The Value Of Perfusion CT For The Reflection Of Angiogenesis In Brain Tumors: Animal Experiment And Clinical Study

Objective: To establish a stable implanted model of VX₂ rabbit brain tumor and study its pathological features. To study the perfusion CT appearance of experimental rabbit brain tumors and validate the perfusion CT method for the reflection of angiogenesis in VX₂ rabbit brain tumors, with correlation to MVD and VEGF. To evaluate the clinical value of perfusion CT in grading and differentiating brain tumors and provide more functional imaging information. To comare the diagnosing efficiency of absolute parameters with those of relative parameters. To comare advantages and disadvantages between perfusion CT and perfusion MRI.Part one: Study on perfusion CT of rabbit VX₂ brain tumor modelMethods and materials: Thirty New Zealand White rabbits which weighted 2³ kg were used in the experiment. They were implanted with 100μl viable VX₂ tumor cells (10⁷/ml) through a hole 5 mm to the right of the sagittal suture and 5 mm posterior to the coronal suture bored by a dental drill. MRI was performed every two days after 10 days of implantation to evaluate the growth of the tumor. Coronal pre- and post-contrast T₁WI (TR 500 ms, TE 20 ms), and T₂WI (TR 4000 ms, TE 90 ms) were performed using a 3 inches surface coil, thickness 2 mm, FOV 8 cm×4 cm , matrix 512×256. The appearance of a nodular enhancement was thought to be the sign of tumor growth.Twenty New Zealand White rabbits with tumor size over 3 mm in diameter were randomly divided into 2 groups according to the tumor growth time with those equal to or less than 3 weeks as group 1 and those more than 3 weeks asgroup 2, and perfusion CT were performed accordingly. Eight rabbits were chosen to perform perfusion CT twice for the auto-control study of changes of parameters in tumors with time. Eight rabbits performed perfusion CT before and after the implantation for the auto-control reproducibility study of the examination. Perfusion CT scans were performed with a GE Hi speed CT/i scanner. Firstly, 1-2 ml OmnipaqueC 300 mgl/ml) was injected intravenously through a hindlimb vein, and 3-mm-thick coronal scans were performed at lmm intervals to select the slice passing the biggest part of the tumor. Then 50 cine mode images were acquired before, during and after intravenous injection of 1.5 ml/kg Omnipaque (300 mgl/ml) at a speed of 0.4 ml/s. Technique parameters were as follows: 80 kV, 80 mA, 1 scan/s, FOV 9.6 cm*9.6 cm.CT measurements of CBV, CBF and PS from tumor, peritumor and contralateral normal tissue regions were obtained using GE perfusion 2 software, and corresponding functional maps were constructed. ROIs were placed as follows : a 36-64 pixel ROI was placed within tumor with highest CBV as a tumor ROI, a copyed ROI was placed to contralateral cerebral hemisphere symmetrically as a normal tissue ROI, a 24-48 pixel ROI was placed near the tumor as a peritumor ROI. After that the animals were sacrificed, and 2% Evans blue (2 ml/kg) were given intravenously in 16 of these animals one hour prior to sacrifice to detect breakdown of the BBB. A semiquantitative scale regarding the leakage of Evans blue was used: grade 0, none; grade 1, slight (barely visible blue coloration) ; grade2, moderate (easily detectable stain) ; grade3, severe (intense blue staining). VEGF expression rate and MVD (stained by factor VI related antigen) were evaluated in immunohistochemical examination of the specimens.Statistical analysis was performed using SPSS 11.0 software package. Mean ± SD was obtained. Student t test and One-Way analysis-of-variance (ANOVA) test were used to compare the parameters. Linear correlation analysis was used tocompare correlations between CBV, CBF, PS and MVD. Spearman's rank correlation was used to compare correlations between CBV, CBF, PS and VEGF, and between PS and blue stain. Statistical significance was declared at a level of 0.05.Results: One of the rabbits dead of anesthetic accident and 7 dead of intracranial infection because of contamination of the cell suspension. Of the 22 rabbits included into the study, tumor grew in 20 rabbits which can be seen clearly on MR imaging. Pathological examination showed characteristics of squamous carcinoma. VEGF was expressed in all tumors with the mean rate of positive cells being (52.51±19.15)% (19.5%⁹2.9%). Mean MVD was 51.3±14.42 (25—81). Evans blue extravasation was observed in all the 16 rabbits with different degree (grade 1 in 1 rabbit, grade 2 in 5 rabbits, and grade 3 in 10 rabbits). Using pearson's linear correlation analysis, positive correlation was found between tumor growth time and volume (r=0.791, />=0.000), MVD and tumor growth time (r=0.875, P=0.000), MVD and tumor volume (r=0.901, /MX000), respectively. Spearman's rank correlation analysis showed positive correlation between VEGF grade and blue stain of the tumor (^=0.594, P=0.015).Tumors can be outlined clearly on functional maps. The mean value of CBV, CBF and PS in tumor region were significantly higher than those of peritumor and normal tissue regions (.P=0.000). Tumor CBV, CBF and MVD in group 2 were significantly higher than those in group 1 (PO.01). The auto-control study of 8 rabbits with tumors in different times 2 weeks apart showed that all the CBV, CBF, and PS in tumors were significantly higher in the second measurements than those in the first measurements (PO.05). We found a significant linear correlation between MVD and CBV (r=0.915, P=0.000), MVD and CBF (r=0.901, P=0.000), and MVD and PS (r=0.459, P=0.042). We also found a rank correlation between PS and blue stain of tumor (^=0.861 , P=0.000). In the auto-controlreproducibility study of the 8 rabbits, the CBV, CBF, and PS of the contralateral normal tissue showed no obvious changes before and after the implantation (P>0.05).Part two: The value of perfusion CT for diagnosis of supertentorial intra-axial mass lesions and comparison with perfusion MR imaging Methods and materials: Perfusion CT was performed on 43 cases of supertentorial intra-axial mass lesions. Twenty-six were male and 17 were female, aged 14⁷2y (averaged 47.ly). All the 27 cases of tumors of neuroepithelial tissue were confirmed surgically and pathologically. According to WHO classification 8 cases were grade II, 8 cases were grade III and 11 cases were grade IV. Among the 10 cases of metastasis 4 cases were confirmed surgically and pathologically and 6 cases were confirmed clinically by the definite history of other primary tumors and typical imaging features of metastasis. Among the 6 cases of brain abscess 2 cases were confirmed surgically and pathologically and 4 cases were confirmed clinically by the absorption of the lesions through antibiotic therapy in follow up study. Among the cases mentioned above 14 cases of tumors of neuroepithelial tissue and 8 cases of metastasis performed perfusion MR imaging.Perfusion CT were performed on a GE Discovery Lightspeed Qx/4i multi-slice spiral scanner. The scan parameters were as follows: 80 kV, 190 raA, matrix 512x512, FOV 23 cm x 23 cm, thickness 5 mm (4 slices per rotation), 1 s per revolution, total scan time 45, 180 images were aquired. A total of 45 ml contrast agent (Ultravist 300) was injected intravenously through a forearm vein by a power injector at a rate of 4 ml/s. The perfusion 2 software was used to create CBV, CBF, and PS maps. ROIs were positioned within lesion, in surrounding edema region, and in contralateral white matter region and their CBV, CBF, and PS values were recorded. The ratio of parameters of lesion and edema to that ofcontralateral white matter were recorded as rCBV, rCBF and rPS.Perfusion MRI were performed on a GE signa 1.5T Twin-speed infinity with excite I unit. Using a single shot gradient recalled echo echoplanar imaging T2*WI sequence. TR 2000ms, TE 80 ms, flip angle 90°, bandwidth 62.5, FOV 24 cm*24 cm, thickness 6.0 mm, matrix 128x128, Nex=l. Atotal of O.lmmol/kg Gd-DTPA was injected intravenously through a forearm vein by a power injector at a rate of 3 ml/s, and then 20 ml isotonic saline was injected to wash the pipe. Functool 2 software was used to create rCBV maps.Rusults: Among the 37 supertentorial intra-axial tumors 10 involed the frontal lobe, 4 involved the parietal, corpus callosum, temporal lobe, and frontal-temporal lobe, respectively, 2 involved parietal-occipital lobe and temporal-parietal-occipital lobe, respectively, 1 involved occipital lobe, frontal-parietal lobe and thalami, respectively, and 4 involved multiple areas with multiple lesions. Among the 6 cases of brain abscess 3 involved parietal lobe and 1 involved frontal, temporal and temporal-parietal-occipital lobe, respectively. Among the 8 cases of grade II tumors of neuroepithelial tissue, mild edema was found in 6 cases; mild and moderate enhancement was found in 1 cases respectively. Among the 8 cases of grade III tumors of neuroepithelial tissue, mild edema was found in 2 cases, and moderate edema in 4 cases; mild enhancement was found in 2 cases, moderate enhancement was found in 1 cases, and obvious enhancement was found in 3 cases; cystic changes were found in 3 cases. Edema, enhancement and necrosis or cystic changes can be seen in all cases of the 11 grade IVtumors of neuroepithelial tissue. Edema could be found in all the 10 metastasis, and 8 of them showed ring-like enhancement with cystic necrosis in the central part of the lesions; 2 cases of them showed nodular enhancement. Amonga the 6 brain abscess edema was found in all cases, and ring-like enhancement was found in 5 cases, and nodular in 1 cases.CBV and CBF of lesions of the grade II tumors of neuroepithelial tissue were lower than that of grade III with statistical significance (PO.05); CBV, CBF and PS of lesions of the grade II and III were lower than that of grade IVwith statistical significance (PO.05).rCBV, rCBF and rPS of lesions of the grade II tumors of neuroepithelial tissue were lower than that of grade III with rCBF having statistical significance (PO.05), and rCBV and rPS having no statistical significance (P>0.05). rCBV, rCBF and rPS of lesions of the grade II and III were lower than that of grade IVwith statistically significance (PO.05).The threshod of CBV > CBF and PS of lesions for diagnosis of grade IV and exclusion of grade II and III were 3.98 (sensitivity 90.9%, specificity 87.5%), 99.02 (sensitivity81.8 % , specificity81.2 % ), and 13.64 (sensitivity90.9 % , specificity87.5%); for diagnosis of grade Hand exclusion of grade III and IV were 2.53 (sensitivity87.5 % , specificity 100 % ), 59.50 (sensitivity87.5 % , specificity94.7%)and 11.95 (sensitivity87.5%, specificity68.4%).CBV and CBF of lesions of high grade tumors of neuroepithelial tissue and metastasis were higher than that of brain abscess with statistical significance (P<0.05); No statistical significance was found between the CBV and CBF of lesions of high grade tumors of neuroepithelial tissue and metastasis (P>0.05). PS of lesions also showed no statistical significance among the three kind of intra-axial masses.rCBF of lesions of high grade tumors of neuroepithelial tissue were statistically higher than that of metastasis and brain abscess (P<0.05).CBV and CBF of surrounding edema region of high grade tumors of neuroepithelial tissue were statistically higher than that of metastasis and brain abscess (PO.05). No statistical significance was found between the CBV and CBF of surrounding edema region of metastasis and brain abscess (P>0.05).rCBV and rCBF of surrounding edema region of high grade tumors of neuroepithelial tissue were statistically higher than that of metastasis and brain abscess (PO.05). No statistical significance was found between the rCBV and rCBF of surrounding edema region of metastasis and brain abscess (P>0.05). rPS of surrounding edema region also showed no statistical significance among the three kind of intra-axial masses(P>0.05).The threshod of CBV, CBF and rCBF of lesions for diagnosis of brain abscess and exclusion of high grade tumors of neuroepithelial tissue and metastasis were 3.12 (sensitivity 100%, specificity 82.1%), 64.52 (sensitivity 83.3%, specificity 85.7%), and 2.91 (sensitivity83.3 %, specificity 100%), respectively.The threshod of CBV, rCBV and rCBF of surrounding edema for diagnosis of high grade tumors of neuroepithelial tissue and exclusion of brain abscess and metastasis were 0.74 (sensitivity 76.5%, specificity 73.3%), 0.86 (sensitivity 76.5%, specificity 86.7%), and 0.74(sensitivity76.5%, specificity 80.0%), respectively; For diagnosis of metastasis and exclusion of high grade tumors of neuroepithelial tissue were 0.79 (sensitivity 77.8%, specificity 70.6%^ 0.75 (sensitivity 77.8%, specificity 82.4%), and 0.74 (sensitivity 88.9%, specificity 76.5%), respectively.No obvious differences were found between the areas under ROC curves of absolute parameters and relative parameters in grading tumors of neuroepithelial tissue and differentiating supertentorial intra-axial mssses.Perfusion MR showed that the rCBV of surrounding edema of high grade tumors of neuroepithelial tissue were higher than that of metastasis with statistical significance (P<0.05).The threshod of rCBV of surrounding edema showed with perfusion MR for diagnosis of high grade tumors of neuroepithelial tissue and exclusion of metastasis was 0.65 (sensitivity 85.7% , specificity 100%).Conclusions:1. A stable model of VX2 rabbit brain tumor has been established with the method of skull drilling. The method was simple and easy to use, with a high tumor growth rate and remarkable angiogenesis. The model is helpful for the study of tumor angiogenesis pathologically and radiologically.2. Perfusion CT can distinguish tumor from peritumor and normal tissue clearly, reflect tumor angiogenesis accurately , and provide useful information for the evaluation of brain tumor.3. Perfusion CT is helpful for grading tumors of neuroepithelial tissue. The threshod of CBV, CBF and PS of lesions for diagnosis of grade IV and exclusion of grade II and III were 3.98, 99.02, and 13.64; for diagnosis of grade Hand exclusion of grade III and IV were 2.53, 59.50 and 11.95.4. Perfusion CT is of value in differentiating supertentorial intra-axial mssses. The threshod of CBV, CBF and rCBF of lesions for diagnosis of brain abscess and exclusion of high grade tumors of neuroepithelial tissue and metastasis were 3.12, 64.52, and 2.91, respectively. The threshod of CBV, rCBV and rCBF of surrounding edema for diagnosis of high grade tumors of neuroepithelial tissue and exclusion of brain abscess and metastasis were 0.74, 0.86 and 0.74, respectively; for diagnosis of metastasis and exclusion of high grade tumors of neuroepithelial tissue were 0.79, 0.75 and 0.74, respectively.5. There were no obvious differences between the areas under ROC curves of absolute parameters and relative parameters in grading tumors of neuroepithelial tissue and differentiating supertentorial intra-axial mssses.6. The threshod of rCBV of surrounding edema showed with perfusion MR for diagnosis of high grade tumors of neuroepithelial tissue and exclusion of metastasis was 0.65. The value of rCBV of surrounding edema showed with perfusion MR in differentiating high grade tumors of neuroepithelial tissue from...