| Glioma is the most common primary brain tumors, mainly characterized by significant heterogeneity of nerve pathology, genetics and prognosis. It is very important for treatment and prognosis to make correct classification, grading diagnosis, comprehensive assessment of tumor extent and its relationship with the surrounding anatomical structure in the early stage. Magnetic resonance spectroscopy and perfusion imaging can provide information on metabolites and tumor angiogenesis in local tissue, which is useful for further understanding the nature of gliomas. The tumor site selected by either biopsy or imaging classification will affect the correct classification, because there is significant different degree of malignancy of glioma tumor in different parts of the glioma. In this study, in combination with relevant pathological and immunohistochemical features, MRI was employed to analyze perfusion characteristics and maximum perfusion metabolite changes of different classes or types of glioma, so as to further define the relationship of tumor MRI blood volume and metabolism changes with tumor microvessel density, microvessel morphological changes and cell proliferative activity.Purpose:The aim of the study was to study the perfusion and characteristics of magnetic resonance spectroscopy in the maximum perfusion region of brain gliomas and assess the value of this two techniques in diagnosis and grading of glioma, explore the relationship of cerebral blood changes and metabolism changes with pathological changes and identify further the possible pathological base of MRS and PWI by analyzing pathological and immunohistochemical features of different grades and histological types of gliomas.Materials and methods:1. Investigative subjects Gloma group:53 glioma patients confirmed by pathology, including 22 low-grade tumors subjects (WHO classification as gradeâ… ~â…¡: 13 patients with astrocytomas, 8 with oligodendroglioma and 1 with ependymoma), 31 high-grade tumors subjects (WHO grade wasâ…¢~â…£: 8 patients with anaplastic oligodendroglioma, 6 with anaplastic astrocytoma and 17 with glioblastoma). Control group: 10 volunteers.2. Scanning sequences and measuringMRI equipment was SIGNA HDx 1.5T scanner by GE (GE Medical System, Milwaukee, WI) and 8-channel (8NV) head coil. After conventional scanning, the raw data of FAIR and DSC perfusion (23 patients have scanned simultaneously FAIR) were acquired. Then, conventional T1WI enhanced sequences were scanned, multivoxel 2D spectroscopy were performed on the maximum perfusion section. In addition, 10 healthy volunteers underwent routine scan and multivoxel 2D spectroscopy scan of the basal ganglia section.Detections were performed on the values of DSC perfusion, the relative maximum cerebral blood volumes (rCBVmax), the relative cerebral flow of the same location of rCBVmax, the rCBV of peritumoral region, the rCBV of marginal zone of T2 hyperintensity respectively. The maximum cerebral blood flow (rCBFmax) of tumor and the rCBF of FAIR were analyzed with Spearman correlation analysis to determine their relationship. Mann-Whitney U test was used to analyze the differences between WHO subgroups or WHO grades. The ratios of Cho / Cr, Cho / NAA, NAA / Cr were measured in the regions of maximum perfusion, peritumor and contralateral normal brain parenchyma, which were then analyzed by Mann-Whitney U test to show the differences between WHO subgroups or WHO grades. Cho/Cr, Cho/NAA and NAA/Cr ratios were determined in cortex, putamen, thalamus and white matter of basal ganglia in normal volunteers and used as control. The metabolites of the symmetry areas in bilateral cerebral hemispheres compared by using Wilcoxon rank-sum test.3. Pathologic staining and indicesThe specimens were processed with routine hematoxylin-eosin (HE) staining and immunohistochemistry for Ki-67 and CD34. CD34 microvessel density and lumen diameter of the microvessel were measured, which were then analyzed by using the binary linear regression analysis to find their relation with rCBV. Ki-67 labeling indices were measured and analyzed using spearman correlation analysis to determine its relationship with Cho/Cr, Cho/NAA, NAA/Cr and Ki-67LI.4. Evaluation of each indexReceiver operating characteristic (ROC) curve analyses were used for analyzing significant difference upon rCBVmax in tumor, Cho/Cr, Cho/NAA, NAA/Cr in the regions of maximum perfusion to determine the optimum threshold value as well as sensitivity and specificity for this threshold value. Combined diagnosis with MR spectroscopy and perfusion weighted imaging were analyzed by the same way.Results:1. There was significant correlation between rCBF values of FAIR and that of DSC perfusion, with correlation coefficient of r=0.84(P<0.01). A significant difference was found on rCBFmax between high-grade glioma and low-grade glioma (Z=ï¹£3.941,P<0.01).DSC perfusion weighted imaging showed significant differences of rCBVmax between high-grade and low-grade glioma(P<0.01), but no significant difference betweenâ…¢andâ…£grade in high-grade glioma. There were no significant differences of rCBV in peritumor region, marginal zone of T2 hyperintensity between high-grade and low-grade glioma. While the rCBVmax of low-grade astrocytomas and low-grade oligodendroglioma showed no significant differences, but the low-grade oligodendroglioma were with more hyperperfusion than the former.The microvessel density, lumen diameter and lumen morphology of different grades and histological subtypes of gliomas were quite different, but binary linear regression and correlation analysis showed binary linear correlation of rCBVmax values with microvessel lumen diameter (X1) and microvessel density (X2). The equation is as follows: Y = 1.047+ 0.219X1 +0.055 X2, F = 10.33, P<0.01, R = 0.557, and t value: 3.91> 2.41.2. There were significant differences upon Cho/Cr and Cho/NAA of the metabolite in maximum perfusion of glioma between the high-grade and low-grade glioma, while no significant difference was found on NAA/Cr. The Lip in the maximum perfusion region of the high-grade gliomas occurred more frequently than that of the low-grade gliomas, but Lac was not significant difference between them. The metabolite ratio of peritumoral area showed no significant difference between the high-grade and low-grade gliomas. While significant difference was found between Cho/Cr, Cho/NAA and NAA/Cr in the maximum perfusion area, peritumoral area and contralateral normal white matter in the same grade. Immunohistochemistry showed that Ki-67 labeling indices representing tumor cell proliferation were significantly different between different grade gliomas and positively correlated with the Cho/Cr and Cho/NAA in the maximum perfusion region of glioma, with correlation coefficient r of 0.419 (P<0.01), 0.306 (P<0.05). There ws no significant difference in aspect of Cho/Cr, Cho/NAA and NAA/Cr between the bilateral cerebral hemispheres of 10 healthy volunteers(Pï¹¥0.05). The range of Cho/Cr was 0.45-4.12, and that of Cho/NAA was 0.29-1.27.3. The sensitivity and specificity of conventional enhanced T1WI for determining glioma grades were 77.40% and 45.50% respectively. When ROC curve analyses determined the optimum threshold value of rCBVmax for 2.95, the sensitivity and specificity for this threshold value were 90.32% and 54.55% respectively. When the optimum threshold values of Cho/Cr and Cho/NAA of maximum perfusion region for 2.08 and 3.37, the sensitivities for these threshold values were 89.29% and 57.14% and specificities were 47.37% and 89.47% respectively. Sensitivity and specificity of combined diagnosis were 100% and 47.37%. No significant difference was seen in respect to the joint area under the ROC curve either by a single index diagnosis or combined diagnosis, but PWI+MRS> rCBVmax> Cho/NAA> Cho/Cr, which showed that combined diagnostic value was better than any single indicator.Conclusions:1. RCBVmax value of the maximum tumor perfusion is correlated not only with microvessel density, but also with microvascular lumen diameter. Binary linear regression analysis showed that the microvessel lumen diameters exert greater impact on the blood volume and that changes of glioma microvessel density and microvascular lumen diameters are important pathological foundation for MR perfusion blood volume change. FAIR perfusion weighted imaging can provide perfusion information consistent with the DSC, which may be helpful for preliminary determination of tumor grade. However, signal to noise ratio of the FAIR sequence on 1.5T MRI is relatively rather low, which limits its application. DSC perfusion can provide a better signal to noise ratio and tumor rCBVmax value helps determine the grade of glioma. Compared with the low-grade astrocytomas, the low-grade tumor oligodendrocytes are more vulnerable to high perfusion. 2. Maximum perfusion reflects the active sites of tumor angiogenesis and more importantly the malignancy of the tumor tissue and can more accurately reflect the different levels of tumor metabolic characteristics. Considering the maximum perfusion as a spectral region of interest may have higher repeatability and reliability compared with previous one considering the enhanced area as a region of interest. Metabolites ratios of Cho/Cr and Cho/NAA of maximum perfusion region in the high- and low-grade gliomas shows significant difference and can be used for identification of high- and low-grade gliomas. Lip peak at maximum perfusion indicated tumor micronecrosis. Characteristic performance of the high-grade gliomas will help determine the tumor grade. Magnetic resonance spectroscopy metabolite ratios of Cho/Cr and Cho/NAA were positively related to tumor cell proliferation index Ki-67. The increase of two ratios suggested the pathological changes including strong tumor cell proliferation and higher degree of malignancy.3. Glioma rCBV values of peritumoral region shows no significant difference with spectral Cho/Cr and Cho/NAA values in the high- and low-grade gliomas, indicating that tumor perfusion and spectroscopic imaging are of low significance in identifying high- or low-grade gliomas. In the meantime, the rCBV values in same grade gliomas at different regions of interest show significant different with Cho/Cr and Cho/NAA values. When there exist difficulties for the conventional magnetic resonance imaging in defining the edema, magnetic resonance spectroscopy or perfusion weighted imaging can help edema delineation of the border, which is helpful for determining the extent of tumor invasion for tumor therapy.4. In diagnosis of high- and low-grade gliomas, magnetic resonance perfusion and magnetic resonance spectroscopy of maximum perfusion region have higher sensitivity and specificity compared with the conventional enhanced T1WI, with sound consistency with microvascular and tumor cell proliferation activity changes, which provides a pathological basis for combined diagnosis with MRI indicators. The joint use of spectrum and perfusion has higher sensitivity and specificity than any one single indicator. Combined use of multiple indicators can provide useful information for glioma treatment plan.
|
PDF Full Text Request