Assessment Of Perfusion Ct(PCT) Parameters As Imaging Biomarkers In Glioma Grading

Background and purpose:Perfusion imaging using CT can provide additional information about tumour vascularity and angiogenesis in different glioma grades. The purpose of our study was to investigate the morphologic characteristics of supratentorial gliomas at different tumoural regions using PCT parameters and to quantitative estimate these imaging parameters for the assessment of the usefulness of PCT in glioma grading.Materials and Methods:PCT was performed in40patients with previously untreated glioma (15low-grade gliomas and25high-grade gliomas) by using a total acquisition time of210seconds on a256-slice CT scanner. Absolute values of PCT parameters values were used when determining the morphologic characteristics of gliomas whilst normalised ratios of the PCT parameters (relative cerebral blood volume [rCBV], relative cerebral blood flow [rCBF], relative mean transit time [rMTT], relative permeability surface area product [rPS]) were used for the final quantitative analysis. The interrelationship of the parameters at different tumoural regions and their diagnostic value in the grading of glioma were assessed. Receiver operating characteristic (ROC) and Student t-test analyses were obtained to evaluate the parameters. Statistical analyses were also performed in astroglial tumours (35patients) after excluding all the oligodendroglial tumours.Results:(1) Mean values of CBV, CBF and PS of HGGs (Grade III and Grade IV) were found to be significantly higher than that of LGGs (Grade II)(p<0.01). No significant difference was observed between the mean of absolute values of Grade III and Grade IV. For most heterogeneous HGGs the following relationship was observed between the absolute values of the respective CBV, CBF and PS parameters:Tumour peripheral region>Tumour central parenchyma region> Contralateral apparent normal white matter (ANWM)>Peritumoural oedema。(2) ROC analyses showed that both rCBV (C-statistic0.972) and rPS (C-statistic0.977) were very similar to each other in differentiating LGGs from HGGs and had higher predictability compared with rCBF and rMTT. rPS showed the highest C-statistic values and hence highest predictability when differentiating Grade II from Grade III and Grade Ⅱ from Grade IV.(3) In astroglial tumour group, differences in rCBV, rCBF and rPS between low-grade and high-grade were statistically significant (p<0.01), with the low-grade tumour group showing a lower mean values than the high-grade group. ROC analyses showed that PS (C-statistic0.983) had highest predictability in differentiating low-and high-grade astroglial as compared to CBV (C-statistic0.967) and CBF (C-statistic0.920).(4) Cut points of>2.37for rCBV (92%sensitivity and100%specificity),>2.31for rCBF88%sensitivity and93.3%specificity), and>3.46for rPS (84%sensitivity and100%specificity) were found to identify the HGGs.Conclusion:PCT can be used for preoperative classification of low-grade and high-grade gliomas as well as astroglial tumours. Both rCBV and rPS proved to be beneficial parameters and hence good imaging biomarkers correlating with glioma grades and providing different valuable information about the tumour biology and haemodynamics. PS having stronger predictability showed slightly better association with glioma grading as compared to CBV and CBF. BACKGROUND and PURPOSE:Glioma angiogenesis and cellular/microvascular proliferation are key features for the progression of tumours. In vivo correlation of perfusion imaging parameters with angiogenic and proliferation markers can help in better understanding the role of perfusion imaging as an imaging biomarker and hence in grading of gliomas. The purpose of this study was to correlate PCT parameters with angiogenic marker vascular endothelial growth factor (VEGF) and proliferation marker Ki-67in gliomas.MATERIALS AND METHODS:PCT was performed on40patients with previously untreated glioma (15low-grade gliomas and25high-grade gliomas) by using a total acquisition time of210seconds on a256-slice CT scanner. We correlated VEGF, Ki-67and WHO grade of the cytoreductive specimens with the PCT parameters (cerebral blood volume, CBV; cerebral blood flow, CBF; Permeability surface area product, PS; Mean transmit time, MTT). Immunohistochemical examinations for hematoxylin-eosin stain, vascular endothelial growth factor (VEGF) expression and Ki-67antigen were performed and analysed under the microscope. Angiogenic marker (VEGF) and proliferation marker (Ki-67) were correlated with perfusion parameters corresponding to the tumour central parenchymal and tumour peripheral regions. Pearson correlation coefficients and multiple regression analyses were performed to assess these correlations. These statistical analyses were also performed in astroglial tumours (35patients) after excluding all the oligodendroglial tumours. RESULTS:When using Pearson correlation and multiple regression analyses, rCBV, rCBF and rPS at both the tumour central and tumour peripheral regions showed statistically significant correlation with WHO grade (p<0.001). With Pearson correlation analysis, rCBV and rPS at both central and peripheral regions, showed positive correlation with VEGF. However, with regression analysis only rPS was statistically significant (p<0.001) and not rCBV; suggesting that rPS is a better predictor for VEGF than rCBV. rPS showed a positive correlation coefficient of r=0.662, p<0.001at the central part and r=0.643, p<0.001at the peripheral region. With Pearson correlation both rCBV and rPS at the central and peripheral regions seemed to correlate positively with Ki-67(p<0.05); however, neither rCBV nor rPS was statistically significant with regression analysis at the0.05level.CONCLUSIONS:rPS is a better predictor of VEGF as compared to rCBV. rPS can help for evaluation of tumour permeability which is more prominent in high-grade gliomas thus helping in glioma grading. rCBV and rPS can also provide additional information about cellular and microvascular proliferation. PCT parameters, rCBV and rPS proved to be important imaging biomarkers in the grading of gliomas.Therefore, PCT can be used as a non-invasive in vivo alternative/adjunct to histological evaluation and may be used to assess tumour angiogenesis.