Magnetic resonance (MR) is a dynamic and flexible technology that allows one to tailor the imaging study to the anatomic part of interest and to the disease process being studied. With its dependence on the more biologically variable parameters of proton density, longitudinal relaxation time (T1), and transverse relaxation time (T2), variable image contrast can be achieved by using different pulse sequences and by changing the imaging parameters. Signal intensities on T1, T2, and proton density-weighted images relate to specific tissue characteristics. For example, the changing chemistry and physical structure of hematomas over time directly affects the signal intensity on MR images, providing information about the age of the hemorrhage. Moreover, with MR's multiplanar capability, the imaging plane can be optimized for the anatomic area being studied, and the relationship of lesions to eloquent areas of the brain can be defined more accurately. Flow-sensitive pulse sequences and MR angiography yield data about blood flow, as well as displaying the vascular anatomy. Even brain function can be investigated by having a subject perform specific mental tasks and noting changes in regional cerebral blood flow and oxygenation. Finally, MR spectroscopy has enormous potential for providing information about the biochemistry and metabolism of tissues. As an imaging technology, MR has advanced considerably the past 10 years, but it continues to evolve and new capabilities will likely be developed.

PHYSICAL PRINCIPLES

      An MR system consists of the following components: 1) a large magnet to generate the magnetic field, 2) shim coils to make the magnetic field as homogeneous as possible, 3) a radiofrequency (RF) coil to transmit a radio signal into the body part being imaged, 4) a receiver coil to detect the returning radio signals, 5) gradient coils to provide spatial localization of the signals, and 6) a computer to reconstruct the radio signals into the final image.       The signal intensity on the MR image is determined by four basic parameters: 1) proton density, 2) T1 relaxation time, 3) T2 relaxation time, and 4) flow. Proton density is the concentration of protons in the tissue in the form of water and macromolecules (proteins, fat, etc). The T1 and T2 relaxation times define the way that the protons revert back to their resting states after the initial RF pulse. The most common effect of flow is loss of signal from rapidly flowing arterial blood.

      The contrast on the MR image can be manipulated by changing the pulse sequence parameters. A pulse sequence sets the specific number, strength, and timing of the RF and gradient pulses. The two most important parameters are the repetition time (TR) and the echo time (TE). The TR is the time between consecutive 90 degree RF pulse. The TE is the time between the initial 90 degree RF pulse and the echo.

 The contrast on the MR image can be manipulated by changing the pulse sequence parameters. A pulse sequence sets the specific number, strength, and timing of the RF and gradient pulses. The two most important parameters are the repetition time (TR) and the echo time (TE). The TR is the time between consecutive 90 degree RF pulse. The TE is the time between the initial 90 degree RF pulse and the echo.

      The most common pulse sequences are the T1- weighted and T2-weighted spin-echo sequences. The T1-weighted sequence uses a short TR and short TE (TR < 1000msec, TE < 30msec). The T2-weighted sequence uses a long TR and long TE (TR > 2000msec, TE > 80msec). The T2-weighted sequence can be employed as a dual echo sequence. The first or shorter echo (TE < 30msec) is proton density (PD) weighted or a mixture of T1 and T2. This image is very helpful for evaluating periventricular pathology, such as multiple sclerosis, because the hyperintense plaques are contrasted against the lower signal CSF. More recently, the FLAIR (Fluid Attenuated Inversion Recovery) sequence has replaced the PD image. FLAIR images are T2-weighted with the CSF signal suppressed.  The TR, matrix size and NEX are the only parameters that affect scan time. Increasing any one of these parameters increases the minimum scan time. Spatial resolution is determined by matrix size, FOV and slice thickness. Increasing matrix size or decreasing FOV and slice thickness increases spatial resolution, but at the expense of either decreased signal-to-noise or increased scan time. To obtain images of high resolution with high signal-to-noise requires longer scan times. All of the scan parameters affect signal-to-noise. The signal within an image can be improved by increasing TR, FOV, slice thickness and NEX or by decreasing TE and matrix size. The most direct way to increase signal is by increasing NEX, but one must keep in mind that increasing NEX from two to four, for example, doubles the scan time, but increases the signal by only the square root of two. Finally, TE does not affect scan time, however, it does determine the maximum number of slices in multislice mode. Increasing the TE or shortening TR decreases the number of slices that can be obtained with one pulse sequence. Specialized techniques for reducing motion and artifacts on the images also have applications for brain imaging. Gradient moment rephasing or flow compensation techniques effectively reduce ghost artifacts resulting from CSF flow. They should be used for T2-weighted spin-echo and gradient-echo acquisitions, but not with T1-weighted imaging because they increase the signal from CSF. Flow compensation techniques do not contribute to SAR (a measure of power deposition), but the extra gradient pulses lengthen the minimum TE, and gradient heating may limit the number of slices, the minimum FOV and slice thickness. Cardiac gating also reduces artifacts from CSF pulsations, resulting in superior object contrast and resolving power in the temporal lobes, basal ganglia and brain stem.

Saturation (SAT) techniques use extra RF pulses to eliminate artifacts from moving tissues outside the imaging volume, such as swallowing or respiratory artifacts, and from unsaturated protons that enter the imaging volume through vascular channels. SAT should be used for T1-weighted imaging of the sella, internal auditory canals, and the spine. The extra RF pulses cost SAR and take time, lengthening the minimum TR or decreasing the maximum number of slices in a multislice mode.

      Methods for eliminating wrap-around or aliasing should be prescribed when imaging small anatomic areas, such as the sella and internal auditory canals, with smaller FOR's. The "no phase wrap" option is most effective in the anterior-posterior direction for sagittal and axial scans. On newer MR systems, many of theses specialized techniques are automatically added to the appropriate sequences.

MR IMAGE CONTRAST

When reviewing an MR image, the easiest way to determine which pulse sequence was used, or the "weighting" of the image, is to look at the cerebrospinal fluid (CSF). If the CSF is bright (high signal), then it must be a T2-weighted imaged. If the CSF is dark, it is a T1-weighted image. Next look at the signal intensity of the brain structures.

      On MR images of the brain, the primary determinants of signal intensity and contrast are the T1 and T2 relaxation times. The contrast is distinctly different on T1 and T2-weighted images. Also, brain pathologies have some common signal characteristics.

      Pathologic lesions can be separated into five major groups by their specific signal characteristics on the three basic images: T2- weighted, proton density-weighted (PD)/FLAIR, and T1-weighted.

BRAIN IMAGING PROTOCOL

      Since studies have shown that T2-weighted images are most sensitive for detecting brain pathology, patients with suspected intracranial disease should be screened with T2-weighted spin-echo and FLAIR images. The axial plane is commonly used because of our familiarity with the anatomy from CT. The other scan parameters include a 256 x 256 matrix, 1 NEX, 22 cm FOR and 5 mm slice thickness for a scan time of less than 4 minutes and a voxel size of 5 x 0.86 x 0.86 mm. A 1-2 mm interslice gap prevents RF interference between slices. 

 If an abnormality is found, additional scans help characterize the lesion. Noncontrast T1-weighted images are needed only if the preliminary scans suggest hemorrhage, lipoma, or dermoid. Otherwise, contrast-enhanced scans are recommended. Gadolinium-based contrast agents for MR are paramagnetic and have demonstrated excellent biologic tolerance. No significant complications or side effects have been reported. It is injected intravenously at a dose rate of 0.1 mmol/kg. The gadolinium contrast agents do not cross the intact blood-brain barrier (BBB). If the BBB is disrupted by a disease process, the contrast agent diffuses into the interstitial space and shortens the T1 relaxation time of the tissue, resulting in increased signal intensity on T1-weighted images. The scans should be acquired between 3 and 30 minutes postinjection for optimal results.

  Contrast enhancement is especially helpful for extra-axial tumors because they tend to be isointense to brain on plain scans, but it also identifies areas of BBB breakdown associated with intra-axial lesions. Gadolinium enhancement is essential for detecting leptomeningeal inflammatory and neoplastic processes. Contrast scans are obtained routinely in patients with symptoms of pituitary adenoma (elevated prolactin, growth hormone, and so forth) or acoustic neuroma (sensorineural hearing loss). To screen for brain metastases in patients with a known primary, contrast-enhanced T1-weighted scans alone are probably sufficient.

      Gadolinium does not enhance rapidly-flowing blood. If vascular structures are not adequately seen on plain scan, the positive contrast provided by gradient-echo techniques or MR angiography may be helpful to confirm or disprove a suspected carotid occlusion or cerebral aneurysm, to evaluate the integrity of the venous sinuses, and to assess the vascularity of lesions. Gradient-echo imaging also enhances the magnetic susceptibility effects of acute and chronic hemorrhage, making them easily observable, even on low and mid-field MR systems.

Although the axial plane is the primary plane for imaging the brain, the multiplanar capability of MR allows one to select the optimal plane to visualize the anatomy of interest. Coronal views are good for parasagittal lesions near the vertex and lesions immediately above or below the lateral ventricles (corpus callosum or thalamus), temporal lobes, sella, and internal auditory canals. The coronal plane can be used as the primary plane of imaging in patients with temporal lobe seizures. Sagittal views are useful for midline lesions (sella, third ventricle, corpus callosum, pineal region), and for the brain stem and cerebellar vermis.

 Scan techniques are slightly different for the sella and cerebellopontine angle. For the sella, the plain and contrast enhanced scans are obtained in the coronal and sagittal planes using a smaller FOR and thin (3mm or less) contiguous or overlapping sections. For patients with a sensorineural hearing loss or suspected acoustic neuroma, contrast enhanced scans with T1-weighting are obtained through the internal auditory canals, again using thin overlapping sections.

CLINICAL INDICATIONS

      As imaging techniques of the brain, MR and CT are both competitive and complimentary. In general, CT performs better in cases of trauma and emergent situations. It provides better bone detail and has high sensitivity for acute hemorrhage. Support equipment and personnel can be brought directly into the scan room. CT scanning is fast. Single scans can be done in 1 second, so that even with uncooperative patients, adequate scans usually can be obtained. CT is more sensitive than MR for subarachnoid hemorrhage. CT is also more sensitive for detecting intracranial calcifications.

MR, on the other hand, functions best as an elective outpatient procedure. Proper screening of patients, equipment, and personnel for ferromagnetic materials, pacemakers, etc. is mandatory to avoid possible catastrophe in the magnet room. If proper precautions are in place, emergency studies can be done, but the set-up time is longer, and the imaging also requires more time. With conventional MR systems, most pulse sequences take a minimum of 2 minutes. Echo-planar capability has become standard on most MR systems, and this advanced technology can acquire sub-second MR scans.

 Due to its high sensitivity for brain water, MR is generally more sensitive for detecting brain abnormalities during the early stages of disease. For example, in cases of cerebral infarction,  brain tumors or infections, the MR scan will become positive earlier than CT. When early diagnosis is critical for favorable patient outcome, such as in suspected herpes encephalitis, MR is the imaging procedure of choice. MR is exquisitely sensitive for white matter disease, such as multiple sclerosis,  progressive multifocal leukoencephalopathy, leukodystrophy, and post-infectious encephalitis. Patients with obvious white matter abnormalities on MR may have an entirely normal CT scan. Other clinical situations where MR will disclose abnormalities earlier and more definitively are temporal lobe epilepsy, nonhemorrhagic brain contusions and traumatic shear injuries.

In general, nonenhancing disease processes are much more apparent on MR than CT. When the blood-brain barrier is damaged, enhancement occurs with both gadolinium and iodinated contrast agents on MR and CT, respectively. As a rule, the degree of enhancement is greater on MR scans.

 For evaluating posterior fossa disease, MR is preferable to CT. The CT images are invariable degraded by streaking artifacts from the bones at the skull base. In conjunction with gadolinium enhancement, MR can reliably detect intracanalicular acoustic neuromas and other schwannomas arising along the cranial nerves within the basal cisterns and foramina of the skull base. Similarly, MR has largely supplanted CT for imaging the sella turcica and pituitary gland.

     The value of MR for defining congenital malformations is unquestioned. The multiplanar display of anatomy gives important information about the corpus callosum and posterior fossa structures.  The superior gray/white contrast allows accurate assessment of myelination.

The phenomenon of flow void within arteries on spin-echo images, the high sensitivity for hemorrhage and hemosiderin deposition,  and the capability of MR angiography give MR distinct advantages over CT for imaging vascular disease. Vascular stenoses or occlusions, aneurysms,  and arteriovenous malformations can be imaged without intravenous contrast media. In cases of cryptic vascular malformations and cavernous angiomas, where the angiogram and CT scan are often negative, MR may reveal small deposits of hemosiderin from prior small hemorrhages.  Diffusion-weighted sequences are highly sensitive for restricted diffusion and cytotoxic edema associated with acute cerebral infarction. By combining conventional MR images with diffusion and perfusion-weighted imaging and MR angiography, a complete workup of vascular disease can be accomplished.

Along with the function of MR as a primary imaging procedure, there are indications for MR as a secondary procedure after the pathology has already been demonstrated by CT. In patients with solitary lesions on CT, in whom the diagnosis of metastatic disease, abscess, or multiple sclerosis would be strengthened by finding additional lesions, MR may resolve the issue. Similarly, in a patient with brain metastases in whom none of the lesions account for the patient's signs or symptoms, MR can help evaluate the particular anatomic area of interest. A potential problem in both of these circumstances is the nonspecificity of white matter hyperintensities, and contrast MR may be necessary to clarify the situation.

MR SPECTROSCOPY

MR spectroscopy provides a measure of brain chemistry. The most common nuclei that are used are 1H (proton), 23Na (sodium), 31P (phosphorus). Proton spectroscopy is easier to perform and provides much higher signal-to-noise than either sodium or phosphorus. Proton MRS can be performed within 10-15 minutes and can be added on to conventional MR imaging protocols. It can be used to serially monitor biochemical changes in tumors, stroke, epilepsy, metabolic disorders, infections, and neurodegenerative diseases. The MR spectra do not come labeled with diagnoses. They require interpretation and should always be correlated with the MR images before making a final diagnosis.

BASIC PHYSICAL PRINCIPLES

The resonant frequencies of nuclei are at the lower end of the electromagnetic spectrum between FM radio and radar. The resonant frequencies of protons range between about 10 MHz at 0.3 T to about 300 MHz on a 7 T magnet. The advantages of higher field strength are higher signal-to-noise and better separation of the metabolite peaks. In a proton spectrum at 1.5 T, the metabolites are spread out between 63,000,000 and 64,000,000 Hertz. Rather than use these large numbers, some very smart person decided to express the resonant frequencies in parts per million (ppm), and he/she positioned NAA at 2.0 ppm and let the other metabolites fall into their proper positions on the spectral line. Then, for unknown reasons, he/she reversed the ppm scale so that it reads from right to left.

For MR imaging, the total signal from all the protons in each voxel is used to make the image. If all the signal were used for MRS, the fat and water peaks would be huge and scaling would make the other metabolite peaks invisible. Since we aren't interested in fat and water anyway, the fat and water are eliminated. Fat is avoided by placing the voxel for MRS within the brain, away from the fat in bone marrow and scalp.Water suppression is accomplished with either a CHESS (CHEmical-Shift Selective ) or IR (Inversion Recovery) technique. These suppression techniques are used with a STEAM or PRESS pulse sequence acquisition. A Fourier transform is then applied to the data to separate the signal into individual frequencies. Protons in different molecules resonate at slightly different frequencies because the local electron cloud affects the magnetic field experienced by the proton.

The STEAM (STimulated Echo Acquisition Mode) pulse sequence uses a 90o refocusing pulse to collect the signal like a gradient echo. STEAM can achieve shorter echo times but at the expense of less signal-to-noise. The PRESS (PointREsolved SpectroScopy) sequence refocuses the spins with a180o rf pulse like a spin echo. Two other acronyms require definition. CSI (Chemical Shift Imaging) refers to multi-voxel MRS. SI (Spectroscopic Imaging) displays the data as an image with the signal intensity representing the concentration of a particular metabolite.

As in MR imaging, the echo time affects the information obtained with MRS. With a short TE of 30 msec, metabolites with both short and long T2 relaxation times are observed. With a long TE of 270 msec, only metabolites with a long T2 are seen, producing a spectrum with primarily NAA, creatine, and choline. One other helpful TE is 144 msec because it inverts lactate at 1.3 ppm.

As a general rule, the single voxel, short TE technique is used to make the initial diagnosis, because the signal-to-noise is high and all metabolites are represented.� Multi-voxel, long TE techniques are used to further characterize different regions of a mass and to assess brain parenchyma around or adjacent to the mass. Multi-voxel, long TE techniques are also used to assess response to therapy and to search for tumor recurrence. The brain metabolites that are commonly seen on the MR spectrum are listed on the right. Each metabolite appears at a specific ppm, and each one reflects specific cellular and biochemical processes. NAA is a neuronal marker and decreases with any disease that adversely affects neuronal integ-rity. Creatine provides a measure of energy stores. Choline is a measure of increased cellular turnover and is elevated in tumors and inflammatory processes. The observable MR metabolites provide powerful information, but unfortunately, many notable metabolites are not represented in brain MR spectra. DNA, RNA, most proteins, enzymes, and phospholipids are missing. Some key neurotransmitters, such as acetylcholine, dopamine, and serotonin, are absent. Either their concentrations are too low, or the molecules are invisible to MRS.

Normal MR spectra obtained from gray matter and white matter are shown on the right. The predominant metabolites, displayed from right to left, are NAA, creatine, choline, and myo-inositol. The primary difference between the two spectra is that gray matter has more creatine.� Hunter's angle is the line formed by the metabolites on the white matter spectrum.The common way to analyze clinical spectra is to look at metabolite ratios, namely NAA/Cr, NAA/Cho, and Cho/Cr. Normal and abnormal values are shown in the chart to the right. By including a known reference solution when acquiring the MR spectral data, absolute concentrations of metabolites can be calculated.

CLINICAL APPLICATIONS

Brain Tumors

MRS can be used to determine the degree of malignancy. As a general rule, as malignancy increases, NAA and creatine decrease, and choline, lactate, and lipids increase. NAA decreases as tumor growth displaces or destroys neurons. Very malig-nant tumors have high metabolic activity and deplete the energy stores, resulting in reduced creatine. Very hypercellular tumors with rapid growth elevate the choline levels. Lipids are found in necrotic portions of tumors, and lactate appears when tumors outgrow their blood supply and start utilizing anaerobic glycolysis. To get an accurate assessment of the tumor chemistry, the spectroscopic voxel should be placed over an enhancing region of the tumor, avoiding areas of necrosis, hemorrhage, calcification, or cysts.

Multi-voxel spectroscopy is best to detect infiltration of malignant cells beyond the enhancing margins of tumors. Particularly in the case of cerebral glioma, elevated choline levels are frequently detected in edematous regions of the brain outside the enhancing mass.Finally, MRS can direct the surgeon to the most metabolically active part of the tumor for biopsy to obtain accurate grading of the malignancy.

A common clinical problem is distinguishing tumor recurrence from radiation effects several months following surgery and radiation therapy. Elevated choline is a marker for recurrent tumor. Radiation change generally exhibits low NAA, creatine, and choline on spectroscopy. If radiation necrosis is present, the spectrum may reveal elevated lipids and lactate.

MRS cannot always distinguish primary and secondary tumors of the brain from one another. As mentioned above, one key feature of gliomas is elevated choline beyond the margin of enhancement due to infiltration of tumor into the adjacent brain tissue. Most non-glial tumors have little or no NAA. Elevated alanine at 1.48 ppm is a signature of meningiomas. They also have no NAA, very low creatine, and elevated glutamates.

Cerebral Ischemia and Infarction

When the brain becomes ischemic, it switches to anaerobic glycolysis and lactate accumulates.Markedly elevated lactate is the key spectroscopic feature of cerebral hypoxia and ischemia. Choline is elevated, and NAA and creatine are reduced.� If cerebral infarction ensues, lipids increase.

Trauma

MR spectroscopy is not routinely used in the acute setting of head injuries. CT and MR imaging demonstrate the fractures and intracranial hemorrhage that require emergent surgical intervention. On the other hand, when the patient has stabilized, MRS is helpful to assess the degree of neuronal injury and predict patient outcomes. Especially in the case of diffuse axonal injury, imaging often underestimates the degree of brain damage. Clinical outcome correlates inversely with the NAA/Cr ratio. The presence of any lactate or lipid indicates a worse prognosis.

Infectious Diseases

As in the case of non-glial tumors, brain abscesses destroy or displace brain tissue, so NAA is not present. The voxel should include the abscess cavity to detect the breakdown products of these lesions. Lactate, cytosolic acid, alanine, and acetate are characteristic metabolites in bacterial abscesses. Toxoplasmosis and tuberculomas show prominent peaks from lactate and lipids.

Clinical investigators of HIV infection and AIDS have been very interested in the potential of MRS for measuring the effects of HIV infection on the brain and neuro-cognitive function. Unfortunately, MRS has not proven very sensitive for detecting HIV encephalitis in the early stages of infection. On the other hand when patients start developing neurocognitive deficits and AIDS dementia complex, the MR spectra become positive, namely with elevated choline and reduced NAA. Choline is the best marker for the white matter abnormalities, and the extent of NAA depletion correlates directly with the degree of dementia. MRS is also very helpful in following patients and assessing the effects of anti-viral therapies.

There is also considerable interest in using MRS to distinguish the common focal brain lesions in AIDS patients. The most helpful marker is choline, which is elevated in lymphoma, but low or absent in toxoplasmosis, tuberuloma, and cryptococcoma. Toxoplasmosis is characterized by markedly increased lactate and lipids and depletion of normal brain metabolites. Tuberculoma and cryptococcoma are similar but with relatively little lactate. The spectrum for PML may be similar to lymphoma, but the imaging features are distinctly different and PML may have elevated myo-inositol.

Pediatric Metabolic Disorders

MRS has a very important role in diagnosing and monitoring patients with metabolic disorders.This group includes a long list of diseases that affect the gray and white matter to varying degrees. The names and terminologies of these disorders are confusing because they were derived from the pathologic literature before their metabolic defects were discovered. As the specific biochemical and enzyme defects are being elucidated, these diseases are being classified more appropriately. The list of disorders is long and beyond the scope of this syllabus. Some of the more important diseases are listed below, along with their specific metabolic markers on MR spectra.

Since most metabolic disorders present in infancy, it is important to understand the normal pediatric MR spectrum. Compared to the adult, newborns have much less NAA, and increased choline and myo-inositol. Progression to the adult pattern follows myelination.

Hepatic Encephalopathy

The spectrum of hepatic encephalopathy is characterized by markedly reduced myo-inositol.� Choline is also reduced, and glutamine is increased. Liver failure results in excess ammonia in the blood. Ammonia is a neurotoxin and causes increased conversion of glutamate to glutamine. Similar metabolic changes are seen in Reye's syndrome, an acute form of liver failure in infants. The metabolic changes of hepatic encephalopathy increase after a TIPS shunt procedure, and they revert back to normal after successful liver transplantation.

Alzheimer's Disease

Although MR spectroscopy is not highly sensitive for detecting early Alzheimer's disease, as the disease progresses, the spectrum becomes abnormal. Specifically, with advancing disease the NAA is reduced and myo-inositol becomes elevated. Since MRS is totally non-invasive and easily obtained, myo-inositol may become an important marker for assessing new therapies for this devastating disorder.

Myo-inositol is also increased in Down's syndrome, a dementia that presents in childhood and is pathogenetically similar to Alzheimer's disease. On the other hand, myo-inositol is not elevated in other adult dementia, so it is a helpful marker to distinguish Alzheimer's disease from the other causes of dementia.