and white matter regions analyzed to evaluate the effects of radiation therapy are shown as circular ROIs (20 pixels). Motor cortex (green), CS (black ring), genu (orange), and splenium (yellow) of corpus callosum, periventricular frontal white matter (black), periventricular occipital white matter (brown), thalamus (blue), basal ganglia (pink), cingulate gyrus (red), and hippocampus (purple). internal reference. The automated MIDAS tool was used to compute parametric maps of NAA, Cr, and Cho. The diffusion-weighted images were co-registered to the non-diffusion weighted (b = 0 s/mm2 ) images to minimize eddy-current and/or subject motion induced artifacts using a 3D affine transformation estimated by maximizing the mutual information between the images as described previously.21 The corrected raw images were combined to compute MD, and FA maps. Unsuppressed water images, corresponding 3D-EPSI maps (NAA, Cr, Cho), DTI maps (MD, FA) and FLAIR images were co-registered to contrast enhanced T1-weighted images using a 3D non-rigid transformation and mutual information by combining affine transformation and discrete sine bases. All image processing procedures were performed using in-house developed interactive data language (IDL) routines (ITT Visual Information Solutions, Boulder, CO).
Data Analysis To evaluate the effect of radiation therapy on normal brain parenchyma, circular bilateral regions of interests (ROIs), (20
pixels each) were drawn on normal appearing gray matter (hippocampus, thalamus, motor cortex, cingulate gyrus, and basal ganglia) and white matter [genu and splenium of corpus callosum, periventricular frontal and occipital white matter, and CS] regions (Fig 1). While placing ROIs, care was taken to avoid the metastatic lesions as well as peri-lesional edema. The ROIs were chosen from these areas as they are sensitive to radiation induced injury and are associated with cognitive, behavior, and motor activities.4, 5 The mean NAA, Cr, Cho, MD, and FA values were computed from each of the selected ROIs at baseline and post-irradiation periods. The NAA and Cho values were normalized with respect to Cr for each ROI to obtain the NAA/Cr and Cho/Cr ratios. To evaluate the subacute effects of radiation, average NAA/Cr, Cho/Cr, MD, and FA from all patients were compared between baseline and 1-month post-irradiation periods. The baseline 3D-EPSI data from one of the brain metastasis patients (patient no. 1, Table 1) was corrupted due to severe motion artifacts. As a result, DTI data were incorporated from 7 patients and 3D-EPSI from 6 patients in the final data analysis. Independent sample t-tests were performed with an assumption that significant differences in 3D-EPSI (NAA/Cr and
Chawla et al: Assessment of radiation injury to normal brain
3
Fig 2. Representative brain regions from a 47-year-old female patient with lung cancer receiving PCI are shown to illustrate the feasibility of high spatial resolution co-registered 3D-EPSI and DTI data. Co-registered post-contrast T1 weighted image (A and F), MD (B and G), FA (C and H), NAA (D and I), and Cho (E and J) maps through the supratentorial (upper-panel) and infratentorial (lower-panel) region. CS (open black ROI and white arrow) and hippocampus (open black ROI and black arrow) are shown as representative regions from supratentorial and infratentorial brain parenchyma, respectively. Cho/Cr) and DTI (MD and FA) values from baseline to postirradiation would reflect radiation induced injury. A P value < .05 was considered significant. To investigate the relationship between EPSI (NAA/Cr and Cho/Cr) and DTI parameters (MD and FA) within each anatomical region, Pearson’s correlation coefficient analyses were performed. Patients receiving PCI did not have any detectable brain lesions on diagnostic MR imaging, while patients with brain metastases had up to three lesions with surrounding perifocal edema encompassing the eloquent locations of the brain that are connected to different regions of the brain through a network of fiber tracts. Thus the lesion load and peri-tumoral edema may affect the metabolic and diffusion status of the normal appearing brain regions as well. Therefore, a sub-analysis was also performed to compare the baseline 3D-EPSI and DTI parameters between patients with brain metastases (n = 10) and patients who did not have any brain lesions (n = 5).
Results Figure 2 demonstrates representative images from a patient with small-cell lung cancer (who received PCI) displaying coregistered DTI (MD and FA) and 3D-EPSI (NAA and Cho) maps through the supra-tentorial and infra-tentorial regions. The baseline and 1-month post-irradiation MD, FA, NAA/Cr, and Cho/Cr values from the hippocampus, CS, thalamus, genu, and splenium of corpus callosum are presented as scatter plots in Figure 3. Significantly increased MD [(.98 ± .01) × 10−3 mm2 /s) vs [(1.16 ± .16) × 10−3 mm2 /s], P = .021), Cho/Cr
4
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
(.24 ± .02 vs .27 ± .02, P = .02) and a trend toward decreased NAA/Cr (1.48 ± .07 vs 1.27 ± .20, P = .06) were observed from the hippocampus post-irradiation. Compared to baseline values, a significant reduction in FA (.35 ± .02 vs .31 ± .03, P = .02) was also observed from the CS after radiation therapy. Additionally, significant increases in MD (.82 × ± .07 × 10−3 mm2 /s vs .95 × ± .13 × 10−3 mm2 /s, P = .04) and Cho/Cr (.26 ± .02 vs .29 ± .02, P = .02) were also observed from genu of the corpus callosum. Increase in MD and Cho/Cr and reduction in FA and NAA/Cr were also observed from the splenium of corpus callosum and thalamus, however, these changes were not significant. No significant differences in the 3D-EPSI and DTI parameters were observed from any of the other regions studied. There was no significant association between the EPSI and DTI parameters from different regions using Pearson’s correlation analysis. When the differences in the baseline 3D-EPSI and DTI parameters between the two groups of patients [patients with (n = 10) and without brain lesion (n = 5)] were evaluated, no significant differences in NAA/Cr, Cho/Cr, and MD values were observed. However, FA values were significantly different (P < .05) from the hippocampus, CS, genu, splenium, thalamus, basal ganglia, and cingulate gyrus regions between the two groups of patients. To account for differences in the baseline FA values, FA values at post-irradiated period were normalized with respect to the baseline values so that each patient acts as its own control. Differences in the baseline FA values from the hippocampus and CS along with variation in the normalized (from baseline) FA values from the 7 patients are shown in Figure 4.
Fig 3. Scatter Plots demonstrate variations in individual patient values of MD (A), FA (B), NAA/Cr (C), and Cho/Cr (D) between baseline (solid circle) and post-irradiation period (open circle) from hippocampus, CS, thalamus, genu and splenium of the corpus callosum. *P < .05.
Predominantly, a decline in FA values from these regions was observed in most patients. However, other regions, such as the cingulate gyrus, motor cortex and frontal and occipital white matter did not demonstrate this trend in FA.
Discussion In this preliminary study, we demonstrated the feasibility of using high spatial resolution 3D-EPSI and DTI data in determining injury to normal appearing regions of the brain subsequent to radiation therapy in the subacute phase. Significant alterations in the EPSI and DTI parameters were observed from these regions due to radiation therapy. Although clinical trials have demonstrated that radiation therapy leads to local control and improved survival,2 concerns about potential damage to the normal brain remain.3 This is further complicated by clinical data which suggests that higher doses of radiation may in fact be more effective for treating the neoplasm. Therefore, early determination of radiation induced injury to the normal brain regions would provide support for safer and more efficacious alternative therapeutic interventions tailored for the individual patient.
While the underlying mechanism of radiation-induced injury is not clear, it has been proposed that demyelination and necrosis in white matter is related to a gradual loss of oligodendrocytes or their precursors after irradiation.30 Stem and precursor cells located in hippocampus are believed to be crucial for normal memory function.31 Recent evidence indicates that the presence of free radicals, impaired antioxidant pool, high metabolic rate, and presence of actively dividing cells in the hippocampal dentate gyrus and subventricular zone increases the vulnerability of these regions to even lower doses of radiation.32 Because the hippocampus is rarely involved in intracranial metastatic disease,33, 34 recent studies have evaluated the feasibility of sparing the hippocampus while simultaneously treating the rest of the brain with radiotherapy.35, 36 In this endeavor, helical tomotherapy, linear accelerator based intensity modulated radiotherapy and proton therapy are currently being explored to reduce radiation induced injury.37, 38 Availability of high-resolution quantitative 1 H MRS and DTI parameters that can assess brain damage, as demonstrated in the present study, can aid in a more accurate dose-painting with these aforementioned techniques.
Chawla et al: Assessment of radiation injury to normal brain
5
Fig 4. Bar diagrams demonstrating the variations in baseline mean values of FA from hippocampus and CS between PCI (white bars, n = 5) and brain metastases (gray bars, n = 10) patients. Error bars indicate ± 1SD, *indicates P < .05. Variations in percentage normalized FA values (from baseline) of individual subjects from the hippocampus and CS region are shown as scatter plots. PCI patients are denoted in black squares and brain metastatses patients by black triangles.
Several studies have reported the potential of 1 H MRS in evaluating both short and long-term effects of radiation therapy to normal brain parenchyma in patients with brain tumors.7–12 All of these studies have reported significant reduction in the NAA and few of these studies8, 9 have reported elevation in the Cho levels from several different regions including the periventricular white matter, thalamus or the occipital lobe. In the present study, we observed a significant increase in Cho/Cr from the hippocampus and genu of corpus callosum. Elevated Cho from the normal brain after irradiation has also been reported earlier.8, 9 Cho containing phospholipids are abundant in myelin and the cell membranes39 and neurodegeneration and demyelination have been described in irradiated brains.40 It is postulated that irradiation induced damage of oligodendrocytes may cause breakdown of myelin sheath and the cell membranes, resulting in rapid membrane turnover and increased Cho levels following irradiation.40 However, Sundgren et al10 reported decreased Cho from normal brain following irradiation. We believe that this discrepancy might result from difference in the clinical follow-up time or larger size of the voxel used in that study. Nevertheless, the alterations in the Cho levels support the notion that radiation therapy disrupts the metabolism of the normal brain. Additionally, we also observed a trend toward a decrease in NAA/Cr from the hippocampus. While the deleterious effects of radiation in the
6
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
human hippocampus have not been reported, a previous study has reported abnormal metabolite levels in the rat hippocampus after proton therapy.41 The complex pathological events after radiation have been confirmed by histopathological studies in animal models42–44 and postmortem human brain specimens.45 Previous 1 H MRS studies7–12 used single voxel or single slice multi-voxel methods and thus were limited to a small region of the brain. By contrast, the 3D-EPSI method allows measurements of metabolite alterations from the entire brain with excellent spatial resolution and signal-to-noise ratio in a relatively short period. A major strength of the 3D-EPSI method is the availability of high-quality spectra from both the supratentorial and infratentorial brain regions which facilitated comprehensive assessment of adverse effects of radiation injury from different regions of the brain in the same session. Significant dose-dependent changes in diffusivity from the corpus callosum have been reported as early as 4 weeks after completion of radiation therapy.26 In another study, Kitahara et al24 reported increased MD and reduced FA from the supraventricular white matter region of brain tumor patients treated with a combination of surgery and chemo-radiation therapy. In a related study, Chapman et al25 reported significant alterations in the diffusivity from the parahippocampal cingulum in patients undergoing conformal fractionated radiation
therapy. Similarly decreased FA was observed from the normal brain regions in pediatric medulloblastoma patients treated with radiotherapy5 and from the supratentorial and infratentorial regions in adult small cell lung cancer patients receiving PCI.46 Our DTI results are in agreement with these studies and imply that DTI can detect changes in the normal brain tissues induced by radiation. Our findings and published reports support the hypothesis that corpus callosum and central semiovale are radiosensitive white matter regions and that radiation causes significant loss to structural integrity characterized by predominant demyelination, axonal degeneration, and incoherent orientation of the fiber tracts. We did not observe any significant correlation between EPSI and DTI parameters in this study. When considering relationships between various parameters, it is important to consider how different pathophysiological phenomena can influence their values. While NAA and Cho levels are influenced by neuronal integrity and cellular density/membrane turnover, several factors, such as viscosity of the medium, cell density, nuclear-cytoplasmic ratio, barriers to diffusion by membrane bound organelles, molecular crowding and presence of active transport affect MD and FA. The absence of a significant correlation between MRS and DTI parameters may be due to the fact that these parameters may be independent of each other since they do not measure the same physiological phenomena. The relatively small sample size in our study may also contribute to the lack of correlation between EPSI and DTI parameters. Multi-parametric data analysis, as performed in the present study allows us to use the unique strengths of different imaging techniques. Given the inherently different metabolic and physiological information that 3D-EPSI and DTI parameters provide, we believe that when used in conjunction, these parameters can play a complementary role and help in broader assessment of the extent of radiation induced injury. Though promising, the results of our initial experience should be treated with caution as these findings are from a relatively small number of patients as we lost a good number of patients on follow-up. While the sample size was small, our findings demonstrate the potential of EPSI and DTI in assessing the adverse effect of radiation therapy. Future studies comprising a larger cohort of patients are warranted to validate our findings. Another limitation of the present study was the variation in radiation doses between patients, as treatment intent differed between patients receiving prophylactic PCI versus curative WBRT for brain metastasis. While it would be of interest to correlate alterations in 1 H MRS and DTI parameters with changes in cognitive function, the present study was not designed to perform the neurocognitive assessment in these patients. Another potential limitation of the current study is the use of metabolite ratios rather than absolute concentrations for the EPSI data. This study was performed using a phased array coil and these coils generally suffer from B1 inhomogeneity, which leads to variable signal intensities across the FOV. Therefore, we normalized the NAA and Cho concentration with respect to that of Cr from the same ROI. We believe that normalization to Cr does not alter the true metabolic alterations as previous studies8, 11 have reported non-significant variation in the Cr concentration from normal brain after irradiation. No significant changes in the Cr concentrations were noted in our study as well (data not shown).
In conclusion, preliminary findings suggest 3D-EPSI and DTI may be used to detect the subacute radiation injury from different regions of the normal brain.
The support of MR imaging coordinators Lisa Desiderio, Kelly Sexton, Mary Giancoli, and Matthew Voluck is gratefully acknowledged. The authors greatly acknowledge the financial support by Bayer Healthcare Pharmaceuticals Inc, Wayne, NJ, which partially funded this work.
References 1. Patchell RA, Tibbs PA, Regine WF, et al. Postoperative radiotherapy in the treatment of single metastases to the brain: a randomized trial. JAMA 1998;280(17):1485-1489. 2. Auperin A, Arriagada R, Pignon JP, et al. Prophylactic cranial irradiation for patients with small-cell lung cancer in complete remission. Prophylactic Cranial Irradiation Overview Collaborative Group. N Engl J Med 1999;341(7):476-484. 3. Davey P. Brain metastases: treatment options to improve outcomes. CNS Drugs 2002;16(5):325-338. 4. Hellstrom NA, Bjork-Eriksson T, Blomgren K, et al. Differential recovery of neural stem cells in the subventricular zone and dentate gyrus after ionizing radiation. Stem Cells (Dayton, Ohio) 2009;27(3):634-641. 5. Khong PL, Kwong DL, Chan GC, et al. Diffusion-tensor imaging for the detection and quantification of treatment-induced white matter injury in children with medulloblastoma: a pilot study. Am J Neuroradiol 2003;24(4):734-740. 6. Ricard D, Taillia H, Renard JL. Brain damage from anticancer treatments in adults. Curr Opin Oncol 2009;21(6):559-565. 7. Virta A, Patronas N, Raman R, et al. Spectroscopic imaging of radiation-induced effects in the white matter of glioma patients. Magn Reson Imaging 2000;18(7):851-857. 8. Kaminaga T, Shirai K. Radiation-induced brain metabolic changes in the acute and early delayed phase detected with quantitative proton magnetic resonance spectroscopy. J Comput Assist Tomogr 2005;29(3):293-297. 9. Esteve F, Rubin C, Grand S, et al. Transient metabolic changes observed with proton MR spectroscopy in normal human brain after radiation therapy. Int J Radiat Oncol Biol Phys 1998;40(2):279286. 10. Sundgren PC, Nagesh V, Elias A, et al. Metabolic alterations: a biomarker for radiation-induced normal brain injury-an MR spectroscopy study. J Magn Reson Imaging 2009;29(2):291-297. 11. Usenius T, Usenius JP, Tenhunen M, et al. Radiation-induced changes in human brain metabolites as studied by 1H nuclear magnetic resonance spectroscopy in vivo. Int J Radiat Oncol Biol Phys 1995;33(3):719-724. 12. Szigety SK, Allen PS, Huyser-Wierenga D, et al. The effect of radiation on normal human CNS as detected by NMR spectroscopy. Int J Radiat Oncol Biol Phys 1993;25(4):695-701. 13. Lee MC, Pirzkall A, McKnight TR, et al. 1H-MRSI of radiation effects in normal-appearing white matter: dose-dependence and impact on automated spectral classification. J Magn Reson Imaging 2004;19(4):379-388. 14. Movsas B, Li BS, Babb JS, et al. Quantifying radiation therapyinduced brain injury with whole-brain proton MR spectroscopy: initial observations. Radiology 2001;221(2):327-331. 15. Ebel A, Soher BJ, Maudsley AA. Assessment of 3D proton MR echo-planar spectroscopic imaging using automated spectral analysis. Magn Reson Med 2001;46(6):1072-1078. 16. Maudsley AA, Darkazanli A, Alger JR, et al. Comprehensive processing, display and analysis for in vivo MR spectroscopic imaging. NMR Biomed 2006;19(4):492-503.
Chawla et al: Assessment of radiation injury to normal brain
7
17. Govind V, Sharma KR, Maudsley AA, et al. Comprehensive evaluation of corticospinal tract metabolites in amyotrophic lateral sclerosis using whole-brain 1H MR spectroscopy. PLoS One 2012;7(4):e35607. doi: 10.1371/journal.pone.0035607. 18. Mathiesen HK, Tscherning T, Sorensen PS, et al. Multi-slice echoplanar spectroscopic MR R2.9 imaging provides both global and local metabolite measures in multiple sclerosis. Magn Reson Med 2005;53(4):750-759. 19. Verma G, Woo J, Chawla S, et al. Whole-brain analysis of amyotrophic lateral sclerosis by using echo-planar spectroscopic imaging. Radiology 2013;267:851-857. 20. Roy B, Gupta RK, Maudsley AA, et al. Utility of multiparametric 3-T MRI for glioma characterization. Neuroradiology 2013;55:603613. 21. Wang S, Kim S, Chawla S, et al. Differentiation between glioblastomas and solitary brain metastases using diffusion tensor imaging. Neuroimage 2009;44:653-660. 22. Wang S, Wu EX, Qiu D, et al. Longitudinal diffusion tensor magnetic resonance imaging study of radiation-induced white matter damage in a rat model. Cancer Res 2009;69(3):1190-1198. 23. Trivedi R, Khan AR, Rana P, et al. Radiation-induced early changes in the brain and behavior: serial diffusion tensor imaging and behavioral evaluation after graded doses of radiation. J Neurosci Res 2012;90(10):2009-2019. 24. Kitahara S, Nakasu S, Murata K, et al. Evaluation of treatmentinduced cerebral white matter injury by using diffusion-tensor MR imaging: initial experience. Am J Neuroradiol 2005;26(9):2200-2206. 25. Chapman CH, Nagesh V, Sundgren PC, et al. Diffusion tensor imaging of normal-appearing white matter as biomarker for radiation-induced late delayed cognitive decline. Int J Radiat Oncol Biol Phys 2012;82(5):2033-2040. 26. Nagesh V, Tsien CI, Chenevert TL, et al. Radiation-induced changes in normal-appearing white matter in patients with cerebral tumors: a diffusion tensor imaging study. Int J Radiat Oncol Biol Phys 2008;70(4):1002-1010. 27. Khong PL, Leung LH, Chan GC, et al. White matter anisotropy in childhood medulloblastoma survivors: association with neurotoxicity risk factors. Radiology 2005;236(2):647-652. 28. Schultheiss TE, Kun LE, Ang KK, et al. Radiation response of the central nervous system. Int J Radiat Oncol Biol Phys 1995;31(5):10931112. 29. Constine LS, Konski A, Ekholm S, et al. Adverse effects of brain irradiation correlated with MR and CT imaging. Int J Radiat Oncol Biol Phys 1988;15(2):319-330. 30. Kurita H, Kawahara N, Asai A, et al. Radiation-induced apoptosis of oligodendrocytes in the adult rat brain. Neurol Res 2001;23(8):869-874. 31. Koehl M, Abrous DN. A new chapter in the field of memory: adult hippocampal neurogenesis. Eur J Neurosci 2011;33(6):1101-1114. 32. Manda K, Reiter RJ. Melatonin maintains adult hippocampal neurogenesis and cognitive functions after irradiation. Prog Neurobiol 2010;90(1):60-68.
8
Journal of Neuroimaging Vol 00 No 0 xxxx 2013
33. Marsh JC, Herskovic AM, Gielda BT, et al. Intracranial metastatic disease spares the limbic circuit: a review of 697 metastatic lesions in 107 patients. Int J Radiat Oncol Biol Phys 2010;76(2):504512. 34. Regine WF, Schmitt FA, Scott CB, et al. Feasibility of neurocognitive outcome evaluations in patients with brain metastases in a multi-institutional cooperative group setting: results of Radiation Therapy Oncology Group trial BR-0018. Int J Radiat Oncol Biol Phys 2004;58(5):1346-1352. 35. Ghia A, Tome WA, Thomas S, et al. Distribution of brain metastases in relation to the hippocampus: implications for neurocognitive functional preservation. Int J Radiat Oncol Biol Phys 2007;68(4):971-977. 36. Gutierrez AN, Westerly DC, Tome WA, et al. Whole brain radiotherapy with hippocampal avoidance and simultaneously integrated brain metastases boost: a planning study. Int J Radiat Oncol Biol Phys 2007;69(2):589-597. 37. Gondi V, Tolakanahalli R, Mehta MP, et al. Hippocampalsparing whole-brain radiotherapy: a “how-to” technique using helical tomotherapy and linear accelerator-based intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys 2010;78(4):12441252. 38. Yuh GE, Loredo LN, Yonemoto LT, et al. Reducing toxicity from craniospinal irradiation: using proton beams to treat medulloblastoma in young children. Cancer J 2004;10(6):386-390. 39. Lampert PW, Davis RL. Delayed effects of radiation on the human central nervous system. ‘‘Early’’ and ‘‘late’’ delayed reaction. Neurology 1964;14:912-917. 40. Belka C, Budach W, Kortmann RD, et al. Radiation induced CNS toxicity—molecular and cellular mechanisms. Br J Cancer 2001;85:1233-1239. 41. Rabinov JD, Cheng LL, Lee PL, et al. MR spectroscopic changes in the rat hippocampus following proton radiosurgery. Stereotact Funct Neurosurg 2006;84(4):147-154. 42. Price RE, Langford LA, Jackson EF, et al. Radiation-induced morphologic changes in the rhesus monkey (Macaca mulatta) brain. J Med Primatol 2001;30(2):81-87. 43. Chiang CS, McBride WH, Withers HR. Radiation-induced astrocytic and microglial responses in mouse brain. Radiother Oncol 1993;29(1):60-68. 44. Benczik J, Tenhunen M, Snellman M, et al. Late radiation effects in the dog brain: correlation of MRI and histological changes. Radiother Oncol 2002;63(1):107-120. 45. Burger PC, Mahley MS, Jr., Dudka L, et al. The morphologic effects of radiation administered therapeutically for intracranial gliomas: a postmortem study of 25 cases. Cancer 1979;44(4):12561272. 46. Welzel T, Niethammer A, Mende U, et al. Diffusion tensor imaging screening of radiation-induced changes in the white matter after prophylactic cranial irradiation of patients with small cell lung cancer: first results of a prospective study. Am J Neuroradiol 2008;29(2):379-383.
