Relationship Between The White Matter Microstructural Dynamic Changes And The Neurocognitive Function Before And After Radiotherapy In Patients With Nasopharyngeal Carcinoma:a DTI-TBSS Study

Part one:A comparative study of neurocognitive function between patients with nasopharyngeal carcinoma before and after radiotherapy and healthy subjectsObjectiveTo evaluate dynamically the overall cognitive function of the patients with nasopharyngeal carcinoma （NPC） before and after radiotherapy （RT） at different time points by using the mini mental state examination （MMSE） scale, and to investigate the effects of RT on the neurocognitive status of the NPC patients and the change trend with time.Materials and Methods1. Subjects One hundred and twenty-eight subjects were investigated in this study, including100NPC patients and28healthy subjects matched for age, gender, and education.100patients [72males,28females, mean age43.73±9.88years （range19-65years）, years of education10.27±3.24（range6-16years）] with pathologically confirmed NPC were included in the study, with staging from TIN0M0to T4N2M0（Union for International Cancer Control, seventh edition,2009）. MRI follow-up time after RT ranged from1week to12years after completion of RT, and the median follow-up time was7.5months. All patients underwent a fractionated radiation therapy with three-dimensional conformal and intensity-modulated techniques （total dose/fraction dose/exposures,66-74Gy/1.8-2.0Gy/30-35times） for the first time. In every patient the same fractionation schedule was followed:one fraction per day, five fractions per week. All of the patients received two to four courses of concomitant chemotherapy during and after RT with one or more agents, such as cisplatin, fluorouracil and gemcitabine, depending on the clinical stage. Excluded from this study were patients with intracranial invasion, intracranial primary tumors or metastases, vascular lesions of the brain, head trauma and intracranial surgery, hypertension, heart disease, diabetes, hyperlipidemia, metabolic diseases, major psychiatric or neurological illness and other underlying disease that could possibly result in cognitive impairment. According to the international and domestic staging systems for classifying radiation-induced brain injury, we divided the NPC patients on the basis of the time before and after completion of RT into four subgroups:group1（pre-RT, or baseline, n=25）; group2（0-6months post-RT, n=25）; group3（>6-12months post-RT, n=25）; group4（>12months post-RT, n=25）. Control group included28healthy subjects,20males,8females, mean age43.57±11.75years （range24-62years）, years of education10.29±3.63（range6-16years）, who had no history of intracranial lesions, brain injury or neurocognitive deficits. All subjects were right-handed, native Chinese speakers. Each participant was fully informed about the purpose, methods, and precautions of the project, and formally agreed to participate and signed the informed consent form.2. Neuropsychological tests For all eligible participants, the Chinese version of the the Mini-Mental State Examination （MMSE） was used to assess their cognitive state in this study. MMSE was administered by a physician during the same period with the MRI examinations.3. Statistic analysisData were analyzed by using the Statistical Package for Social Sciences （SPSS for Windows, Version13.0, Chicago, IL, USA）. One-way analysis of variance （ANOVA） was performed to compare the differences in the age and years of education among different groups. Chi-square test was performed to assess the differences of gender among different groups. Cognitive performance was evaluated as a change in scores over time. The Kruskal-Wallis H test with multiple comparisons using Bonferroni post hoc test was used to determine the statistical differences of the MMSE scores across the different groups. A two-side P value<0.05was considered to be statistically significant.Results1. There was no significant difference in the age （F=0.908, P=0.462）, years of education （F=0.253, P=0.907） and gender （χ2=1.191, P=0.880） among the groups before and after RT at different time points in NPC patients and control group.2. There was no significant significantly difference in the MMSE scores （χ2=5.429, P=0.246） among the different groups.Conclusion1. No statistically significant decrease in MMSE scores over the period of follow-up was found after RT with three-dimensional conformal and intensity-modulated techniques compared with pre-RT in NPC patients and control group.2. The neurocognitive function after three-dimensional conformal intensity modulated radiotherapy for NPC patients showed dynamic change. The MMSE scores temporary decline occurred within6months after RT, but then will recover gradually, and will recovered to the level of the pre-RT after12month.3. Our results supports the safety of focal RT using three-dimensional conformal and intensity-modulated techniques in conventional fractionation （≤2Gy）, commonly prescribed total doses （66-74Gy） and exposures （30-35times） for treatment of nasopharyngeal carcinoma. The modern techniques may have no significant impact on overall neurocognitive performance of normal-appearing WM in NPC patients. Part two:White matter microstructural changes before and after radiotherapy in patients with nasopharyngeal carcinoma:a DTI-TBSS studyObjective1. To noninvasively investigate white matter （WM） microstructural dynamic changes before and after radiotherapy （RT） at different times by using diffusion tensor imaging （DTI） and tract based spatial statistic （TBSS） in comparison to the baseline and healthy subjects in nasopharyngeal carcinoma （NPC） patients.2. To explore whether white matter （WM） microstructural changes before and after RT at different times associate with neurocognitive outcome of NPC patients monitored using the Mini-Mental State Examination （MMSE）.Materials and Methods1. SubjectsOne hundred and seven subjects were investigated in this study, including81NPC patients and26healthy subjects matched for age, gender, and education.81patients [60males,21females, mean age44.26±9.91years （range19-65years）, years of education10.27±3.30（range6-16years）] with pathologically confirmed NPC were included in the study, with staging from T1N0M0to T4N2M0（Union for International Cancer Control, seventh edition,2009）. MRI follow-up time after radiotherapy ranged from1week to4years and7months after completion of radiotherapy, and the median follow-up time was7.5months. All patients underwent a fractionated radiation therapy with three-dimensional conformal and intensity-modulated techniques （total dose/fraction dose/exposures,66-74Gy/1.8-2.0Gy/30-35times） for the first time. In every patient the same fractionation schedule was followed:one fraction per day, five fractions per week. All of the patients received two to four courses of concomitant chemotherapy during and after RT with one or more agents, such as cisplatin, fluorouracil and gemcitabine, depending on the clinical stage. Excluded from this study were patients with intracranial invasion, intracranial primary tumors or metastases, vascular lesions of the brain, head trauma and intracranial surgery, hypertension, heart disease, diabetes, hyperlipidemia, metabolic diseases, major psychiatric or neurological illness and other underlying disease that could possibly result in cognitive impairment. According to the international and domestic staging systems for classifying radiation-induced brain injury, we divided the NPC patients on the basis of the time before and after completion of RT into four subgroups:group1（pre-RT, or baseline, n=23）; group2（0-6months post-RT, n=21）; group3（>6-12months post-RT, n=20）; group4（>12months post-RT, n=17）. Control group included26healthy subjects,18males,8females, mean age43.73±11.67years （range24-62years）, years of education10.23±3.49（range6-16years）, who had no history of intracranial lesions, brain injury or neurocognitive deficits. All subjects were right-handed, native Chinese speakers. Each participant was fully informed about the purpose, methods, and precautions of the project, and formally agreed to participate and signed the informed consent form.2. Neuropsychological testsFor all eligible participants, the Chinese version of the the Mini-Mental State Examination （MMSE） was used to assess their cognitive state in this study （the same as the first part）. MMSE was administered by a physician during the same period with the MRI examinations.3. DTI data acquisitionAll MR imaging data were acquired using a3.0T GE clinical scanner （SIGNA EXCITE GE Medical Systems, Milwaukee, WI, USA） with an eight-channel head coil. The routine MRI brain protocol including axial T1-weighted images （TR/TE, 600/15ms）, T2-weighted images （TR/TE,5200/140ms）, and T2-weighted fluid attenuated inversion recovery （TR/TE/IR,9000/120/2100ms） was obtained for every subject to detect intracranial lesions. DTI scans were performed employing a single-shot echo-planar imaging sequence and array spatial sensitivity encoding technique with the following parameters:repetition time （TR）12000ms, echo time （TE）75.5ms, field of view （FOV）24×24cm, matrix128R128, slice thickness3mm with no interslice gap, NEX=1, flip angle90°. Images were collected along25non-collinear diffusion gradient directions with a b-value of1000sec/mm2, and one set of null images with b=0sec/mm2was acquired.4. DTI data processing and TBSS analysisDTI data was analyzed by using FSL （FMRIB Software Library, www.fmrib.ox.ac.uk/fsl, version4.19） tools. The procedure of DTI data processing included:First, diffusion tensor images were corrected for head movement and eddy current distortion by using FDT tool of the FSL software. Second, mask image for each brain was created by using each subject’ B0image with BET tool of the FSL software. Then, the diffusion tensor was calculated on a voxel-by-voxel basis by using dtifit. Maps of fractional anisotropy （FA）, mean diffusivity （MD）, axial diffusivity （λ‖=λ1） and radial diffusivity [λ⊥=（λ2+λ3）/2] were obtained. And then follow the tract based spatial statistic （TBSS） processes, first running the next steps: tbss_1preproc, tbss_2reg, tbss_3postreg, tbss_4prestats. Subsequently, threshold-free cluster enhancement （TFCE） in randomise was used to perform the multisubject analysis of FA, MD, λ‖and λ⊥respectively, permutation-based correction for multiple comparisons at P<0.05.5. Statistic analysisPatial correlation analysis with age and years of education as covariates was performed to investigate the underlying relationship between the patients’MMSE scors and the mean FA, MD, λ‖and λ⊥values in significantly different areas that revealed by pre-RT vs0-6months post-RT in NPC patients. A two-side P value less than0.05was considered as significant correlation.Results TBSS results1. FA valuesCompared with pre-RT group, the mean FA values in the right frontal lobe, right parietal lobe and right occipital lobe white matter reduced significantly in post-RT0-6m group （P<0.05）, then increased gradually. Until one year after RT （group4）, the FA level in the right frontal lobe and right parietal lobe white matter had remained significantly lower than that in the pre-RT group, while the FA level in the right occipital lobe white matter was slightly lower than pre-RT, but was not significantly different.Compared with pre-RT group, the mean FA values in the right cerebellum and left parietal lobe white matter increased significantly in post-RT0-6m group （P<0.05）, then reduced gradually. Until one year after RT （group4）, the FA level in the right cerebellum had remained significantly higher than that in the pre-RT group, while the FA level in the left parietal lobe white matter had been restored and was not significantly different from that in the pre-RT group.2.λ‖valuesCompared with pre-RT group, the λ‖values in significantly different areas in the bilateral occipital lobe, temporal lobe and parietal lobe white matter increased significantly in post-RT0-6m group （P<0.05）, then reduced gradually. Until one year after RT （group4）, the λ‖level in the bilateral temporal lobe, left occipital lobe, part of the bilateral parietal lobe white matter was still significantly higher than that in the pre-RT group; while the λ‖level in the right occipital lobe and rest of the bilateral parietal lobe white matter was close to the values in the pre-RT group and there was no significant difference.3.λ⊥valuesCompared with pre-RT group, the mean λ⊥values in the right occipital lobe, left frontal-parietal junction area and left parietal lobe white matter increased significantly in post-RT0-6m group （P<0.05）, then reduced gradually. Until one year after RT （group4）, the λ⊥level in the left parietal lobe white matter had remained significantly higher than that in the pre-RT group; while the λ⊥level in the right occipital lobe and left frontal-parietal junction area white matter was close to the values in the pre-RT group and was not significantly different.Compared with pre-RT group, the mean λ⊥values in the right cerebellum and left centrum semiovale reduced significantly in post-RT0-6m group （P<0.05）, then increased gradually. Until one year after RT （group4）, the λ⊥level in the above brain areas had remained significantly lower than that in the pre-RT group.4. MD valuesCompared with pre-RT group, the MD values in significantly different areas in the right occipital lobe, bilateral temporal lobe, the right occipital-temporal junction area, the left parietal lobe, the left centrum semiovale, the left frontal-parietal junction area white matter increased significantly in post-RT0-6m group （P<0.05）, then reduced gradually. Until one year after RT （group4）, the MD level in the right occipital lobe white matter was still significantly higher than that in the pre-RT group; while the MD level in the rest brain areas mentioned above was close to the values in the pre-RT group and there was no significant difference.Correlation resultsFA value in the right occipital （r=0.483, P=0.027）, and left parietal （r=0.445, P=0.034） white matter that revealed by groups comparison between pre-RT vs post-RT0-6m correlated positively with MMSE score of NPC patients. λ‖value in the right parietal white matter （r=0.498, P=0.025） correlated positively with MMSE score of NPC patients. λ‖value in the left frontal-parietal junction area white matter （r=-0.482, P=0.027） correlated negatively with MMSE score of NPC patients. MD value in the left frontal-parietal junction area （r=-0.487, P=0.025） correlated negatively with MMSE score of NPC patients. MD value in the left parietal （r=0.514, P=0.035） correlated positively with MMSE score of NPC patients.Conclusion1. As the only noninvasive method for characterizing changes in tissue microstructural organization and diffusivity of water changes in brain tissue in vivo, DTI may be used in sensitive discovering and dynamical monitoring the subtle abnormalities of the radiation-induced WM microstructure, earlier than conventional cranial MRI, following RT for NPC patients.2. For NPC patients, radiotherapy may produce extensive white matter microstructure damage, involving multiple brain regions and not just limited to the temporal lobe. Furthermore, WM microstructure abnormalities are complex, dynamic and temporary, brain tissue injury is most serious within6months after radiotherapy, then will repair gradually, and will be close to the level of the pre-RT after12month.3. The FA value decreased significantly, and λ⊥, MD,λ‖value increased significantly in multiple WM regions within6months after RT compared with pre-RT in NPC patients, suggesting that WM microstructural damage in these regions may be caused mainly by brain edema and demyelination, and may be accompanied by mild axonal degeneration.4. The mean FA, MD, λ‖and λ⊥values in significantly different areas that revealed by pre-RT vs0-6months post-RT in NPC patients were correlated significantly with their MMSE score. It indicates that extensive WM microstructure damage after radiotherapy affects the patient’s cognitive function, and DTI in vivo measurement of WM microstructure changes may provide sensitive imaging biomarkers for early evaluation of radiation-induced brain injury and cognitive dysfunction.