European Journal of Radiology 83 (2014) 185–190

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

Prognostic indices for cerebral venous thrombosis on CT perfusion: A prospective study Rakesh Kumar Gupta a,∗ , J.R. Bapuraj b , N. Khandelwal c , Dheeraj Khurana d a

Department of Radiodiagnosis and Imaging, MMIMSR, Mullana, Ambala, India Department of Radiology, Division of Neuroradiology, University Hospital, University of Michigan, 1500 E Medical Center Drive, Ann Arbor, MI 48109, United States c Department of Radiodiagnosis and Imaging, Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, India d Department of Neurology, Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, India b

a r t i c l e

i n f o

Article history: Received 4 June 2013 Received in revised form 31 August 2013 Accepted 22 September 2013 Keywords: Computed tomography (CT) Relative cerebral blood flow (rCBF) Relative cerebral blood volume (rCBV) Relative mean transit time (rMTT) Cerebral venous sinus thrombosis (CVST)

a b s t r a c t Purpose: We determined the prognostic significance of CT perfusion characteristics of patients with cerebral venous sinus thrombosis (CVST) and assessed the change in perfusion parameters following anticoagulation therapy. Materials and methods: 20 patients with CVST diagnosed on non-contrast computed tomography (NCCT), magnetic resonance imaging (MRI), and magnetic resonance venography (MRV) were included in this study. The initial CT perfusion study was performed at the time of admission. The following perfusion parameters: relative cerebral blood flow (rCBF), relative cerebral blood volume (rCBV), and relative mean transit time (rMTT) were calculated in the core and periphery of the affected area of the brain. Follow-up CT perfusion studies were performed at 1 month following anticoagulation therapy and the perfusion parameters thus obtained were compared with pre-treatment results. Receiver operating characteristic (ROC) curve analysis was performed to determine the prognostic significance of perfusion parameters. Results: All patients in this study showed areas of hypoperfusion on CT perfusion. To determine the favorable clinical outcome on basis of perfusion parameters, ROC curve analysis was performed which showed that the optimal threshold for rCBF > 60.5%, rCBV > 75.5%, and rMTT < 148.5% correlated with better clinical outcomes. Post treatment perfusion parameters showed significant correlation in core of the lesion (p < 0.05) than in the periphery. Conclusion: CT perfusion studies in CVST are a good prognostic tool and yield valuable information regarding clinical outcome. © 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction The reports in the literature highlight variations in the outcome after cerebral venous sinus thrombosis [1–5]. There are no standardized clinical parameters or imaging criteria which can reliably predict clinical outcome of patients of CVST Magnetic resonance (MR) perfusion [6], Xenon computed tomography (CT) [7] and Single Photon emission computed tomography (SPECT) [8,9] have been used to evaluate cerebral perfusion in CVST. However, all of these techniques have significant drawbacks such as limited availability and patient discomfort. There is, therefore a need for a rapid, readily available technique which can identify and quantify the presence and extent of a perfusion deficit

∗ Tel.: +91 9729292317. E-mail addresses: [email protected] (R.K. Gupta), [email protected] (J.R. Bapuraj). 0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.09.027

in CVST. Perfusion CT is an easily available, reproducible quantitative imaging technique that allows rapid evaluation of cerebral circulation. The present study was performed using CT perfusion in patients of CVST to characterize the perfusion deficits and evaluate its potential as a prognostic marker. 2. Material and methods This study was approved by our institute’s ethics committee. Written informed consent was obtained from the patients or their families before the patients were enrolled in the study. 2.1. Design This prospective study was designed to assess the prognostic value of CT perfusion characteristics in predicting the clinical outcome and to assess the changes after the anticoagulation therapy in CVST.

186

R.K. Gupta et al. / European Journal of Radiology 83 (2014) 185–190

Patients presenting to emergency services of our hospital with clinical suspicion of CVST were evaluated by noncontract CT (NCCT) head and MRI of the brain. The diagnosis of CVST was confirmed on MRV examination. Confirmed cases of CVST were subjected to CT perfusion study before initiation of anticoagulation therapy. Second assessment with CT perfusion was done one month after the initiation of anticoagulation therapy.

2.2. Patient selection Twenty patient of confirmed CVST were included in this study irrespective of age and sex. Patients with history of prior stroke, allergy to iodinated contrast media and conditions which can alter the perfusion parameters like arteriovenous malformations, invasion of the dural sinus by the adjacent tumor, circulatory low flow states resulting from the blood volume depletion, dehydration, cardiac disease, and thyroid disease were excluded from this study.

2.3. Data acquisition Non-contrast head CT was done on four slice multidetector row CT scanner (Lightspeed; GE Medical System, Milwaukee, Wisconsin, USA). MRI was performed on a 1.5 T scanner (Magnetom Vision, Seimens Medical System, Erlangen, Germany). The MR examination included T1WI (axial and sagittal) [TR 510 ms, TR: 14 ms, Flip angle 90◦ ], T2 WI (axial) [TR: 2735 ms, TE: 102 ms, Flip angle 90◦ ], T2 FLAIR (coronal) [TR 32–40 ms, TE: 8–12 ms IT 1700–2200 ms], and 2D time of flight MRV [TR 32–40, TE: 8–12 ms, Flip angle 50–70◦ , slice: 1.5–3 mm matrix 144 × 256, NEX 2]. CT perfusion scans were obtained on fourslice multidetector row scanner (Lightspeed; GE Medical system, Milwaukee, Wisconsin, USA). After the unenhanced CT of the brain, four contiguous 5 mm thick sections were selected in the abnormal area of parenchymal abnormalities detected on CT/MRI of the brain. A total of 50 mL of nonionic contrast was injected at 5 mL/s, with a 5 s preparation delay before the acquisition of images in cine mode for 50 s at 4 images per second per rotation with scan parameters: slice thickness 5 mm at an interval of 0 mm, 120 kVp, 200 mAs and matrix size 512 × 512. The calculated CT effective dose was 10 mSv and the absorbed dose to the brain was about 349 mGy.

groups were compared. Prognostic indices based on CT perfusion parameters were retrospectively evaluated. 2.4. Statistical analysis Statistical analysis was done by applying paired-t test and Pearson’s correlation. Prognostic indices based on CT perfusion parameters were retrospectively evaluated by using receiver operating characteristic (ROC) curve. The area under the ROC curve (AUC) and first derivate of the ROC curve corresponding to the best perfusion parameter were determined. Unpaired t-test and binary logistic regression analysis was used to correlate the parameters with clinical outcome. All statistical analysis was performed by using commercially available software (SPSS 10.0). 3. Results Out of the 20 patients included in our study, 3 were male (15%), and 17 were female (85%). The mean age was 29.6 years, with an age range from 21 to 59 years. Twelve of the patients in this study were postpartum. Most common presenting clinical symptoms was headache (n = 17). 3.1. Cross sectional imaging Sixteen patients (80%) NCCT and MRI showed evidence for a venous infarct with hemorrhage. 3 (15%) patients showed ahypodense area suggestive of ischemia or vasogenic edema without hemorrhage. One patient (5%) had a normal NCCT, without discernable abnormal parenchymal areas of hemorrhage or low attenuation suggesting vasogenic edema or ischemia. The MRI, however, was suggestive of venous infarct with a well-defined area of T1 hypointensity with corresponding T2/FLAIR hyperintense signals. All the patients had diagnostic quality magnetic resonance venography (MRV) studies. MRV demonstrated occlusion of the major dural sinuses and cortical veins including superior sagittal sinus in 7 patients (35%), left transverse and sigmoid sinus 6 patients (30%), right transverse and sigmoid sinus 5 (25%), left transverse sinus in one patient (5%) and cortical veins in one patient (5%). 3.2. CT perfusion findings

2.3.1. Data interpretation The data was analyzed by two experienced attending neuroradiologists (NK, JRB) and one fellow (RKG). The CT perfusion scans were analyzed at the workstation (Advantage Windows; GE medical systems) with available software (4.0 GE CT perfusion) which calculated cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT) on color coded maps. The software required placement of small regions of interest (ROIs) on one artery and one vein to generate arterial input functions and venous outflow functions, respectively, for the deconvolution analysis. The color-coded CBF, CBV, MTT and time to peak (TTP) maps thus generated were visually examined for areas of hypo perfusion. The mean values of rCBF, rCBV, and rMTT were calculated by placing the 100 mm2 regions of interest (ROIs) at the core and periphery in region of the parenchymal abnormalities and compared with mirror image ROI placed in the contralateral hemisphere. Follow-up CT perfusion was done at 1 month with the same data acquisition parameters as the initial CT perfusion. At one month follow-up, the images were similarly analyzed and quantitative perfusion parameters were calculated. Neurological deficit was determined based on NIHSS score 0 (0–1 = no neurological deficit, 2 or more = neurological deficit present) and patients were divided into two groups. CT perfusion parameters in the two

CT perfusion studies demonstrated areas of hypoperfusion in all the patients. The area of perfusion abnormally was identified by asymmetric color distribution in the color-coded perfusion maps of CBF, CBV and MTT. 100 mm2 diameter ROI markers were used to assess the perfusion parameters of the abnormal areas and at the similar level the contra lateral normal brain. On all perfusion scans 80 ROIs were placed in the central core of the lesion and 38 ROIs were placed in the periphery of the lesion and the parameters were derived. The CT perfusion studies were repeated on follow-up. The sections obtained on follow-up were matched with the pre-therapy perfusion scans and the parameters were again derived. The perfusion scores from each ROI in areas of central core and the periphery of the lesion were evaluated. In pre-treatment studies, areas of perfusion deficit showed that rCBF were significantly reduced in core and periphery of the lesion (68.5 ± 39.0%, 76.8 ± 43.5% respectively). rCBV was also reduced in core and periphery of the lesion (82.8 ± 43.4%, 86.9 ± 43.8% respectively). There was marked prolongation of rMTT in core and periphery of the lesion (158.2 ± 72.6%, 142.4 ± 55.7%) respectively. Mean perfusion parameters in the core and periphery of the lesion in pre-treatment studies were compared and results are depicted in the Table 1.

R.K. Gupta et al. / European Journal of Radiology 83 (2014) 185–190 Table 1 Mean perfusion parameters in core and periphery of affected areas before treatment.

Core of lesion Periphery of lesion

rCBF

rCBV

rMTT

68.5 + 39.0% 76.8 + 43.6% p < 0.001

82.8 + 43.4 86.9 + 43.8 p = 0.014

158.2 + 72.6 142.4 + 55.7 p = 0.829

187

Scatter plot showing overlap of rCBV and rCBV2 300 250 200 150

Scatter plot showing overlap of rCBF and rCBF2 250

100 50

200

0

150

rCBV

100

rCBV2

Fig. 2. Scatter plot showing overlap of rCBV and rCBV2 in lesions in pretreatment studies (rCBV – relative cerebral blood volume in core of lesion, rCBV 2 – relative cerebral blood volume in periphery of lesion).

50 0

rCBF

rCBF2

Fig. 1. Scatter plot showing overlap of rCBF and rCBF2 in lesions in pretreatment studies (rCBF – relative cerebral blood flow in core of lesion, rCBF 2 – relative cerebral blood flow in periphery of lesion).

There was a good correlation for rCBF and rCBV values in the core and periphery of the lesion with a p value of 60.5%, for rCBV > 75.5% and for rMTT it was 40% decrease in the rCBF had severe brain damage. Koenig et al. [16] in their study in arterial stroke have found that a value of 0.48 for rCBF and 0.60 for rCBV as a best thresholds for the discrimination of the infarct and peri-infarct regions in arterial stroke. However, the authors also mentioned that because of overlap of data from infarcted and non-infarcted regions, these values do not reflect the real threshold for a given tissue compartment to survive an ischemic injury. Based on these results we propose that there is no discrete of region of infarct and penumbra in CVST and any region of perfusion deficit showing rCBF > 60.5% and rCBV 75.5% and rMTT < 148.5% would be associated with good clinical outcome. We also found that rCBF is the best parameter to assess prognosis (p < 0.001). Post treatment CT perfusion showed significant improvement in the perfusion parameters and these changes were statically significant in the core of the lesion, The p values for rCBF, rCBV, rMTT were 0.014, 0.021 and 0.005 respectively. The improvement of parameters in the periphery of the lesions however were not statically significant (p > 0.05). Given these observations we believe that lesions following venous thrombosis showed significant improvement following treatment. The best quantitative evaluation of improvement could be easily performed by CT perfusion studies. A repeat CT perfusion study was therefore justified. The calculated CT dose was well within the established norms for the CT technology available for the study. The ease of performing these studies, patient comfort, cost and the short time of acquisition contributed to the fidelity of data obtained and the successful completion of the study. The role of MR imaging (MRV studies) was secondary to the CT perfusion studies. The MRV was performed for anatomic localization of the venous sinuses involved. A TOF flight study was performed because it entailed short acquisition times, which was

190

R.K. Gupta et al. / European Journal of Radiology 83 (2014) 185–190

comfortable to most patients who were imaged in an emergent clinical situation. To date, to the best of our knowledge, there are no prospective studies which have assessed perfusion parameters before and after treatment in patients of CVST. Therefore we propose CT perfusion may prove to be a valuable investigation in assessing the lesions associated with CVST. 5. Conclusion CT perfusion, performed in patients of diagnosed CVST is a valuable diagnostic and prognostic tool. Parameters derived from perfusion CT could qualitatively assess the hemodynamic state of the affected portions of the brain. rCBFappears to be the best parameter to assess the perfusion deficit. Based on our results, any region of perfusion deficit showing rCBF > 60.5% and rCBV > 75.5% and rMTT < 148.5% would be associated with good prognosis. Following appropriate treatment, CT perfusion would be invaluable in assessing improvement of deranged perfusion parameters. Perfusion studies can be rapidly acquired in emergent settings, are reproducible and are readily available in most emergency departments. Therefore these studies can prove to be invaluable as an aid to assess patients of CVST. Conflict of interest I confirm that, in this research work, there was no possible conflict of interest References [1] Bousser MG, Chiras J, Bories J, Castaigne P. Cerebral venous thrombosis: a review of 38 cases. Stroke 1985;16(2):199–213.

[2] Einhäupl KM, Villringer A, Meister W, et al. Lancet 1991;338(September (8767)):597–600. Erratum in: Lancet 1991 Oct 12;338(8772):958. [3] Krayenbuhl HA. Cerebral venous and sinus thrombosis. Clin Neurosurg 1966;14:1–24. [4] Milandre L, Pellissier JF, Vincentelli F, Khalil R. Deep cerebral venous system thrombosis in adults. Eur Neurol 1990;30(2):93–7. [5] Nagpal RD. Dural sinus and cerebral venous thrombosis. Neurosurg Rev 1983;6(3):155–60. [6] Doege CA, Tavakolian R, Kerskens CM, et al. Perfusion and diffusion magnetic resonance imaging in human cerebral venous thrombosis. J Neurol 2001;248(7):564–71, http://dx.doi.org/10.1007/s004150170133. [7] Shinohara Y, Takagi S, Kobatake K, Gotoh F. Influence of cerebral venous obstruction on cerebral circulation in humans. Arch Neurol 1982;39(8):479–81. [8] Hayashi T, Miyazaki H, Toda Y, Ishiyama N. Single photon emission computed tomography of reversible magnetic resonance imaging abnormalities of deep cerebral venous thrombosis accompanied with malignant glioma. Surg Neurol 2007;67(2):195–9, http://dx.doi.org/10.1016/j.surneu.2006.05.061. [9] Schmiedek P, Einhaupl K, Moser E, Bull U, Leinsinger G, Kreisig T. Cerebral blood flow in patients with sinus vein thrombosis. In: Einhaupl K, Kempski O, Baethmann A, editors. Cerebral sinus Thrombosis. New York/London: Plenum press; 1990. p. 75–83. [10] Fries G, Wallenfang T, Hennen J, et al. Occlusion of the pig superior sagittal sinus, bridging and cortical veins: multistep evolution of sinus-vein thrombosis. J Neurosurg 1992;77(1):127–33, http://dx.doi.org/10.3171/jns.1992.77.1.0127. [11] Fujita K, Kojima N, Tamaki N, Matsumoto S. Changes in regional cerebral blood flow in intracranial venous hypertension. Neurol Med Chir (Tokyo) 1986;26(2):111–5. [12] Hossmann KA. Viability thresholds and the penumbra of focal ischemia. Ann Neurol 1994;36(4):557–65, http://dx.doi.org/10.1002/ana.410360404. [13] Kurokawa Y, Hashi K, Okuyama T, Uede T. Regional ischemia in cerebral venous hypertension due to embolic occlusion of the superior sagittal sinus in the rat. Surg Neurol 1990;34(6):390–5. [14] Nakase H, Heimann A, Kempski O. Alterations of regional cerebral blood flow and oxygen saturation in a rat sinus-vein thrombosis model. Stroke 1996;27(4):720–7, discussion 728. [15] Ungersbock K, Heimann A, Kempski O. Cerebral blood flow alterations in a rat model of cerebral sinus thrombosis. Stroke 1993;24(4):563–9, discussion 569–570. [16] Koenig M, Klotz E, Luka B, Venderink DJ, Spittler JF, Heuser L. Perfusion CT of the brain: diagnostic approach for early detection of ischemic stroke. Radiology 1998;209(1):85–93.