CT uses photons and a collecting detector to record tissue density. On standard CT, air (the least dense) is black and bone (the most dense) is white. All other tissues are a shade of gray (see Table 18.2). Conventional CT uses several rings of detectors to acquire several images simultaneously. A computer translates the detections into a 2-D picture. This is either displayed on a computer monitor or printed onto x-ray film. Radiation exposure is clinically insignificant (approximately 5 rads). For extra protection, scans are angled to avoid the lens of the eye. If urgently needed in pregnancy, lead aprons are worn. The recent introduction of spiral CT further reduces the radiation exposure, and provides true 3-D imaging. An advantage of 3-D imaging is that no areas are missed between slices.

Modern CT scanners can generate brain images that range from 0.5 to 10 mm in thickness, with 3 to 5 mm used most commonly. The slice thickness of a CT image is an important variable in clinical scanning. Thinner slices allow visualization of smaller lesions. However, the thinnest sections have less contrast (i.e., the signal intensity difference between gray and white matter is less) because the signal/noise is lower. It also takes longer to complete the examination because more slices must be acquired. Therefore, there is more chance of patient movement degrading the images. The longer scan time also decreases the number of patients that can be examined in a day. Thicker sections (or slices) have greater contrast but smaller lesions may be missed. There is also greater artifact due to increased volume averaging. This is particularly true in the base of the skull and may obscure brain stem and mesial temporal structures.

MRI is based on manipulating the small magnetic field around the nucleus of the hydrogen atom (proton), a major component of water in soft tissue. To take an MRI of a patient’s soft tissues, the patient must be placed inside a large magnet. The strength of the magnet is measured in tesla (T). Most clinical systems have a field strength of 1.5 T, although 3.0 T systems are becoming more available for clinical work. (More powerful systems are often used in research settings.) A
midfield system is generally 0.5 T, and low-field units range from 0.1 to 0.5 T. The greater signal available with 1.5 and 3.0 T systems allows higher resolution images to be collected. However, this increased detail is costly because high-field systems are more expensive than mid- or lowfield systems. Also, many patients feel uncomfortable while lying inside these huge enclosing magnets (see Fig. 18.1A). Open design magnets are now available that provide less feeling of confinement to the patient (Fig. 18.1B).

To create an MRI, the patient’s hydrogen atoms are exposed to a carefully calculated series of radio frequency (RF) pulses while the patient is within the scanner’s magnetic field. These RF pulses change the magnetization of the hydrogen atoms, generating tiny electric signals that are picked up by a receiver placed close to the area being scanned (the imaging coil). A head coil that encloses the head, allowing maximum signal pickup, is used to obtain brain images. The magnetic field gradients needed to acquire the image are created by huge coils of wire embedded in the magnet, driven with large current audio amplifiers similar to those used for musical concerts. These can create a great deal of noise during the scan, and may distress the unprepared patient.

In MRI there are a wide variety of image types possible, each sensitive to a different physical aspect of tissue. This is quite different from CT, where the images reflect a single parameter, the density of tissue. The pulse sequence used to acquire the MRI (the combination of RF and magnetic field pulses used by the computer to create the image) determines the unique information the image will contain. Clinical MRI most commonly uses the spin echo (SE) or the fast spin echo (FSE) sequences. The expected appearances of tissues using the most common SE imaging methods are summarized in Table 18.3. T1-weighted images are traditionally considered best for displaying anatomy, whereas T2-weighted images are best for displaying pathology. However, both pathology and cerebrospinal fluid (CSF) will appear bright, making it difficult to visualize pathology near the ventricles. A variation on the T2-weighted scan has been developed (called fluid attenuated inversion recovery [FLAIR]) where CSF is dark, making pathology near CSF-filled spaces much easier to see. FLAIR MRI is extremely useful in neuropsychiatry (see Fig. 18.2).

Two other methods that are useful in clinical imaging are gradient echo (GE) or gradient refocused echo (GRE) and diffusion-weighted (DW) MRI. GE imaging (also called susceptibility weighted imaging) is very sensitive to anything in the tissue causing magnetic field inhomogeneity, such as hemorrhage or calcium. These images have artifacts at the interfaces between tissues with very different magnetic susceptibility, such as bone and brain. The artifacts at the skull base are sometimes severe. DW MRI is sensitive to the speed of water diffusion, and may be able to visualize areas of ischemic stroke in the critical first few hours after onset. It is also showing potential in the imaging of other conditions, including neurodegenerative conditions and traumatic brain injury (TBI).

CT is more widely available and is less expensive than MRI (see Table 18.4). Several types of pathology, particularly calcification, acute hemorrhage, and bone injuries, are better imaged with CT. Otherwise, MRI is preferred because it has a much higher resolution, can provide images in any plane of section, does not have artifacts near bone, and is sensitive to more types of pathology. CT has limited contraindications surrounding contrast agent administration: Allergy to iodine,

contrast dyes, or shellfish, creatinine ≥1.5 mg per dL, or metformin administration on the day of contrast scan. Contraindications to MRI include magnetic metals in the body, history of welding without screening x-rays, or implanted electrical, mechanical, or magnetic devices.

Structural brain images can be acquired either with or without intravenous administration of a contrast agent. These intravascular agents normally do not

penetrate into the brain, as they cannot pass through an intact blood–brain barrier (BBB). If this barrier is disrupted, the contrast agent leaks into surrounding tissue and changes its appearance on imaging. Conditions where the BBB may open include tumors, inflammatory or autoimmune conditions such as lupus or multiple sclerosis, and infectious processes. Other conditions where a contrast agent is useful include evaluation of suspected aneurysms, arteriovenous malformations, or other vascular processes (e.g., temporal arteritis).⁵,⁶,⁷

The neuroradiologist needs very clear clinical information on the imaging request form (not just “rule out pathology” or “new-onset mental status changes”). If a lesion is suspected in a particular location, the neuroradiologist should be informed of this or give enough clinical data for selection of the best imaging method and parameters to view suspicious areas. The neuroradiologist and technical staff also need information on the patient’s current condition (e.g., delirious, psychotic, easily agitated, or paranoid). This may eliminate difficulties with patient management during the scan.