Oxford Textbook of Neuroimaging

Edited by Massimo Filippi Neuroimaging Research Unit, and Department of Neurology, Institute of Experimental Neurology, Division of Neuroscience, San Raffaele Scientific Institute, Vita-Salute San Raffaele University, Milan, Italy

Series Editor Christopher Kennard

1 1

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Oxford University Press 2015 The moral rights of the authors‌ have been asserted First Edition published in 2015 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2015933911 ISBN 978–0–19–966409–2 Printed in China by Asia Pacific Offset Ltd. Oxford University Press makes no representation, express or implied, that the drug dosages in this book are correct. Readers must therefore always check the product information and clinical procedures with the most up-to-date published product information and data sheets provided by the manufacturers and the most recent codes of conduct and safety regulations. The authors and the publishers do not accept responsibility or legal liability for any errors in the text or for the misuse or misapplication of material in this work. Except where otherwise stated, drug dosages and recommendations are for the non-pregnant adult who is not breast-feeding Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work. CHAPTER 1 Computed tomography Qinghua Hou, Elizabeth Tong, Cong Gao, and Max Wintermark

Introduction and limit the ionizing radiation harm. Unlike conventional X-ray imaging, CT images are not developed directly from the attenu- Since its first introduction in 1971, computed tomography (CT) has ation of projection X-rays. Instead, slices of specific areas of the evolved rapidly. With the evolution from fan beam to cone beam, body are imaged and divided into a matrix of voxels, the intensity and from axial to helical mode, the increase in the number of detec- of which is calculated from multiple projections from an X-ray tors (from a single detector array to 256/320 detector arrays), and source to a panel of detectors, both rotating in the CT gantry. the advent of dual energy technology, CT now offers very short The attenuation value of an individual voxel depends upon its scan time (even for extensive anatomical coverage), improved spa- contents and can be quantitatively scored in Hu (Hu; called after tial and temporal resolution, and reduced patient dose, both in the inventor of CT), which ranges from –1000 Hu to +1000 Hu, terms of radiation and contrast agent. CT is the imaging modal- with water in the middle (Hu = 0), and air (–1000 Hu) and bone ity of choice in the emergency setting. It is fast, its requirements (+1000 Hu) at the extremities of the scale. Modern CT scan- are low, and it is available 24 hours a day, 7 days a week. Modern ners can reconstruct axial, sagittal, coronal, oblique, 3D, or even CT scanners offer volumetric acquisitions, which improve the 4-dimensional (4D) images. capability to evaluate intracranial and spinal structures in reconstructed three-dimensional (3D) views. With the development of post-progressing software, CT has expanded its applications Single-slice versus multislice computed beyond the evaluation of structural anatomy to include functional tomography scanners analyses such as perfusion-CT (PCT) to assess cerebral blood sup- In very early single-slice CT scanners, a single row of detectors ply (Table 1.1). were arrayed in a line opposite to the X-ray tube inside the CT gan- Despite these technological advances, the basic imaging principle try. These scanners used a narrow beam of X-rays (pencil width), of CT has not been changed and relies on X-ray physics. The radia- which was collimated and developed very little scatter when it tion dose concerns associated with CT remains, although it has passed through the human body. With such scanners, one gantry been alleviated by a number of dose-reduction techniques imple- rotation could only sample one slice. mented by the imaging manufacturers. Nowadays, the radiation In multislice CT scanners, detectors are arrayed linearly in multi- dose of a non-contrast head computed tomography (NCT) ranges ple rows, so that the machine can cover multiple slices in one rota- from 1.7 to 2.5 mSv [1,2], lower than the background radiation for tion (up to 320 slices with the most advanced CT technology). Also, a person living in Boston for a year (approximately 3 mSv) [3]‌. The a fan beam X-ray is used. The benefits of this geometry are not only estimated lifetime risk of death from cancer that is attributable to limited to obtaining a larger coverage with one single rotation, but a single NCT is 0.01–0.02% [4]. With the advent of dual-energy can also shorten the scanning time. With multislice CT scanners, CT, the radiation dose associated with CT may be reduced further the spatial and temporal resolution, and the contrast resolution are when contrast-enhanced examination is applied [5]. also greatly improved. The latest generation of CT scanners (e.g. 320-dector row CT) Computed tomography technology uses a cone beam of X-rays with a flat panel detector for captur- ing the images. Cone-beam CT scanners can produce 3D images CT is an imaging technique based on X-rays, a form of elec- with a single rotation of the gantry [6]‌, and also time-resolved tromagnetic radiation with a wavelength range of 0.01–10 nm. CT-angiogram (4D CTA) showing the dynamic progression of the X-ray beams can traverse the human body and be used to create contrast through arteries and veins [7]. an image (they can penetrate, but do no harm), while at the same time the photon energy it carries makes it an ionizing radiation, which can cause harm to living tissue (in proportion to the Image reconstruction algorithms absorption). CT imaging uses a specific range of wavelengths of For a considerable time, filtered back-projection has dominated X-rays (wavelength below 0.2–0.1 nm) to obtain good penetration the field of CT image reconstruction. Filtered back-projection runs 4 Section 1 fundamentals of neuroimaging techniques

Table 1.1 Comparison of CT, MRI, and single photon emission tomography (SPECT)/positron emission tomography (PET) imaging

CT MR SPECT/PET Spatial resolution 0.5 mm 1 mm 4–6 mm Scan time seconds 20–40 minutes 10–15 minutes [22] Availability 24/7 24/7 Business hours Contrast Iodinated-based Gadolinium-based Radionuclide tracer Vascular imaging CTA and CT-venogram MRA, contrast-enhanced (CE)-MRA, MRV Not applicable

Functional imaging Perfusion and blood–brain Perfusion and diffusion, 2O consumption, and Blood supply and metabolic rates barrier permeability evaluation metabolism spectrum analysis, connectivity evaluation, blood–brain barrier imaging. analysis, etc. Artefacts Streak artefact (Fig. 1.1), Geometric distortion caused by susceptibility Photon scatter and attenuation beam-hardening. artefacts, chemical-shift artefacts, Gibbs’ artefacts Major limitations Ionizing radiation, low sensitive Longest scan time, metal implant and pace maker Ionizing radiation, limited availability in to posterior fossa lesions contra-indication the emergency setting

the projection images back through the image to obtain a rough Computed tomography imaging modalities approximation of the original object (back-projection), and uses a high-pass filter to eliminate the image blurring caused by such NCT ‘back-projection’. Routine NCT is performed in the axial plane, parallel to the orbit- More recently, iterative reconstruction algorithms have been omeatal line, with 2.5-mm thick slices. For cervical spine imaging, developed, originally to reduce the noise of images. These algo- 0.625-mm thick slices are obtained, while 1.5-mm thick slices are rithms consist of a series of sequential reconstructions and the standard for imaging the thoracic and lumbar spine. corrections—forward- and back-projection reconstruction steps, NCT is used for the initial evaluation of many intracranial and and comparison of the projection data with the real measured raw spinal lesions, especially in the acute setting. NCT is the first-line data and correction of deviation in projection between each step. examination for acute traumatic brain injury (TBI) and stroke, The correction is based on the statistical counting of the detected since it is especially suited for detecting ischaemia, haemorrhage, photons in the reconstruction process. The iterative process can be and skull fracture. performed in the raw data domain, in the image domain alone, or NCT is less sensitive than magnetic resonance imaging (MRI) in both [8,9]. for the assessment of posterior fossa and brainstem lesions, due to Iterative reconstruction can reduce the radiation dose associated hard-beam artefacts (Fig. 1.1), and for the diagnosis of conditions with CT studies by 20–40% without compromising image quality such as haemorrhagic or non-haemorrhagic diffuse axonal injury, [10–13]. One of the downsides of iterative reconstruction lies in the contusion, early infarction, encephalitis, and brain tumour. longer image reconstruction time. Contrast-enhanced computed tomography Enhancement after injection of iodinated contrast indicates intra- Single-source versus dual-source computed vascular enhancement due to an increased permeability of the tomography blood–brain barrier, which leads to a leakage of the contrast material into the interstitium. By analysing the enhancement patterns Dual-energy CT is an emerging CT modality. There are currently associated with contrast agent injection, contrast-enhanced com- two ways used for dual-energy scanning: puted tomography (CECT) can help differentiate lesions such as 1. Two X-ray tubes and two detectors (mounted on a CT gantry vascular malformations, neoplasms, and active inflammation with a mechanical offset of 90°) with the two X-ray tubes being (infectious and non-infectious) [16]. operated at different voltages. Computed tomography angiography 2. A single X-ray tube that can rapidly switch its peak voltage and Computed tomography angiography (CTA) with multiplanar refor- one detector. matted images, maximum intensity projection images (MIP), and Through the simultaneous acquisition of two image series with 3D reconstructions of axial source images provide images compa- different kVp (e.g. 40 and 140 kVp), dual-energy CT facilitates rable with, or even superior to, those obtained with digital subtrac- material separation. For example, it can be used to generate vir- tion angiography (DSA) [17,18]. Unlike time-of-flight magnetic tual non-contrast images of the brain from a contrast-enhanced resonance angiography (MRA), CTA is less susceptible to turbulent CT; this also allows the removal of bone and calcium from a CTA or slow-flow artefacts. Spatial resolution of CTA is approximately [14,15]. twice that of gadolinium-enhanced MRA [19]. CTA has become chapter 1 neuroimaging: computed tomography 5

CT hyperdensity Acute intracranial haemorrhages Acute haemorrhage always appears as a hyperdense area in any region of the brain parenchyma, ventricles, subarachnoid space, or extra-axial space. Over time, intracranial haemorrhages become isodense (in the subacute phase) [23] and hypodense (in the chronic phase) [24]. Of note, low attenuation within an acute haematoma may indicate hyperacute, ongoing bleeding [25]. Vascular clot An acute thrombus can be detected on NCT as an area of hyperdensity, featuring the ‘dense artery sign’ in the setting of a M1 occlusion of the middle cerebral artery (MCA) (Fig. 1.2), and the ‘dot sign’ in the setting of a M2 or M3 occlusion. Venous sinus thrombosis also appears hyperdense on NCT, while featuring a delta sign or trian- gular lack of contrast filling on post-contrast CT. Calcifications Calcifications are very hyperdense. Some calcifications in the brain are within normal limits, e.g. pineal gland calcifications, choroid plexus calcifications, falx calcifications, and basal ganglia calcifi- Streak artefact on non-contrast CT caused by intravascular stent (arrow). Fig. 1.1 cations/mineralization. Basal ganglia calcifications (and cerebel- lar dentate nuclei calcifications) are particularly pronounced in the modality of choice to non-invasively assess cervical and intrac- patients with Fahr’s disease [26]. Calcifications are also seen in a ranial vessels. Dynamic CTA provides dynamic, real-time infor- variety of pathological conditions, e.g. vascular malformations, mation, similar to that offered by DSA, which is rarely used for chronic haematomas, brain infections, brain tumours, and congen- diagnostic purposes. ital malformations, such as Sturge–Weber syndrome and tuberous sclerosis. PCT Others PCT evaluates capillary, tissue-level circulation, which is beyond the Highly cellular neoplasms, such as lymphomas, can appear as resolution of traditional anatomic imaging. PCT is most commonly hyperdense. carried out using dynamic sequential scanning of a pre-selected slab of the brain (modern CT scanners offer whole-brain cover- CT hypodensity age), during the injection of a bolus of iodinated contrast material Oedema as it travels through the vasculature. By recording the wash-in and Subacute infarcts, infections, and brain tumours alter the permea- wash-out of the contrast bolus through the cerebral vasculature, bility of the blood–brain barrier, resulting in vascular leakiness and several parameters describing the cerebral perfusion, i.e. cerebral water extravasation into the interstitial space, or vasogenic oedema. blood flow (CBF), cerebral blood volume (CBV), mean transit The accumulation of interstitial water decreases the attenuation of time (MTT), time to peak (TTP), and blood–brain barrier perme- brain tissue and features a hypodensity. Likewise, chronic ischae- ability (BBBP), can be quantitatively evaluated on reconstructed mia causes encephalomalacia and gliosis, which also present as colour maps. hypodensity on CT. PCT limitations and pitfalls include the following: Hydrocephalus 1. Complex post-processing and long-lasting experience in inter- Hydrocephalus relates to enlarged ventricles. In obstructive hydro- pretation of the images. cephalus, ventricles upstream of an obstructive lesion are selectively 2. Due to arterial input function (AIF) delay, vascular stenosis dilated (Fig. 1.3). In communicating hydrocephalus, all ventricles may result in an overestimation of the area of cerebral perfusion are dilated. abnormality [20]. It is therefore of importance that PCT should Mass effect be performed and interpreted with concurrent CTA. Mass effect refers to the deviation or distortion of normal structures 3. Although higher than MR perfusion-weighted imaging [21] and by an abnormal structure or mass, and the swelling accompanying SPECT/PET, the resolution of PCT maps is still relatively low. it. Mass effect is typically observed in the case of large intracra- Small infarcts may be missed, and white matter changes can be nial haematomas and brain tumours, as well as in the setting of misdiagnosed. malignant infarct. Early mass effect is seen as local effacement of sulci. Progressive mass effect results in midline shift, ventricular Common CT features entrapment, and herniation (Fig. 1.4). Ventricular entrapment typi- CT findings are categorized into hyperdense, isodense, and cally occurs when there is significant midline shift, the foramen of hypodense, depending on their density compared with that of nor- Monroe is compressed, and the drainage of the lateral ventricles mal brain tissue (approximately 40 Hu). Normal cerebrospinal fluid is impaired. There are different types of herniation—subfalcine, (CSF) has a density of around 5 Hu. Bone density is greater than uncal, downward transtentorial, upward transtentorial, and tonsil- 350 Hu, fat density is –100 Hu, and that of air is –1,000 Hu. lar and transcalvarial herniation types. Severe herniation can cause 6 Section 1 fundamentals of neuroimaging techniques

(a) (b)

Fig. 1.2 Hyperdense MCA sign caused by an acute intravascular clot. This finding on non-contrast CT (a, arrow) corresponds to a segmental filling defect on CT-angiography (b,c, arrows) and on digital subtraction angiography (d, arrow).

secondary ischaemic lesions, by compression of neighbouring to acute ischaemia can be improved by setting the window centre arteries (anterior cerebral artery by subfalcine herniation, posterior level around 30 Hu and narrowing the window width to 8–10 Hu cerebral artery by uncal and downward transtentorial herniation), [28]. The role of NCT in the setting of an acute stroke is mainly to and is lethal without prompt surgical decompression. rule out the presence of haemorrhage [29] and mimics of stroke, such as infection, inflammation, and neoplasm. Early signs of CT clinical applications ischaemic stroke on NCT include sulcal effacement, blurring of Ischaemic stroke cortical gray–white matter differentiation, and in 50% of patients NCT is the imaging modality of choice to initially assess patients [30] a dense artery sign, as described previously. The Alberta Stroke suspected of acute stroke, despite its limited sensitivity to ischae- Program Early CT Score (ASPECTS) quantifies the extent of the mia in the first few hours following symptom onset [27]. Sensitivity ischaemic changes visible on CT and helps to determine whether

(a) (b)

Fig. 1.3 Obstructive hydrocephalus caused by colloid cyst in the foramen of Monroe (arrow). The obstruction led to sequential enlargement of bilateral ventricles (a, arrowhead) and interstitial oedema in periventricular tissues (b, star). chapter 1 neuroimaging: computed tomography 7

(a) (b)

Fig. 1.4 Midline shift and uncal herniation caused by a right insular/temporal abscess. Non-contrast CT clearly demonstrates the mass effect caused by this abscess (a, star) with shift of the midline to the left (a, arrow), entrapment of the temporal horn of the right lateral ventricle, and right uncal herniation (b, arrowhead) compressing the midbrain. an acute stroke involves more than one-third of the MCA territory. The safety of the combined use of CTA and PCT in stroke patients The latter represents a contraindication to the administration of IV has been well documented [40]. Iodinated contrast administration tissue plasminogen activator (tPA), together with the presence of for CTA/PCT does not cause significant renal injury nor does it haemorrhage on the admission NCT [29]. In the subacute phase of interfere with the safety and efficacy of the tPA treatment [41]. an ischaemic stroke, a typical ‘wedge-shaped’ hypodensity can be seen on NCT (Fig. 1.5). Intracranial haemorrhage CTA is routinely used to detect the exact site of arterial occlusion The AHA/ASA guidelines indicate that either NCT or MRI may be in stroke patients. Compared with DSA, CTA has a 98.4% sensitiv- used in the initial diagnostic evaluation of an intracranial haemor- ity and 98.1% specificity for detecting proximal intracranial arterial rhage (ICH) [42]. occlusion [23], and an 85% sensitivity and 93% specificity in detect- The majority of spontaneous ICHs are related to arterial hyper- ing significant extracranial arterial stenosis [30]. In addition to the tension. Hypertensive ICHs (Fig. 1.6) tend to involve the basal luminal evaluation, CTA is capable of evaluating the arterial wall, ganglia, thalamus, cerebellum, and brainstem, resulting from the detecting vasculopathies (for instance, dissection), and characteriz- rupture of micro-aneurysms of perforating vessels in these regions. ing carotid atherosclerotic plaques [31]. Intracranially, CTA source When multiple haematomas are seen, if the haematomas have an images have increased sensitivity for acute ischaemic changes and irregular shape, or if they occur in unusual locations, secondary can assess the collateral circulation distal to an arterial occlusion. ICH should be considered. Finally, when the stroke CTA scan extends below the origin of the Subarachnoid haemorrhage (SAH) occurs in the setting of aortic arch to include the left atrium or the whole heart, CTA can trauma and of aneurysmal rupture (Fig. 1.7). Sensitivity of NCT to detect the presence of a left atrial or ventricular thrombus. SAH in the first 24 hours is close to 100%. After 48 hours NCT sen- PCT is increasingly recognized as a diagnostic tool in patients sitivity decreases to less than 85% [43]. In the presence of a negative with suspected stroke. Its applications in the setting of stroke NCT after 48 hours, a lumbar puncture is then required to detect include, but are not limited to: xanthochromia. CTA is used in patients with acute ICH for the purpose of iden- 1. Improving the sensitivity and accuracy of stroke diagnosis (in tifying secondary causes of ICH [44]. CTA has a sensitivity and some cases, a lesion on PCT leads to more careful scrutiny and specificity comparable with DSA in terms of detecting intracranial identification of a vascular occlusion that was not evident (Fig. aneurysms [45,46]. CTA presents a number of advantages in terms 1.5), particularly in the M2 and more distal MCA branches) of surgical planning—CTA data can easily be uploaded to neuro- [32–35]. navigation systems [47], and show precisely the relationship of the 2. Excluding stroke mimics [36]. aneurysm to the skull base [48,49], calvarium, and adjacent veins 3. Better assessment of the ischaemic core [34] and collateral [50]. CTA is also useful in planning an endovascular intervention. flow [18]. CTA and CECT can detect the ‘spot sign’, corresponding to the acute extravasation of contrast in the haematoma (Fig. 1.6). A ‘spot 4. Prediction of haemorrhagic transformation and malignant sign’ is indicative of haematoma expansion [51] and predictive of a oedema [37, 38]. poor outcome. Currently, different PCT software use different methods to process PCT has been used to detect a ‘penumbra’ surrounding haema- the data, and extract the infarct core and penumbra information. tomas. The existence of such a penumbra surrounding haemato- Standardization of PCT processing and interpretation is lacking. mas is still very much debated [52,53]. In a series of patients who The use of a relative MTT threshold of 145% of the contralateral underwent surgical haematoma evacuation, PCT after surgery values to delineate the total ischaemic area is suggested, and an demonstrated a correction of the perfusion deficit surrounding the absolute CBV threshold of 2 mL/100 g to differentiate the infarct haematoma, suggesting that the latter represents reduced oxygen core and the penumbra within the total ischaemic area [39]. demand of tissue damaged by pressure and clot components [52]. (a) (b)

(e) (f)

Fig. 1.5 Typical CT imaging workup in a patient suspected of acute ischaemic stroke. (a) NCT shows mild effacement of the right lentiform nucleus. (b) PCT shows the right MCA to be ischaemic, with a mixture of ischaemic core (red) and penumbra (green). (c) CT-angiography demonstrates the right MCA occlusion (arrow) responsible for the ischaemia, later confirmed on digital subtraction angiography (d, arrow). (e,f) Follow-up NCT and PCT demonstrate that, in the absence of early recanalization, the whole ischaemic territory infarcted.

(a) (b) (c)

Fig. 1.6 ‘Spot sign’ on contrast-enhanced CT. (a) NCT shows an intraparenchymal haematoma in the left thalamus with extension in the third ventricle. (b) Contrast-enhanced CT shows a ‘spot sign’ (arrow) within the parenchymal haematoma, indicating active extravasation of contrast. (c) The size of the parenchymal haematoma and the intraventricular extension have progressed significantly, as typically seen when the ‘spot sign’ is present. chapter 1 neuroimaging: computed tomography 9

(a) (b) (c)

Fig. 1.7 Aneurysmal subarachnoid haemorrhage. (a) NCT demonstrates subarachnoid haemorrhage in the left Sylvian fissure (arrow), the left ambient cistern, and the interhemispheric fissure. A relative hypodensity suggests the presence of an aneurysm in the left posterior communicating artery region (arrowhead), later confirmed on CT-angiography maximal intensity projections (b,c; arrows for suprclinoid internal carotid artery and arrowheads for posterior communicating artery).

PCT is also used to detect vasospasm, one of the complications scrutinized for traumatic injury. Thin-slice CT with multipla- of subarachnoid haemorrhage that typically occurs 4–7 days fol- nar reformats increases the sensitivity of the technique to subtle lowing SAH, and it adds prognostic information regarding delayed skull fractures, e.g. of the temporal bone. cerebral ischaemia and poor outcome [54,55]. Epidural haematomas (EDHs) are associated with skull fractures, Another complication of SAH is communicating hydrocephalus, and generally result from a traumatic injury to branches of the mid- which can be diagnosed early on NCT. dle meningeal artery. An EDH appears on NCT as a lenticular or biconvex blood collection. EDHs can cross dural insertions (falx, Traumatic brain injury tentorium), but do not usually cross suture lines. CT is the imaging modality used to screen trauma patients in the Subdural haematomas (SDHs) occur in elderly patients with acute phase [56]. Head NCT is indicated in trauma patients with brain atrophy. They usually occur in the absence of skull frac- loss of consciousness or post-traumatic amnesia if one or more of tures. They have a crescentic shape on NCT. SDHs can cross the following is present—age greater than 60 years, headache, vom- sutures. They do not cross, but extend along dural attachments iting, drug or alcohol intoxication, deficits in short-term memory, (falx, tentorium). Acute and subacute SDHs may be subtle and physical evidence of trauma above the clavicle, Glasgow Coma difficult to detect. An appropriate, narrow CT window when Scale score less than 15, post-traumatic seizure, focal neurological reviewing NCT can help, together with a subtle loss of normal deficit, or coagulopathy [57]. brain sulci or the buckling of gray–white matter junction caused Skull fractures may be linear, comminuted, or depressed (Fig. 1.8). by the SDH. They are demonstrated as discontinuities with or without dis- Traumatic SAH is usually more localized than aneurysmal SAH, placement on the CT bone windows, and may be associated with and generally has its epicentre at the site of the coup. As a result of extra-intracranial air and/or haemorrhage. Whenever a bone contrecoup injury, traumatic SAH may also occur contralateral to fracture is identified, underlying brain parenchyma should be the side of direct impact [25].

(a) (b)

Fig. 1.8 Mildly depressed left frontal and temporal bone fractures. Bone windows of thin non-contrast CT slices in coronal and axial planes demonstrate the discontinuity and mild depression of left frontal and temporal bones (a and b, arrows). 10 Section 1 fundamentals of neuroimaging techniques

Table 1.2 Rotterdam Classification score (a) (b)

Predictor value Score Basal cisterns Normal 0 Compressed 1 Absent 2 Midline shift No shift or shift <5 mm 0 Shift>5 mm 1 Epidural mass lesion Present 0 (c) (d) Absent 1 Intraventricular blood or traumatic SAH Absent 0 Present 1

Note: to make the grading numerically consistent with the grading of the motor score of GCS and with Marshall CT classification, the final score is the sum of the scoring items +1. Modified from Maas, A.I., Hukkelhoven CW, Marshall LF, et al. (2005). Prediction of outcome in traumatic brain injury with computed tomographic characteristics: a comparison between the computed tomographic classification and combinations of computed tomographic predictors. Neurosurgery, 57(6), 1173–82. Copyright (2005), with permission from Lippincott Williams & Wilkins.

Parenchymal contusions on NCT appear as focal low-density (e) (f) areas, except if they are haemorrhagic, in which case they appear as hyperdense. Contusions can be seen deep to a skull fracture or on the opposite side of the brain, as part of a contrecoup injury. The above CT features of TBI constitute the core of the Rotterdam CT classification (Table 1.2). Penetrating injuries can damage cervical and intracranial arteries and veins. Skull fractures, especially skull base fractures, can also result in vascular injuries. The latter can be evaluated by CTA (or CT-venogram). CTA can detect traumatic dissection and pseudoa- neurysms with high sensitivity. PCT abnormalities have been reported after TBI. Focal perfusion abnormalities are observed in and around contusions [59], and can be detected before structural abnormalities become apparent. PCT abnormalities might be useful to triage patients with extra-axial blood collections who need aggressive surgical treatment versus those who can be treated conservatively [60]. In Fig. 1.9 Spine fractures. (a) Lateral plain film and (b) anteroposterior myelogram TBI patients, baseline PCT hyperaemia is associated to a favour- demonstrate two compression fractures (L1 and L4, arrows) in a severely able outcome, while oligaemia is the hallmark of an unfavourable osteoporotic spine. (c,d,e) CT and (f) MRI allow better characterization of the outcome [61]. PCT can provide information about cerebral vascu- posterior wall and any compromise of the spinal canal by fracture fragments (arrowheads). lar auto-regulation [60], and therefore can be used as a monitoring tool in TBI patients. Modern CT with multiplanar reformats is the gold standard for the evaluation of the cervical spine for fractures, and has a sensitiv- Others ity that is two or three times that of plain films [62]. Also, CT is MRI, and not CT, is the standard-of-care for the evaluation of more accurate than plain films in diagnosing thoracolumbar spine patients with suspected epilepsy, infections, or brain tumours. CT traumatic injuries (Fig. 1.9) [63]. CT can provide indirect signs to in these patients is only obtained to rule out haemorrhage, hydro- suggest ligamentous injury, but when these are suspected MRI is cephalus, or mass effect, especially in the emergency setting. CT is the preferred method of evaluation. usually obtained with contrast in these cases. chapter 1 neuroimaging: computed tomography 11

18. Tan JC, Dillon WP, Liu S, et al. (2007). Systematic comparison of Conclusions perfusion-CT and CT-angiography in acute stroke patients. Annals of CT is the fastest neuroimaging modality and has been established Neurology, 61(6), 533–43. as the first line of examination in the setting of stroke and TBI. 19. Moore C, Heck D, Beauchamp N, et al. (2003). Detection on CT angi- With its exquisite spatial resolution and functional capabilities, ography of an accessory middle cerebral artery simulating a fusiform coupled with new developments, such as whole-brain coverage and aneurysm on MR angiography. AJR American Journal of Roentgenology, 180(2), 544–5. dual-energy scanning, CT will remain the test of first choice in the 20. Waaijer, A, van der Schaaf IC, Velthuis BK, et al. (2007). Reproducibility evaluation of neurological disorders in the emergency setting. of quantitative CT brain perfusion measurements in patients with symptomatic unilateral carotid artery stenosis. AJNR American Journal References of Neuroradiology, 28(5), 927–32. 1. Wintermark M, Maeder P, Verdun FR, et al. (2000). Using 80 kVp versus 21. Coutts SB, Simon JE, Tomanek AI, et al. (2003). Reliability of assessing 120 kVp in perfusion CT measurement of regional cerebral blood flow. percentage of diffusion-perfusion mismatch.Stroke , 34(7), 1681–3. AJNR American Journal of Neuroradiology, 21(10), 1881–4. 22. Wintermark M, Sesay M, Barbier E, et al. (2005). Comparative overview 2. Cohnen M, Wittsack HJ, Assadi S, et al. (2006). Radiation exposure of of brain perfusion imaging techniques. Stroke, 36(9), e83–99. patients in comprehensive computed tomography of the head in acute 23. Lev MH, Farkas J, Rodriguez VR, et al. (2001). CT angiography in the stroke. AJNR American Journal of Neuroradiology, 27(8), 1741–5. rapid triage of patients with hyperacute stroke to intraarterial throm- 3. Konstas AA, Wintermark M, and Lev MH. (2011). CT perfusion bolysis: accuracy in the detection of large vessel thrombus. Journal of imaging in acute stroke. Neuroimaging Clinics of North America, 21(2), Computer Assisted Tomography, 25(4), 520–8. 215–38, ix. 24. Jakubik LD, Cockerham J, Altmann AR, et al. (2003). The ABCs of pedi- 4. Berris, T, Gupta, R, and Rehani, MM. (2013). Radiation dose from atric laboratory interpretation: understanding the CBC with differential cone-beam CT in neuroradiology applications. AJR American Journal of and LFTs. Pediatric Nursing, 29(2), 97–103. Roentgenology, 200(4), 755–61. 25. Le TH, and Gean AD. (2009). Neuroimaging of traumatic brain injury. 5. Postma AA, Hofman PA, Stadler AA, et al. (2012). Dual-energy Mount Sinai Journal of Medicine, 76(2), 145–62. CT of the brain and intracranial vessels. AJR American Journal of 26. Ogi S, Fukumitsu N, Tsuchida D, et al. (2002). Imaging of bilateral strio- Roentgenology, 199(Suppl. 5), S26–33. pallidodentate calcinosis. Clinical Nuclear Medicine, 27(10), 721–4. 6. Anas EM, Kim JG, Lee SY, et al. (2011). High-quality 3D correction of 27. Grotta JC, Chiu D, Lu M, et al. (1999). Agreement and variability in ring and radiant artifacts in flat panel detector-based cone beam volume the interpretation of early CT changes in stroke patients qualifying for CT imaging. Physics in Medicine and Biology, 56(19), 6495–519. intravenous rtPA therapy. Stroke, 30(8), 1528–33. 7. Orrison WW, Jr, Snyder KV, Hopkins LN, et al. (2011). Whole-brain 28. Lev MH, Farkas J, Gemmete JJ, et al. (1999). Acute stroke: improved dynamic CT angiography and perfusion imaging. Clinical Radiology, nonenhanced CT detection—benefits of soft-copy interpretation by 66(6), 566–74. using variable window width and center level settings. Radiology, 8. Winklehner A, Karlo C, Puippe G, et al. (2011). Raw data-based iterative 213(1), 150–5. reconstruction in body CTA: evaluation of radiation dose saving poten- 29. Jauch EC, Saver JL, Adams HP Jr, et al. (2013). Guidelines for the tial. European Radiology, 21(12), 2521–6. early management of patients with acute ischemic stroke: a guideline 9. Maass C, Meyer E, and Kachelriess M. (2011). Exact dual energy mate- for healthcare professionals from the American Heart Association/ rial decomposition from inconsistent rays (MDIR). Medical Physics, American Stroke Association. Stroke, 44(3), 870–947. 38(2), 691–700. 30. Liebeskind DS, Sanossian N, Yong WH, et al. (2011). CT and MRI 10. Ren Q, Dewan SK, Li M, et al. (2012). Comparison of adaptive statistical early vessel signs reflect clot composition in acute stroke.Stroke , 42(5), iterative and filtered back projection reconstruction techniques in brain 1237–43. CT. European Journal of Radiology, 81(10), 2597–601. 31. Kim SJ, Ryoo S, Hwang J, et al. (2012). Characterization of the infarct 11. Korn A, Fenchel M, Bender B, et al. (2012). Iterative reconstruction in pattern caused by vulnerable aortic arch atheroma: DWI and multide- head CT: image quality of routine and low-dose protocols in compari- tector row CT study. Cerebrovascular Disease, 33(6), 549–57. son with standard filtered back-projection.AJNR American Journal of 32. Schramm P, et al. (2004). Comparison of perfusion computed Neuroradiology, 33(2), 218–24. tomography and computed tomography angiography source images 12. Becce F, Ben Salah Y, Verdun FR, et al. (2013). Computed tomography of with perfusion-weighted imaging and diffusion-weighted imaging in the cervical spine: comparison of image quality between a standard-dose patients with acute stroke of less than 6 hours’ duration. Stroke, 35(7), and a low-dose protocol using filtered back-projection and iterative 1652–8. reconstruction. Skeletal Radiology, 42(7), 937–45. 33. Kloska SP, Schellinger PD, Klotz E, et al. (2004). Acute stroke assessment 13. Katsura M, Sato J, Akahane M, et al. (2013). Comparison of pure and with CT: do we need multimodal evaluation? Radiology, 233(1), 79–86. hybrid iterative reconstruction techniques with conventional filtered 34. Wintermark M, Fischbein NJ, Smith WS, et al. (2005). Accuracy of back projection: image quality assessment in the cervicothoracic region. dynamic perfusion CT with deconvolution in detecting acute hemi- European Journal of Radiology, 82(2), 356–60. spheric stroke. AJNR American Journal of Neuroradiology, 26(1), 104–12. 14. Gupta R, Phan CM, Leidecker C, et al. (2010). Evaluation of dual-energy 35. Scharf J, Brockmann MA, Daffertshofer M, et al. (2006). Improvement CT for differentiating intracerebral hemorrhage from iodinated contrast of sensitivity and interrater reliability to detect acute stroke by dynamic material staining. Radiology, 257(1), 205–11. perfusion computed tomography and computed tomography angiogra- 15. Thomas C, Andreas Korn, Dominik Ketelsen, et al. (2010). Automatic phy. Journal of Computer Assisted Tomography, 30(1), 105–10. lumen segmentation in calcified plaques: dual-energy CT versus stand- 36. Keedy A, Soares B, and Wintermark M. (2012). A pictorial essay of ard reconstructions in comparison with digital subtraction angiography. brain perfusion-CT: not every abnormality is a stroke! Journal of AJR American Journal of Roentgenology, 194(6), 1590–5. Neuroimaging, 22(4), e20–33. 16. Garg RK, Desai P, Kar M, et al. (2008). Multiple ring enhancing brain 37. Ryoo JW, Na DG, Kim SS, et al. (2004). Malignant middle cerebral artery lesions on computed tomography: an Indian perspective. Journal of infarction in hyperacute ischemic stroke: evaluation with multiphasic Neurological Sciences, 266(1–2): 92–6. perfusion computed tomography maps. Journal of Computer Assisted 17. Ledezma CJ, and Wintermark M. (2009). Multimodal CT in stroke Tomography, 28(1), 55–62. imaging: new concepts. Radiology Clinics of North America, 47(1), 38. Hom J, Dankbaar JW, Soares BP, et al. (2011). Blood–brain barrier per- 109–16. meability assessed by perfusion CT predicts symptomatic hemorrhagic 12 Section 1 fundamentals of neuroimaging techniques

transformation and malignant edema in acute ischemic stroke. AJNR 51. Huynh TJ, Demchuk AM, Dowlatshahi D, et al. (2013). Spot sign American Journal of Neuroradiology, 32(1), 41–8. number is the most important spot sign characteristic for predicting 39. Wintermark M, Flanders AE, Velthuis B, et al. (2006). Perfusion-CT hematoma expansion using first-pass computed tomography angiogra- assessment of infarct core and penumbra: receiver operating characteris- phy: analysis from the PREDICT study. Stroke, 44(4), 972–7. tic curve analysis in 130 patients suspected of acute hemispheric stroke. 52. Etminan N, et al. (2012). Perfusion CT in patients with spontaneous Stroke, 37(4), 979–85. lobar intracerebral hemorrhage: effect of surgery on perihemorrhagic 40. Lima FO, Lev MH, Levy RA, et al. (2010). Functional contrast-enhanced perfusion. Stroke, 43(3), 759–63. CT for evaluation of acute ischemic stroke does not increase the 53. Herweh, C, Beseoglu K, Turowski B, et al. (2007). Evidence against a risk of contrast-induced nephropathy. AJNR American Journal of perihemorrhagic penumbra provided by perfusion computed tomogra- Neuroradiology, 31(5), 817–21. phy. Stroke, 38(11), 2941–7. 41. Aulicky P, Mikulík R, Goldemund D, et al. (2010). Safety of performing 54. Killeen RP, Mushlin AI, Johnson CE, et al. (2011). Comparison of CT CT angiography in stroke patients treated with intravenous thromboly- perfusion and digital subtraction angiography in the evaluation of sis. Journal of Neurology, Neurosurgery & Psychiatry, 81(7), 783–7. delayed cerebral ischemia. Academic Radiology, 18(9), 1094–100. 42. Morgenstern LB, Hemphill JC 3rd, Anderson C, et al. (2010). Guidelines 55. Kunze E, Pham M, Raslan F, et al. (2012). Value of Perfusion CT, for the management of spontaneous intracerebral hemorrhage: a guide- Transcranial Doppler Sonography, and Neurological Examination line for healthcare professionals from the American Heart Association/ to Detect Delayed Vasospasm after Aneurysmal Subarachnoid American Stroke Association. Stroke, 41(9), 2108–29. Hemorrhage. Radiology Research and Practice, 2012, Article ID 43. Edlow JA, Malek AM, and Ogilvy CS. (2008). Aneurysmal subarachnoid 231206. hemorrhage: update for emergency physicians. Journal of Emergency 56. Provenzale JM. (2010). Imaging of traumatic brain injury: a review of Medicine, 34(3), 237–51. the recent medical literature. AJR American Journal of Roentgenology, 44. Delgado Almandoz JE, and Romero JM. (2011). Advanced CT imaging 194(1), 16–19. in the evaluation of hemorrhagic stroke. Neuroimaging Clinics of North 57. Jagoda AS, Bazarian JJ, Bruns JJ Jr, et al. (2009). Clinical policy: neuro- America, 21(2), 197–213, ix. imaging and decisionmaking in adult mild traumatic brain injury in the 45. Delgado Almandoz JE, Schaefer PW, Forero NP, et al. (2009). Diagnostic acute setting. Journal of Emergency Nursing, 35(2), e5–40. accuracy and yield of multidetector CT angiography in the evaluation 58. Maas AI, Hukkelhoven CW, Marshall LF, et al. (2005). Prediction of out- of spontaneous intraparenchymal cerebral hemorrhage. AJNR American come in traumatic brain injury with computed tomographic characteris- Journal of Neuroradiology, 30(6), 1213–21. tics: a comparison between the computed tomographic classification and 46. Yeung R, Ahmad T, Aviv RI, et al. (2009). Comparison of CTA to DSA combinations of computed tomographic predictors. Neurosurgery, 57(6), in determining the etiology of spontaneous ICH. Canadian Journal of 1173–82; discussion 1173–82. Neurological Sciences, 36(2), 176–80. 59. Wintermark M, van Melle G, Schnyder P, et al. (2004). Admission 47. Schmid-Elsaesser R, Muacevic A, Holtmannspötter M, et al. (2003). perfusion CT: prognostic value in patients with severe head trauma. Neuronavigation based on CT angiography for surgery of intracranial Radiology, 232(1), 211–20. aneurysms: primary experience with unruptured aneurysms. Minimally 60. Wintermark M, Chioléro R, van Melle G, et al. (2004). Relationship Invasive Neurosurgery, 46(5), 269–77. between brain perfusion computed tomography variables and cer- 48. Marro B, Galanaud D, Valery CA, et al. (1997). Intracranial aneu- ebral perfusion pressure in severe head trauma patients. Critical Care rysm: inner view and neck identification with CT angiography virtual Medicine, 32(7), 1579–87. endoscopy. Journal of Computer Assisted Tomography, 21(4), 587–9. 61. Metting Z, Rödiger LA, Stewart RE, et al. (2009). Perfusion computed 49. Villablanca JP, Hooshi P, Martin N, et al. (2002). Three-dimensional tomography in the acute phase of mild head injury: regional dysfunction helical computerized tomography angiography in the diagnosis, charac- and prognostic value. Annals of Neurology, 66(6), 809–16. terization, and management of middle cerebral artery aneurysms: com- 62. Diaz JJ, Jr, , Gillman C, Morris JA Jr, et al. (2003). Are five-view plain parison with conventional angiography and intraoperative findings. films of the cervical spine unreliable? A prospective evaluation in blunt Journal of Neurosurgery, 97(6), 1322–32. trauma patients with altered mental status. Journal of Trauma, 55(4), 50. Kaminogo M, Hayashi H, Ishimaru H, et al. (2002). Depicting 658–63; discussion 663–4. cerebral veins by three-dimensional CT angiography before surgical 63. Hauser CJ, Visvikis G, Hinrichs C, et al. (2003). Prospective validation of clipping of aneurysms. AJNR American Journal of Neuroradiology, computed tomographic screening of the thoracolumbar spine in trauma. 23(1), 85–91. Journal of Trauma, 55(2), 228–34; discussion 234–5.