| [Objective]1. To investigate the 1H-MRS imaging and metabolite levels changes of normal appearing temporal white matter in nasopharyngeal carcinoma after radiotherapy with compare to group of Pre-RT.2. To investigate the DTI and water diffusion parameters changes of normal appearing temporal white matter in nasopharyngeal carcinoma after radiotherapy with compare to group of Pre-RT;3. Through analyze the radiation-induced brain damage in DTI and 1H-MRS changes in patients with NPC after RT in different times, to investigate the application of DTI and'H-MRS in monitoring brain radiation injury in the long-term follow-up after radiotherapy.[Materials and Methods]1. Investigate objectsThere were 75 patients (52males,23females; average age 45.6 years; age range 19-71 years) included in the study from April to October 2010. All patients had NPC confirmed by biopsy. Written informed consent was obtained before all examinations, which were performed before or up to 6 years after completion of RT. They staging from T1N0M0 to T4N2M0 (Union for International Cancer Control, seventh edition,2009). All patients received a single course of RT with 3-dimensional conformal techniques (total doses/daily fraction/exposures/,66-74GY/1.8-2.0GY/30-35 times). All of the patients underwent 2 to 4 courses of concomitant chemotherapy during and 2 months after RT with one or more agents, such as cisplatin, cyclophosphamide, and gemcitabine. Excluded from the study were patients with intracranial invasion, brain tumors or metastases, vascular lesions of the brain, heart disease, or diabetes.A total of 75 spectral and DTI examinations were performed. Based on previous MR spectroscopy and DTI studies that reported adverse effects of radiation therapy, we divided the dataset on the basis of the time before and after completion of RT into 6 groups in order to evaluate the transient reaction effectively:group 1 (pre-RT, control group, n=18); group 2 (0-3 months, n=16); group 3 (>3-6 months, n=12); group 4 (>6-9 months, n=10); group 5 (>9-12 months, n=8); group 6 (>12 months, n=11).2. MRI studiesAll conventional MR imaging,2D CSI MR spectroscopy, and DTI scans were obtained using a 3.0 T GE clinical scanner (SIGNA EXCITE II GE Medical Systems, Milwaukee, Wis.). The conventional MR imaging brain protocol included axial and sagittal T1-weighted images (TR/TE,600/15ms), axial T2-weighted images (TR/TE, 5200/140ms), axial T2-weighted fluid attenuated inversion recovery (TR/TE/IR, 9000/120/2100ms), and axial and coronal postgadolinium T1-weighted images (Gd-DTPA,0.2 mmol/kg).DTI scans were obtained before post-contrast T1-weighted images using a spin-echo echo-planar imaging sequence with a repetition time of 8000 ms, echo time of 87 ms,240×240 mm2 field of view,128×128 matrix, and 5-mm slice thickness, with a 1.0 mm gap. Diffusion-sensitizing gradient encoding was applied in 25 directions using a higher-diffusion weighting factor of b=1000 s/mm2, and one set of null images with b=0 s/mm2 was acquired. The scan time was approximately 4 minutes. The position and orientation of the DTI slices on a mid-sagittal image were used to prescribe the slice position and orientation for the follow-up examination.In each case,2D CSI MR spectroscopy was performed as the last sequence more than 30 minutes after contrast-enhanced MR imaging. Spectroscopic data were obtained using point-resolved (PRESS) 2D CSI performed with the following parameters:Probe-P, extended dynamic range, TR/TE 1200/144 ms, FOV 240×240 mm, thickness 10 mm, matrix 16×16, NEX 1, scanning time 6 minl5s(unilateral). The region of interest (ROI) was placed on a new axial T2-weighted images sequence The bilateral ROI contained temporal WM approximately 2×2cm2～2×4cm2. Six presaturation bands were placed around the big region of interest (ROI) to minimize contamination by osseous and cerebrospinal fluid (CSF)-containing structures. Automatic pre-scanning was performed before each spectroscopic scan to ensure that full width-at-half-maximum (FWHM) was kept at 12±3 Hz and water saturation was above 98%.3. Post-processingThe 2D CSI MR spectroscopy data were processed off-line using standard software on a workstation (FUNCTOOL 4.4; GE Medical Systems). First, we made sure that bilateral temporal lobe WM had no abnormal signal intensity in T1WI and T2-FLAIR sequence. Then we chose 3 fixed minimum voxel in each temporal lobe WM. Metabolite values of N-acetylaspartic acid (NAA), choline (Cho), and creatine (Cr) were calculated automatically from the area under each metabolite peak using the FUNCTOOL software. Metabolite ratios (NAA/Cho, NAA/Cr and Cho/Cr) were automatically calculated simultaneously. We accepted the average of these 3 ratio values in the bilateral temporal lobe WM as the final metabolite ratio value.Post-processing, including eddy current correction, was performed using FUNCTOOL. We first identified the slice as close to the spectral location T2WI anatomy as possible. One uniform circular regions of interest (ROI) were placed in each temporal lobe WM in the b=0 images. ROIs were identified from the b=0 images rather than the mean diffusibility (MD) or fractional anisotropy (FA) maps, which avoids the problem of using the dependent variable. The size of the ROIs was 30±3 mm2. FA, MD,ADC,λ⊥,λ2 andλ3 for each ROI were calculated. Mean diffusivity, FA, axial diffusivity (λ‖), and radial diffusivity (λ⊥) were derived according to the following equations:For each subject, the measures of FA, MD, ADC,λ‖andλ⊥were averaged across all the ROIs of each temporal WM region bilaterally.4. Statistical analysisData were analyzed using the Statistical Package for Social Sciences (SPSS for Windows, Version 13.0, SPSS, Chicago, IL). The date of NAA/Cho,NAA/Gr,Cho/Cr,FA,MD,ADC,λ‖,λ⊥, tested by Levence firstly. Between group comparisons was performed using one-way analysis of variance (one-way ANOVA) test with the post-hoc LSD-t test. The Brown-Forsythe or Welch test with the post-hoc Dunnett's T3 test was used when the variance of parameter was heterogeneous. One-way analysis of variance (one-way ANOVA) was used to analyze the difference of the age of radiotherapy group and control group, using Pearsonχ2 test comparing whether the gender differences between the groups. A P-value of 0.05 was considered statistically significant.[RESULTS]1. Radiotherapy group compared with the control group no statistically significant difference in age (F=1.888, P=0.108), compared between the groups there were no significant gender differences statistically significant (χ2=9.287, P=0.505).2. NAA/Cho ratiosThe mean NAA/Cho ratio of Pre-RT group was 1.271±0.23, The NAA/Cr ratios of groups2-5 after radiotherapy were 0.952±0.158,1.005±0.182,1.032±0.101,0.999±0.160 respectively, and the one-way ANOVA shows difference between the 6 groups was statistically significant. The NAA/Cho ratio decreased significantly within 1 year after RT (groups 2-5) compared with the pre-RT group (group 1). After 12 months (group 6) the NAA/Cho ratio was close to the values in the pre-RT group, but had not fully recovered. There were no significant differences in multiple comparisons of the NAA/Cho ratio between groups 2,3,4 and 5. The NAA/Cho ratio of groups 2 decreased significantly compared with the group 6. The minimum value of each parameter was observed at 0-3 months (group 1, average 2.2 months) after RT, and again at >9-12 months (group 5, average 11.6 months) after RT with an interval of about 9.4 months.3. NAA/Cr ratiosThe mean NAA/Cr ratio of Pre-RT group was 1.696±0.338, The NAA/Cr ratios of groups2-5 after radiotherapy were 1.328±0.33,1.371±0.283,1.375±0.285,1.346±0.172 respectively, and the one-way ANOVA shows difference between the 6 groups was statistically significant. The NAA/Cr ratios decreased significantly within 1 year after RT (groups 2-5) compared with the pre-RT group (group 1). After 12 months (group 6) the NAA/Cr ratio was 1.609±0.221, closed to the values in the pre-RT group, but had not fully recovered. There were no significant differences in multiple comparisons of the NAA/Cr ratio between groups 2,3,4 and 5. The NAA/Cr ratios of groups (2,3 and 5) decreased significantly compared with the group 6. Variations in both NAA/Cho and NAA/Cr values over time were similar.4. Radial diffusivity (λ⊥) and axial diffusivity (λ‖) valuesThe meanλ⊥value of Pre-RT group was 6.075±0.341, The meanλ⊥values of groups2-5 after radiotherapy were 6.700±0.379,6.976±0.527,6.621±0.388,6.751±0.460 respectively, and the one-way ANOVA shows difference between the 6 groups was statistically significant. Theλ⊥mean values increased significantly within 1 year after RT (groups 2-5) compared with the pre-RT group (group 1) and group 6. After 12 months (group 6), the meanλ⊥was slightly higher than pre-RT, but was not statistically significantly different (P=0.331). There were no significant differences in multiple comparisons ofλ⊥mean values between groups 2,3,4 and 5. Maximum values ofλ⊥were observed at >3-6 months (group 3, average 4.6 months) and >9-12 months (group 5, average 11.6 months) after RT, with an interval of about 7 months.The meanλ‖value of Pre-RT group was 12.252±0.713, The meanλ‖values of groups2-6 after radiotherapy were 11.764±0.574,11.842±0.471,11.569±0.552,12.050±0.614,12.100±0.529 respectively. The meanλ‖decreased within 9 months after RT (groups 2,3, and 4) compared with the pre-RT group (group 1). The meanλ‖values more than 9 months after RT (groups 5 and 6) were not significantly different from the pre-RT group. Minimalλ‖values of were seen at 0-3 months (group 2, average 2.2 months) and >6-9 months (group 5, average 7.9 months) after RT, with an interval of about 5.7 months.5. FA valuesThe mean FA value of Pre-RT group was 0.452±0.030, The mean FA values of groups2-6 after radiotherapy were 0.379±0.028,0.382±0.028,0.392±0.029,0.388±0.022,0.423±0.023 respectively. The FA values in all five groups that received RT were significantly lower than in the pre-RT group (group 1). The variation in FA values was similar to that seen in the ratios of NAA/Cho and NAA/Cr. Minimal values of FA also were observed at 0-3 months (group 2) and >9-12 months (group 5) after RT, with an interval of about 9.4 months. Although the FA level had been restored partially 1 year after RT (group 6), it remained significantly lower than in the pre-RT group (P=.014).6. Cho/Cr ratio, MD and ADC valuesThe Cho/Cr ratio, MD and ADC values of in post-RT groups were not significantly different from those in the pre-RT group. Variations in both MD and ADC values over time were similar with Radial diffusivity (λ⊥).[Conclusion]1. Changes in metabolite levels and water diffusion parameters within visually normal-appearing WM shown on conventional MR images can be investigated quantitatively using diffusion-tensor imaging and proton MR spectroscopy.2. MRSIn the present study, metabolic alterations of temporal lobe NAWM after irradiation were dynamic and transient, consistent with the findings of other studies. NAA/Cho and NAA/Cr values decreased significantly within 1 year after RT compared with pre-RT. The greatest decline in both ratios occurred at 0-3 months after RT. Values rose slightly between 3 and 12 months, and then increased to near pre-RT values after 12 month. We also suggested that the decrease in NAA was due to neuronal damage, neuronal cell apoptosis, and neuronal dysfunction.A transient reduction of NAA/Cr also was detected by proton MR spectroscopy. Regional volume changes may cause a decrease in the NAA concentration due to an increase in water content by edema, The investigators suggested that this oscillation may be caused by blood-brain barrier disruption and repair processes. After 12 months, the NAA/Cho and NAA/Cr values had not recovered fully in our study. In our opinion, apoptosis of small partial neurons contributed to the irreversible injury that occurred in the neuronal mitochondrial. NAA cannot be restored to normal levels because the neurons cannot regenerate.There were no changes in Cho/Cr values in our study. Reports regarding the changes in Cho/Cr in normal brain tissue after irradiation have not been consistent. Some studies reported that Cho/Cr decreased, some reported that the ratio was increased, and still others found that Cho/Cr was unaltered. The diverse findings with regard to Cho/Cr were observed during RT, or in the acute reaction and early delayed reaction periods after RT. The specific reason for this is unclear.3. DTIIn our study, theλ⊥value increased significantly and theλ‖value decreased significantly after RT. The recovery time forλ⊥was 1 year, but forλ‖it was just 9 months. We suggest that demyelination accompanying axonal injury is the main effect in NAWM of the temporal lobe after RT. White matter in the brain is known to be highly vulnerable to radiation. Demyelination and axonal injury are the hallmarks of WM injury, each playing a different role in neurologic function.The increase inλ⊥and decrease inλ‖in our study-occurred simultaneously. In a transgenic mouse model of irreversible and reversible dysmyelination, there was an increase inλ⊥and a significant decrease inλ‖and FA as the result of a severe loss of myelin, and most axons were of small diameter and unmyelinated. During the recovery period, a progressive increase inλ‖and FA and a decrease inλ⊥were attributed to an enlarged oligodendrocyte population, newly formed myelin sheaths around additional axons, and a gradual increase in axonal caliber under electron microscopy.After completion of RT, the ionizing effects on brain tissues are sustained for a period of time. The injured brain tissues then recover gradually, but the duration and extent of recovery is uncertain, and few DTI studies have been done. Theλ⊥andλ‖changes versus time after RT, and the cycle time were not the same in our study, perhaps due to differences in the speed with which myelin and axons were repaired. The different degrees of demyelination and axonal injury should be considered as well. Because of the invasive nature of histologic sampling, it has not been possible to quantify CNS demyelination and axonal injury. Therefore, further research and long-term investigations are needed.Our results showed that FA in all five groups was significantly lower after RT than pre-RT. Radiation-induced transient NAWM changes consisting of small vessel injury (alterations in permeability of the small vessel wall), and result in edema and consequent demyelination. The transient reduction of FA might be attributed to demyelination and subsequent remyelination. In our study, minimal values of FA were observed at 0-3 months and again at 9-12 months after RT, with an interval of about 9.4 months. It is possible that FA cannot recover totally, or that our observation period was too short.There were no statistically significant changes in MD and ADC following RT. In our opinion, MD does not have sufficient sensitivity to detect diffusion changes in NAWM, possibly because MD is the average ofλ1,λ2,λ3.Another study confirmed our results, indicating that the sensitivity of MD is lower than that of FA.[CONCLUSION]Two-dimensional CSI, MR spectroscopy, and Diffusion Tensor imaging can provide non-invasive, quantitative methods to detect alterations of metabolism and structural changes in WM after RT. Metabolic alterations and water diffusion characteristics of NAWM in the temporal lobe of patients with NPC after RT were dynamic and transient. Variations over time were the same for NAA/Cho, NAA/Cr, and FA. Theλ⊥andλ‖changes reflected demyelination and axonal injury, and had different recovery times. The potential mechanism of these dynamic changes may consist of irreversible injury of small partial neurons, blood-brain barrier disruption and repair, demyelination and remyelination, and axonal injury and repair processes. |
