Magnetic resonance imaging contrast agents: theory and application to the central nervous system.

J Neurosurg 73:820-839, 1990

Magnetic resonance imaging contrast agents: theory and application to the central nervous system RICHARD A. BRONEN, M.D., AND GORDON SZE, M . D .

Department of Diagnostic Radiology, Yale University School of Medicine, New Haven, Connecticut ~" The theoretical aspects of magnetic resonance (MR) imaging contrast agents are reviewed, and their current applications to the central nervous system (CNS) and their future applications are discussed. Profound differences exist between contrast agents used for MR imaging and computerized tomography (CT). In MR imaging, the contrast agents are not imaged directly but rather act on adjacent protons to shorten TI and T2 relaxation times. This in turn results in signal intensity changes. The lanthanide metal, gadolinium, in the form of gadopentetate dimeglumine, has been found to be both safe and efficacious as the only currently approved contrast agent for MR imaging. Magnetic resonance imaging revolutionized the detection and treatment of disease affecting the brain and spine. Initially, it was thought that signal characteristics on MR imaging would allow differentiation of specific pathology. It was soon found that MR studies were able to detect more abnormalities but were less able to characterize them. The recent development of contrast agents for MR imaging has allowed this modality to surpass CT for the evaluation of most CNS lesions. At present, contrast-enhanced MR imaging is generally accepted as the study of choice for evaluating acoustic neurinomas, pituitary lesions, meningeal disease, primary and secondary brain tumors, active multiple sclerosis, intradural spinal neoplasms, intramedullary spinal disease, and postoperative states in both the spine and brain. Even when contrast-enhanced CT can detect the same abnormalities, evaluation of the lesions in multiple planes on MR imaging can sometimes yield invaluable information, especially prior to surgery. Future developments of contrast material for MR imaging include non-gadolinium compounds, intrathecal contrast media, cerebral blood flow and volume evaluation, and, possibly, antibody-labeled contrast agents. KEY WORDS magnetic resonance imaging 9 contrast medium central nervous system lesion 9 brain neoplasm ~

T

HE recent Food and Drug Administration (FDA) approval of gadopentetate dimeglumine (Magnevist),* also formerly known as gadoliniumdiethylenetriaminepenta-acetic acid (Gd-DTPA), paves the way for contrast-enhanced magnetic resonance (MR) imaging to become a significant form of neuroimaging. Numerous studies have already indicated that postcontrast M R imaging is superior to all other modalities, including noncontrast-enhanced M R for certain conditions. 1~ This review examines the theoretical aspects of M R contrast enhancement, the specific properties of gadolinium (Gd) as a contrast agent, the present indications for GdD T P A in clinical practice, and the future prospects for contrast-enhanced M R imaging.

* Gadopentetate dimeglumine supplied by Berlex Imaging, Wayne, New Jersey.

820

9 gadolinium

9

Theoretical Aspects of M R Contrast Agents

Magnetic resonance spin-echo signal intensity depends on the equation: I = (H)f(v)[e-TE/T2][1 -- e-'rR/T'], where I is signal intensity, H is hydrogen or proton density, and f(v) is a function of both the speed and fraction of protons moving through the imaged region. 2~ Alterations of the following factors will increase signal intensity: shortened T~, prolonged T2, increased proton density, or a change in flow function. Agents which alter these parameters will alter contrast. The Tz and T2 relaxation times are the easiest parameters to manipulate. To understand how agents alter TI and T2, one must be familiar with the principles of magnetism.

Magnetism Magnetism results when a charged particle is in motion; for example, proton/electron spinning, electron J. Neurosurg. / Volume 73/December, 1990

Magnetic resonance imaging contrast agents

Fie. 1. Relationship of induced magnetism with an external magnetic field. Diamagnetic agents (D) have a weak negative magnetism while paramagnetic agents (P) have a positive magnetic effect. Superparamagnetic materials (SP) achieve a much greater positive magnetization than do paramagnetic materials. The arrows indicate that all three substances lose their magnetization completely when the external applied field is turned off. Ferromagnetic substances (F) retain positive magnetism even when the external field returns to zero (point a).

orbiting, or tumbling of a charged molecule. The particle generates a magnetic field called a "magnetic dipole moment" with both direction and magnitude. Electrons have a markedly larger magnetic moment than protons and thus form the basis for MR imaging contrast agents. Electrons with paired spins cancel each other's magnetic moments producing a net magnetic moment of zero. Unpaired electron spins produce a net magnetic dipole moment with the magnitude proportional to the number of unpaired electrons per molecule. 2~176 Magnetic susceptibility is the ability of a substance to become magnetized when placed in an external field. This can be quantified as a ratio of induced magnetization in the substance to that of the external field2~176 (Fig. 1). Magnetic susceptibility induces proton relaxation, resulting in contrast enhancement. Magnetic behavior can be classified into four categories: diamagnetic, paramagnetic, ferromagnetic, and superparamagnetic 2~176(Fig. 1). Almost all tissues are diamagnetic and have very weak negative susceptibility because they have paired electrons and therefore induce very little change in the local field. Compounds with unpaired electrons have a positive magnetic dipole moJ. Neurosurg. / Volume 73/December, 1990

ment and are either paramagnetic, ferromagnetic, or superparamagnetic. In an applied magnetic field a positive magnetic moment is induced in these compounds resulting in an attractive force. When the external field is removed, paramagnetic molecules resume random orientation, resulting in a net magnetization of zero. Paramagnetic agents include chromium, manganese (Mn), copper, Gd, organic free radicals, and molecular oxygen. 2~176 These are the most clinically interesting molecules since they are positive contrast agents and are water-soluble and therefore can be injected intravenously. Ferromagnetic substances retain a net magnetism after the external field is returned to zero. This is due to grouping of paramagnetic atoms or molecules into a crystal. With closely spaced molecules the dipole moments align parallel to each other. Ferromagnetic substances include iron (Fe), nickel, and cobalt. 2~176 Superparamagnetic substances are composed of "micro" ferromagnetic particles. They have a much greater magnetic susceptibility than paramagnetic substances because of their crystalline structure. When an external magnetic field is removed, however, they become ran821

R. A. Bronen and G. Sze domly oriented similar to paramagnetic molecules and thus retain no magnetization. The particulate nature of superparamagnetic substances makes them useful for imaging the reticuloendothelial system. ~~176176 Proton Relaxation Enhancement and Signal Intensity As previously mentioned, a ehange in T~ and T2 relaxation times will result in an alteration of signal intensity. Substances with strong magnetic susceptibility, such as paramagnetic agents, alter the local magnetic field within an external field, resulting in proton relaxation enhancement. The Tj relaxation time is shortened when the fluctuating magnetic fields match the Larmor frequency and facilitate the transition between high- and low-energy spin states. The fluctuating magnetic fields are created by the paramagnetic molecular tumbling and electron motion. The T2 relaxation time is shortened because inhomogeneities in the local magnetic field cause rapid dephasing of the transverse magnetization of local protons. A non-uniform field causes protons to precess at different frequencies with loss of transverse-phase coherence. 2~176 Proton relaxation enhancement shortens both T1 (increase in signal) and T2 (decrease in signal) relaxation times. These are conflicting changes in signal intensity. However, T~ effects predominate at lower, pharmacological concentrations of paramagnetic substances while T2 shortening predominates at higher levels 2~176 (Fig. 2). T2 shortening with the subsequent decrease in signal is the dominant effect of superparamagnetic and ferromagnetic substances with their much higher magnetic susceptibility. In clinical neuroimaging, the aim is to achieve signal enhancement by utilizing small doses of paramagnetic agents with predominant T~ shortening effects. It is important to realize that the increase in signal on a T~weighted image is not due to the paramagnetic agent itself but to the agent's effect on neighboring protons. Unlike the situation with iodinated contrast in computerized tomography (CT), increasing the dose of contrast medium will eventually lead to negative rather than positive enhancement when T2 shortening effects predominate56 (Fig. 2). Clinical Contrast Agents Soluble paramagnetic species include transitional metal complexes, the lanthanide (rare earth) complexes, organic free radicals, and molecular oxygen. 2~176 The transitional metals are strongly paramagnetic because the 3d orbit contains unpaired electrons. Examples of transition metals include Mn and copper. The lanthanide metals are paramagnetic because of unpaired electrons in the 4f orbit. The most clinically relevant ion in this series is Gd with seven unpaired electrons. It is one of the strongest paramagnetic ions. The greater the number of unpaired electrons, the stronger the magnetic moment, and therefore the greater the relaxation potential tends to be. Since these metal cations are very toxic, the metal is united with a ligand or chelated into 822

F i - - _ i T1W(500/14) 2000 I e ~ , D T2W(2000/88) 2400

/~.-~k,\

= ;ili~i...__~_il~i 1600

\\

I 200

0 I-

0.001

I

0.010

I

0.100

Concentration

I

1.000

---Xi---,,

IO000

(mmolar)

FIG. 2. Oadolinium-DTPAconcentrationversussignalintensity for T~-weighted(T1W) and T2-weighted(T2W) pulse sequences on a 1.0-teslasystem.The relationshipis nonlinear, with a decrease in signal at high concentrations. (Adapted from RungeVM, SchaibleTF, GoldsteinHA, et al: Gd-DTPA clinical efficacy.Radiographics 8:147-159, 1988, with permission.)

a metal complex which is nontoxic. Diethylenetriaminepenta-acetic acid (DTPA), ethylenediaminetetra-acetic acid (EDTA), and tetra-azacyclododecane tetra-acetic acid (DOTA) combined with Gd, Mn, Fe, and chromium have been investigated as parenteral agents. Gadopentetate dimeglumine, the chelate of DTPA with Gd, is the only potentially useful imaging substance at present approved by the FDA. Comparison of Gd complexes has shown that GdDTPA has greater T~ relaxivity (17%) 3 and contrast enhancement (45%) 55 than Gd-DOTA. Although GdDOTA has much greater in vitro stability (10 28) than Gd-DTPA (1023) or Gd-EDTA (1017), both Gd-DTPA and Gd-DOTA have a much greater safety index than iodinated contrast-enhanced CT. 3 A Belgian study47has recently shown Gd-DOTA to be safe and efficacious in clinical studies of central nervous system (CNS) lesions in 38 patients. Gadolinium-EDTA is almost as toxic as free Gd, which may be related to a greater amount in the form of free Gd ion due to Gd-EDTA's lower formation constant. Nitroxide stable free radicals are organic molecules with an unpaired electron which is shielded from bonding by bulky alkyl groups. Like Gd compounds, nitroxides have the advantage of low acute animal toxicity and the potential for conjugation with target-specific agents}~ They have been used successfully in some animal studies as contrast agents? ~ Molecular oxygen acts as if it had two unpaired electrons as opposed to two paired electrons; thus, it is paramagnetic. 2~176 While very safe, molecular oxygen is a poor contrast agent. Its concentration is difficult to manipulate. In vivo, oxygen may rapidly be converted to diamagnetic oxyhemoglobin.2~

Gadopentetate Dimeglumine: Physical Properties Gadolinium is chelated to DTPA because the free

ion is very toxic. There is no detectable in vivo disso-

J. Neurosurg. / Volume 73/December, 1990

Magnetic resonance imaging contrast agents ciation of the ion from DTPA, due to the high formation constant (1023) of the compound. 79 As a dimeglumine salt, Gd-DTPA has an osmolality of 1940 mOsm/ kg, six times that of plasma. Gadolinium is distributed in the intravascular and extracellular fluid compartments. TM It does not cross the blood-brain barrier or cell membranes because of its charge, high molecular weight, and hydrophilic nature. Gadolinium-DTPA is excreted by glomerular filtration, with 83% being eliminated within 6 hours. Minimal excretion occurs via the gastrointestinal tract. Gadolinium-DTPA has shown very low toxicity in both animal and clinical trials. 44"56"59"64 The 50% lethal dose (LDso), measuring the acute intravenous toxicity in animals, is 5 to 15 mmol/kg. This is 50 to 150 times the 0.1 mmol/kg dose in clinical patients, as opposed to a safety index (animal LDso/clinical dose) of 10 for iodinated contrast material. It is also a very poor activator of the complement system, which is thought to play a role in the induction of anaphylactoid reactions. There have been no significant adverse effects with Gd. 44'56'59'64The incidence of adverse reactions related to Gd-DTPA is as follows: headache 5%, nausea 3.4%, vomiting 2%, and irritation at site of injection 2%; all other reactions have occurred in less than 1% of cases (which includes two cases of hypotension in a study of 442 subjects). One study 59 noted a similar percentage of adverse reactions (21%) occurring in both the 60 patients receiving Gd-DTPA and the 28 receiving placebo (p > 0.99). A transient rise in the serum Fe and bilirubin levels in some patients is thought to be due to a minimal increase in hemolysis by the spleen because of Gd-DTPA-induced changes in the erythrocyte membrane. No mutagenic or teratogenic effects have been seen at levels up to 1.0 mmol/kg. Studies have shown Gd-DTPA to be safe and efficacious for imaging of the brain 56'59'64and spine, 67'76 and in pediatric patients?'1949 In the United States, gadopentetate dimeglumine was approved by the FDA for imaging of the brain in June, 1988, for the spine in May, 1989, and in pediatric patients aged 2 years or older in August, 1989. The effectiveness of Gd-DTPA as a contrast agent is determined by several factors. These include its hydration number, magnetic moment, correlation time, and its proximity to water protons. 5'56 Compared to other substances, Gd-DTPA is one of the most suitable choices in each of these respects. Gadolinium is one of the few metal ions which still allows access of water molecules to the ion after chelation. With chelation, however, the effective proton relaxation enhancement decreases. The concentration of Gd-DTPA must be about twice as high as free Gd to produce the same proton relaxation enhancement. 79 Parameters determining signal enhancement include paramagnetic concentration, pulse sequence, and field strength, as well as the choice of paramagnetic agent. Although the Gd-enhanced relaxation of both T~ and T2 results in opposite effects on signal intensity, the increased signal from shortened T~ relaxation predom-


inates on T~-weighted (short repetition time (TR), short echo time (TE)) spin-echo sequences and at lower concentrations of Gd. At the recommended dose of 0.1 mmol/kg, these T~ effects are dominant. At concentrated levels, such as in the dependent portion of the urinary bladder, T2 effects predominate with loss of signal from the urine (Fig. 2). 52 Therefore, the contrast agent dosage is important for obtaining optimal studies. The optimal TR for a spin-echo pulse sequence is one between the TI of the tissue before and after contrast enhancement. Comparison of Gd-DTPA with iodinated agents shows many similarities but a few important differences. Both Gd-DTPA and iodinated contrast medium have similar molecular weight and pharmacokinetics with rapid intravascular and extracellular compartment distribution as well as excretion primarily by the kidneys. Both will enhance intracranial lesions when the blood-brain barrier is broken. However, a much smaller dose of Gd-DTPA is needed, which results in a much greater safety index and a relative paucity of side effects. Runge, et al., 52 suggested that Gd-DTPA-enhanced MR imaging is more sensitive than iodinated contrast-enhanced CT at detecting early disruption in the bloodbrain barrier.

Clinical Applications: The Brain Enhancement of Normal Brain Structures Knowledge of normal contrast patterns is important to differentiate normal from pathological enhancement and to confirm adequate enhancement with Gd-DTPA (see Figs. 5, 7, 8, and 9). The intracranial contents with an intact blood-brain barrier do not enhance with Gd-DTPA. The following structures have an incomplete barrier and routinely show enhancement: pituitary gland, infundibulum, intracavernous portions of the third through sixth cranial nerves, cavernous sinuses, choroid plexus, and the retinal choroid. 33 Peak enhancement is at 3 minutes, but enhancement persists for approximately 1 hour. The pineal, falx cerebri, tentorium, and walls of the large vessels show enhancement inconsistently. Enhancement is seen in small vessels, usually venous. Rapidly flowing blood (in arteries) is imaged as a signal-void area which does not enhance. There is a 5 % quantitative increase in the signal of gray matter after injection of Gd that cannot be visualized qualitatively. Inner and middle ear structures, including the facial nerve, do not normally take up contrast medium. 13~33Extracranially, the mucosa of the paranasal sinuses, nasopharynx, and oropharynx consistently show enhancement with GdDTPA. Overview of Clinical Utilization Gadolinium-DTPA has the potential to increase detection of intracranial lesions, improve diagnostic ability, and reduce imaging time. Five reports 4'21~29'59'64 involving 283 patients have shown that Gd-DTPA823

R. A. Bronen and G. Sze

F1G. 3. Magnetic resonance images of an anaplastic glioma. Left: Long-TR/TE axial MR image showing increased signal abnormality and associated mass effect in the right frontal and parietal lobes. It is impossible to separate "gross" lesion from edema. Center: Short-TR/TE coronal image demonstrating the mass effect on the right frontal sucli and right lateral ventricle. It is again difficult to separate the lesion from edema. Hyperintensity (arrow) at the posterior margin is consistent with hemorrhage. Right: Gadolinium-DTPAenhanced short-TR/TE image revealing an enhancing lesion consistent with tumor. This contrast-enhanced scan clearly separates gross tumor from edema. Hemorrhage cannot be distinguished from enhancement.

enhanced MR studies provided additional information in up to 50% of patients when compared to precontrast MR images. Gadolinium-DTPA increased the contrast of lesions, helped distinguish tumor from edema and normal brain, differentiated intra-axial from extra-axial lesions, and improved differential diagnosis. A blinded retrospective MR study 56 of 55 lesions confirmed that postcontrast MR images improved diagnostic accuracy, particularly of metastases and meningiomas, as compared to the precontrast TI- and T2-weighted MR studies, with a decrease in false negatives from 8.6% to 0.7%. The authors also substantiated that contrastenhanced MR imaging was better than noncontrast MR studies at distinguishing approximate tumor boundaries from edema. Since Gd-DTPA is functionally similar to iodinated contrast material, enhancement patterns are similar. Gadolinium-DTPA MR images may show enhancement slightly earlier and more frequently than contrastenhanced CT. A greater number of lesions may be seen with Gd-DTPA in patients with active multiple sclerosis, metastases, hemangioblastomas, meningeal lesions, ventricular abnormalities, pituitary adenomas, and intracanalicular neurinomas.4' 10,~6,~7.26.28,29,58.59,64,65,74However, CT remains the best modality to diagnose calcified lesions and small calvarial abnormalities. Some predictions for contrast-enhanced MR imaging have been overly optimistic. Postcontrast T~-weighted images have a lower sensitivity in general than precontrast T2-weighted images. 4'29'64 T2-weighted studies are very sensitive to changes in tissue water content which 824

may be due to minimal disruption of the blood-brain barrier. Gadolinium-DTPA, like iodinated contrast material, is a macromolecule which will accumulate in the extravascular space only if there is a significant breakdown in the blood-brain barrier. Thus, postcontrast TIweighted images are generally used in conjunction with T2-weighted images for screening purposes. For example, inactive multiple sclerosis plaques often do not enhance with contrast m e d i u m . 25'26 Nevertheless, detection of these lesions on the T2-weighted sequence is often very important, particularly in the initial diagnosis. As with iodinated contrast material, Gd-DTPA may obscure certain types of pathology. On contrast-enhanced T l-weighted images, hemorrhage within a lesion may be missed (Fig. 3), Enhancing skull-base lesions, such as glomus tumors or bone metastases, may be obscured by the adjacent bright signal of fat or bone marrow. 27'67 While it was originally feared that lesions which are normally visible on T2-weighted images by virtue of their prolonged T2 relaxation times may become isointense to brain with Gd-induced T2 shortening, 4 this appears to be a rare situation. As stated previously, the dosage of contrast medium is important for obtaining optimal studies. The optimal dose for imaging intracranial tumors in a comparison study of 11 patients has been found to be 0.1 mmol/ kg. 44 If one-half of this dose is used, there is inadequate tumor enhancement. A 0.2-mmol/kg dose is safe and may be useful in selected cases such as metastases, but is greater than the current recommended dose. 44,64 J. Neurosurg. / Volume 73/December, 1990


FIG. 4. Magnetic resonance images of metastasis from lung. Metastatic lesions are not detected on the long-TR/TE image (left) or the long-TR/short-TE study (center), but are easily visualized on the Gd-DTPAenhanced short-TR/TE image (right).

Pathological Conditions Primary Brain Tumors. Gadolinium-DTPA enhancement represents the site of breakdown of the blood-brain barrier and not necessarily the tumor margin. Microscopic tumor is known to occur within areas of edema in patients with high-grade gliomas. Contrast is helpful in separating "gross" tumor margins from surrounding edema, a differentiation which is often difficult on noncontrast MR images (Fig. 3). This may beimportant for surgical planning and for posttherapeutic evaluation. Enhancement patterns are variable, similar to CT. Noncontrast T2-weighted studies are more sensitive, especially for the nonenhancing gliomas. 4'21'64 Stack, et al., 64 reported that 40 primary brain tumors all showed an increased signal on T2-weighted images, whereas six lesions were poorly identified on Gd-DTPAenhanced T,-weighted images. Gadolinium-DTPA-enhanced MR imaging has detected subependymal spread of a glioma missed with contrast-enhanced CT.29 Metastases. Contrast-enhanced MR imaging is equal to or better than contrast-enhanced CT in the evaluation of intracranial metastases: more lesions are detected with better contrast. 4,28,29,58,64,73This is particularly true of lesions adjacent to bone such as leptomeningeal metastases and of punctate metastases? TM In a study of 50 consecutive patients with suspected intraparenchymal metastases imaged with MR and CT studies (both contrast-enhanced), short-TR contrastenhanced MR images showed more lesions in eight patients. 73 Contrast-enhanced Tj-weighted studies can be more sensitive than noncontrast T2-weighted images for imaging small sites of blood-brain barrier breakdown without significant edema. In six patients from this same series, 73 short-TR postcontrast MR images detected metastases not seen on noncontrast MR (Fig. J. Neurosurg. / Volume 73 /December, 1990

4). Punctate metastases, particularly in the gray matter, can be more easily detected after Gd administration; these lesions often appear isointense under all sequence parameters if contrast material is not employed. In addition, noncontrast T2-weighted images can be equivocal because the nonspecific signal abnormalities in the periventricular area and white matter often found in older patients can mask metastases? 8 Finally, if several metastases are adjacent to each other, contrast-enhanced MR imaging is much better at discriminating the actual number of lesions than noncontrast MR imaging. Without contrast enhancement, the copious edema associated with metastases often obscures small lesions. Since neurosurgical resection of single intraparenchymal metastases is sometimes performed, administration of Gd can be essential in the MR evaluation. Although most metastases may be detected at 5 minutes postinjection, the peak detection time may be 30 minutes or longer after injection. 4,28,58 Metastatic lesions to the skull were investigated by West and colleagues8~ in 14 patients and 60 control subjects. The normal intradiploic space shows a wide variability in signal intensity due to fatty marrow, vascular spaces, and focal cortical bone. There is generally some symmetry of the intradiploic space from side to side, although it is inhomogeneous. The signal intensity of the calvaria changes with age. Bone marrow in young children appears as a uniform hypointense signal within the calvaria and clivus. 45 After the age of 7 years, children exhibit increasing fatty infiltration of the marrow. Thus, the signal intensity within the calvaria or clivus becomes more hyperintense on short-TR images. The normal diploic space does not enhance except for diploic venous channels and dura related to pacchionian granulations. Calvarial metastases generally enhance with Gd-DTPA but, since the intradiploic 825


FIG. 5. Magnetic resonance images of a tentorial meningioma. Left: The meningioma (arrow) is isointense with gray matter on this short-TR/TE coronal image. It is not easily seen and impossible to differentiate from an intra-axial lesion. Right: With Gd-DTPA enhancement, the lesion is easily detected and definitely extra-axial. There is normal enhancement of the choroid plexus (large arrow) and pineal (small arrow).

spaces are heterogeneous, subtle metastases may be missed if postinfusion studies are not compared to preinfusion images. Noncontrast studies are better at detecting tumor extension to fatty marrow regions, such as the clivus. 8~

Extra-Axial Tumors. Because extra-axial tumors are often isointense with normal brain on multiple pulse sequences, noncontrast M R imaging can be less sensitive than contrast-enhanced imaging by CT or MR study. Enhancement occurs immediately post-injection and persists for at least 1 hour. Gadolinium-DTPA enhancement may occasionally obscure lesions (such as glomus tumors or chordomas) which are adjacent to the bright signal of the nasopharynx or bone marrow. Computerized tomography remains the best imaging modality for demonstrating calcification, hyperostosis, and bone erosion. 1,6.21,27,64 Gadolinium-DTPA-enhanced MR studies of meningiomas have significant advantages over noncontrast MR images and contrast-enhanced CT scans in four areas (Fig. 5): 1) contrast-enhanced MR imaging is the best modality for evaluating difficult regions such as the planum sphenoidale, parasellar area, and the foramen magnum; 2) since arteries and patent venous sinuses can be seen coursing through or adjacent to these tumors on contrast-enhanced MR images, vascular displacement and encasement is easily noted; 3) multiplanar evaluation may provide important additional information for the surgeon; and 4) dural extension (en plaque meningioma) may be better detected, especially on slightly delayed studies when contrast washes out of the normal dura (however, differentiation of tumor from reactive change may be difficult). 1,21,27,64 Large neurinomas can be easily detected on either contrast-enhanced CT or MR imaging, with or without contrast enhancement. Acoustic neurinomas smaller 826

than 1.5 cm are better seen on MR images than CT scans (Fig. 6). Several studies indicate that MR imaging is equal to or more successful than air-cisternography CT for detecting acoustic neurinomas.10'65 The use of postcontrast MR imaging was evaluated by Curati, et al., 1~ in 10 patients with 12 presumed acoustic neurinomas (three surgically proved). Computerized tomography detected only five of the 12 lesions; the lesions not detected were usually less than 1.5 cm in diameter. Air-cisternography CT was positive in five of six patients who underwent this type of procedure. Magnetic resonance imaging was positive in all cases; the detection rate with noncontrast MR studies was equivalent to that after contrast enhancement; However, the lesions were markedly more conspicuous. In another study by Stack, et al.,65 involving 20 patients with 21 acoustic neurinomas, all five tumors less than 8 m m in diameter could only be detected on CT with air cisternography. Noncontrast MR images detected all 21 tumors, although three were visualized poorly. Contrastenhanced MR imaging clearly delineated the tumor and distinguished it from the surrounding structures in all 21 patients, improving confidence of diagnosis. Eighteen acoustic neurinomas were confirmed surgically. These studies by Curati, et al., and Stack, et al., have demonstrated that with low-strength magnet systems, postcontrast MR imaging appears to be better than CT, and equivalent or better than CT with air cisternography or precontrast MR imaging. Since magnets of high field strength yield better resolution, the proportional benefit of contrast-enhanced MR imaging over noncontrast MR studies is less certain. However, the general feeling throughout the neuroradiology community is that postcontrast MR imaging is a great benefit for small lesions, regardless of the field strength. In addition to a potential increased detection rate, the increase in clarity of the lesion leads to improved ob-



m intracanalicular rt-TR/TE coronal image (right) are ~tected within the short-TR/TE cotows an enhancing : internal auditory an acoustic neuri-

server confidence in the diagnosis. Thus, contrast-enhanced MR imaging appears to be the best type of study for small intracanalicular acoustic neurinomas, small facial-nerve and other neurinomas, and recurrent tumors. ~,t~ Pituitary A d e n o m a s . Macroadenomas are well visualized on noncontrast MR studies. After Gd administration, they will enhance; however, extension into the cavernous sinus may be masked on postcontrast MR studies.~4 Microadenomas can also be visualized on noncontrast MR but the use of contrast can improve detection (Fig. 7). At intermediate field strengths, Dwyer, et al., ~7 showed that contrast-enhanced MR imaging was superior to both plain MR and postcontrast CT studies in detecting microadenomas in patients with Cushing's disease. Ten of 12 microadenomas were seen with GdDTPA MR enhancement, whereas eight were detected with noncontrast MR studies and only four with CT. Doppman and colleagues 16 used a 1.5-tesla magnet and Gd-DTPA for MR evaluation of eight patients with surgically proven adrenocorticotropic hormone-secreting microadenomas. Microadenomas were found in five of the eight patients at surgery. Tumor was found in the other three only after hemihypophysectomy and microscopic examination. Noncontrast as well as postJ. Neurosurg. / Volume 73/December. 1990

contrast MR studies were negative in these three; there was one false-positive contrast-enhanced MR image. Three microadenomas were imaged both by noncontrast and postcontrast MR images. The contrast-enhanced MR images detected two microadenomas not found by noncontrast MR studies. Newton, et al.,43 examined 1 1 patients with surgically proven microadenomas. Noncontrast MR studies correctly identified the lesion in two of six patients with Cushing's disease, in three of four patients with hyperprolactinemia, and in a single patient with acromegaly. Contrast-enhanced MR imaging correctly detected another two patients with Cushing's disease and an additional patient with hyperprolactinemia. In the one falsepositive examination of a patient with Cushing's disease (wrong site identified) obtained by both noncontrast and postcontrast MR imaging, contrast-enhanced CT and venous sampling were better at localizing the correct side of the tumor. These studies 16'17'43 concluded that contrast-enhanced MR imaging is indicated for evaluating microadenomas if the noncontrast MR study is normal. However, incidental nonfunctioning cysts or adenomas in patients with Cushing's disease can lead to false-positive results. There are two time periods to image pituitary microadenomas with Gd-DTPA: 1) immediately (3 rain827


FIG. 7. Magnetic resonance images of a hyperprolactin pituitary microadenoma: short-TR/TE coronal images of the pituitary glands. Left: Noncontrast image showing subtle enlargement of the left side of the gland without a focal signal abnormality. Right: Gadolinium-enhanced study, obtained immediately postinjection, clearly identifying the hypointense microadenoma with a hyperintense center on the left side of the gland (black arrows). It is easy to measure the size of the 6-ram microadenoma on this study. The adjacent normal pituitary tissue (white arrows) is markedly enhanced. There is normal enhancement of the cavernous sinus more laterally.

utes postinjection) when the microadenoma enhances to a very limited extent compared to the adjacent normal pituitary gland and therefore the t u m o r appears hypointense (Fig. 7); and 2) delayed (55 minutes postinjection) when there is delayed enhancement of the microadenoma and mild decrease in the enhancement of the normal pituitary resulting in relative hyperintensity of the microadenoma. 6'~7 In practical terms, imaging of microadenomas after injection of contrast medium should be performed immediately and must be accomplished quickly. Acquisition parameters for rapid postcontrast pituitary M R studies are still in a state of transition. If no abnormality is seen in a patient strongly suspected of having a microadenoma, a delayed image can be performed approximately 1 hour after contrast injection, as recommended by Dwyer, et al., ~7 although other investigators have not found delayed examinations to be helpful. 16'43

Meningeal Disease. Gadolinium-DTPA may mildly enhance normal meningeal structures. 33'74In a series of 20 patients without meningeal disease, Sze, eta/., TM found fine linear enhancement of normal meninges in short segments, especially in the parasagittal and anterior portions of the cranium. Plain MR imaging and contrast-enhanced CT are comparatively insensitive to meningeal disease. Recent findings suggest that Gd-DTPA-enhanced M R images are more sensitive at detecting tumors and inflammatory and traumatic changes of the leptomeninges, n'38'74 While subtle enhancement adjacent to the skull may not be seen on postcontrast CT scans, these changes are easily detected with M R imaging (Fig. 8). Mathews, eta/., 38 evaluated experimental bacterial meningitis in four dogs with postcontrast CT and MR 828

FIG. 8. Magnetic resonance images showing meningeal enhancement in bacterial meningitis. Upper."The noncontrast axial short-TR/TE (left) and long-TR/TE (right) images do not disclose a meningeal abnormality, although mild prominence of the ventricles is present. A shunt track is seen entering the right occipital horn. Lower Left: This postcontrast axial short-TR/TE image demonstrates marked enhancement of the meninges surrounding the entire brain (arrows). Enhancement of the falx may be a normal finding, although in this case the enhancement is pathological. Lower Right: After treatment with antibiotics, this follow-up postcontrast axial image shows resolution of the previously noted meningeal enhancement. If fibrous change in the meninges occurs, however, enhancement after meningitis or other meningeal pathology may be permanent. The contrast enhancement of vessels with slow-flowing blood and of the superior sagittal sinus walls is normal. studies. Computerized tomography demonstrated abnormal enhancement in only one of the four animals, 28 hours after induction of meningitis; however, M R imaging was able to demonstrate abnormal enhancement in all four dogs between 13 and 29 hours after inoculation. The areas of abnormal enhancement on the M R images were found to demonstrate severe inflammatory changes histologically. In areas that were mildly inflamed at pathological examination, there was no abnormal enhancement on M R evaluation. Complications of meningitis such as ventriculitis and cerebritis were more effectively demonstrated on MR images than CT scans. Unenhanced M R images were not useful in detecting meningitis. Sze, et a/., TM found abnormal meningeal enhancement in 52 of 175 contrast-enhanced M R studies of the brain. Twenty-two patients had carcinomatosis meningitis, 14 cases were postoperative or posttraumatic, while other etiologies included infection, sarcoid-



FIG. 9. Imaging of the residual cerebellar astrocytoma 1 year postoperatively. Upper: Axial contrast-enhanced computerized tomography (CT) scan showing abnormal tumor enhancement in the cerebellar vermis. As commonly encountered on CT, this scan of the posterior fossa is degraded by artifact from adjacent bone. Center: Noncontrast long-TR/ short-TE (left) and long-TR/long-TE (right) magnetic resonance (MR) images. There is no clear distinction between residual tumor and postoperative changes. Lower: ShortTR/TE MR images without (left) and with (right) Gd-DTPA enhancement. Gadolinium easily detects the enhancing residual tumor nodules, with a cystic component. Enhancement of the cystic rim probably indicates neoplasm in the wall. Note the normal enhancement of the infundibulum (arrow). osis, subdural hematoma, subacute infarction, or adjacent skull or parenchymal metastasis. Compared with the short linear enhancement seen normally, abnormal meninges demonstrated longer segments of enhancement which were sometimes nodular and more peripheral (that is, away from the parasagittal region). Abnormal enhancement was seen in up to two-thirds of patients with proven leptomeningeal carcinomatosis.

J. Neurosurg. / Volume 73 / December, 1990

When dura adjacent to the skull was involved, contrastenhanced MR imaging demonstrated abnormal meningeal enhancement better than postcontrast CT because bone obscures adjacent contrast enhancement on CT. Contrast-enhanced CT was equal to contrast-enhanced MR imaging in detection of meningeal nodules. Postoperative Lesions. Residual or recurrent tumor in patients who have undergone surgery may be difficult to detect and differentiate from postoperative changes on MR imaging. Contrast-enhanced MR imaging has had a major impact on these cases, regardless of the etiology of the original t u m o r . 2'27'29'3~ In a study of 15 postoperative patients, Bird, et al.,2 reported six patients in whom postcontrast MR studies showed recurrent tumor not identified on the noncontrast MR images (Fig. 9). A few studies have evaluated the temporal evolution of MR enhancement in the postoperative state] 8'69 As with postcontrast CT, MR enhancement in the hypervascular brain parenchyma adjacent to the resection can be seen postoperatively for several months. Contrast-enhanced MR imaging differs from contrast-enhanced CT in two ways, however. First, contrast enhancement is often seen slightly earlier and somewhat later than in postcontrast CT. Second, as hemorrhage evolves, it becomes hyperintense on MR imaging and thus is more easily confused with Gd-DTPA contrast enhancement, whereas on CT, hemorrhage at the resection site often presents less of a problem after the first few days since the blood becomes isodense. For these reasons, in the immediate postoperative situation, it is important to perform Gd-DTPA MR imaging even sooner than postcontrast CT. Initially, during the first few days after an operation, very little MR enhancement is seen at the operative site. However, after several days, thin linear enhancement is noted at the edge of the resected tissue. This linear appearance is difficult to confuse with tumor. With time, however, the enhancement can become progressively nodular and can be easily confused with tumor. From approximately 5 or 6 days to several months after surgery, enhancement that is seen at the operative site can be due to either tumor or to postoperative change. During this time, it is difficult to differentiate one from the other. In the chronic postoperative condition, enhancement at the postoperative site fades; however, the time course for this evolution differs between postcontrast CT and postcontrast MR imaging. With postcontrast CT, enhancement due to surgical change is generally not seen after 6 months; with contrast-enhanced MR imaging, however, enhancement is often seen 6 to 8 months postoperatively. Therefore, contrast-enhanced MR evaluation for possible residual or recurrent tumor must be viewed with skepticism if performed within 8 months after an operation.18'69

Infection. Contrast enhancement of cerebral infections is similar on both CT and MR imaging. Animal studies suggest that Gd-DTPA enhances more conspicuously and at an earlier stage in the development of 829


FIG. 10. Axial magnetic resonance images in a case of multiple sclerosis. Left: Short-TR/TE image showing relative insensitivity to plaques, which are hypointense. Center." Long-TR image showing good sensitivity to plaques, which appear hyperintense. This study cannot distinguish active from inactive lesions. Right: Gadolinium-DTPA-enhanced short-TR/TE image revealing active plaques with disruption of the bloodbrain barrier. Fewer lesions are identified than in the study shown center.

cerebritis; this has been observed as early as 1 day postinoculation in two of five dogs. 54Contrast enhancement was more easily detectable at 10 minutes than at 2 minutes after Gd-DTPA injection. Multiple Sclerosis. Enhancement in multiple sclerosis is thought to be due to breakdown of the bloodbrain barrier in active lesions. Enhancement of lesions on MR studies has been correlated with clinical symptoms in patients with active disease. Gadolinium-DTPA can often separate active from nonenhancing inactive disease as opposed to noncontrast T2-weighted images (Fig. 10). On long-TR images, both appear as areas of high signal. Gadolinium-enhanced MR is probably more sensitive than contrast CT. Grossman, et a[., 26 reported enhancement in all of nine symptomatic patients studied with Gd-DTPA as opposed to four of the same nine studied with contrast-enhanced CT. Most lesions were seen on the immediate postinjection images, although some became more visible on delayed studies. A follow-up study by Grossman, et al., 25 in 13 of these patients found that multiple sclerosis lesions are dynamic, with both active (enhancing) and inactive (nonenhancing) lesions showing marked changes on serial MR studies. Research by Kuharik, et al.,35 complements the findings by Grossman and colleaguesY '26 Experimental allergic encephalomyelitis, a demyelinating disease with similarities to multiple sclerosis, was induced in two of 12 dogs by Kuharik's group. During the acute phase, periventricular white matter lesions were imaged on T~weighted MR images. These lesions enhanced with Gd830

DTPA and were found to represent new active plaques of demyelination at necropsy. One dog had pathologically proven subacute demyelinating lesions, 6 weeks old. While these subacute plaques were indistinguishable from the acute lesions using long-TR studies, postcontrast examinations were able to separate the nonenhancing subacute lesions from the enhancing acute lesions. As clinical symptoms resolved, enhancement of lesions disappeared. Infarction. Noncontrast T2-weighted MR imaging is more sensitive to early infarction than CT or postcontrast MR imaging. T2-weighted MR studies are very sensitive to the increase in water content seen acutely in ischemic tissue. After several days, capillaries lacking an intact blood-brain barrier are seen at the periphery of an infarct. Although enhancement by Gd-DTPA has been seen as early as 16 hours experimentally39 and 22 hours in patients, 32 clinically the abnormal appearance is seen best several days after infarction and usually persists up to 6 to 8 weeks. 77 Postcontrast MR imaging can detect infarctions earlier and with better enhancement than postcontrast CT but may not yield any more significant information than CT. 32 The enhancement pattern may help with the differential diagnosis of lesions seen on noncontrast MR imaging. This is especially true in separating chronic infarcts or asymptomatic white matter lesions (often called "UBO's" or "unidentified bright objects") which do not enhance from lesions with active disruption of the blood-brain barrier, such as subacute infarctions, neoplasms, or inflammatory lesions. J. Neurosurg. / Volume 73/December, 1990

Magnetic resonance imaging contrast agents Skull-Base Lesions. Skull-base lesions are sometimes difficult to image on CT scans since the bone at the skull base may obscure or mask pathology. Sagittal or direct coronal planes are often the most optimal means of imaging skull-base lesions, yet they can be difficult to obtain with routine CT scanning. Due to its multiplanar capability, MR imaging has been particularly useful in these situations. Although contrast can obscure lesions adjacent to fat, recent reports ~2'5~ indicate that contrast-enhanced MR studies can be more useful than noncontrast MR studies in some cases involving the skull base. In a report on 27 patients with extracranial head and neck lesions, Robinson and colleagues 5~described three conditions in which Gd-enhanced MR imaging was more useful than noncontrast MR studies. First, perineural tumor spread at the level of the skull base was better seen in one patient. Second, Gd-enhanced studies more accurately documented neoplasm extending to the leptomeninges in three patients. Third, tumor could be more easily differentiated from inflammatory disease of the nasal and paranasal sinuses in four of their patients. Daniels, et al., 12 reported Gd-enhancement in nine of 10 patients who had facial-nerve paralysis but no facial-nerve tumors. Three of the four patients had idiopathic Bell's palsy and the other six had facial-nerve paralysis after temporal bone surgery. Thus, the authors stressed that enhancement of the intratemporal facial nerve is a nonspecific finding and can be seen with inflammatory changes as well as with tumors. This type of information could not be acquired on noncontrast MR or contrast-enhanced CT examinations. Vogl, et al., TM examined 26 patients with paragangliomas comparing postcontrast CT with noncontrast and postcontrast MR studies. There were 16 glomus jugulare tumors, three glomus tympanicum tumors, and seven carotid body tumors in their study. They claimed that, in three of five tumors less than 1.5 cm in size, contrast MR imaging was able to detect tumors not seen on nonenhanced MR studies. They found that MR imaging was better than CT because it displayed anatomy, vascularity, and tumor delineation better in all cases except the three with glomus tympanicum tumors. However, it was not stated whether the detection rate for CT and MR imaging was the same. Dynamic contrast-enhanced MR studies using gradient echo imaging can help confirm the diagnosis of very vascular lesions such as paragangliomas by their typical enhancement-versus-time profile/8 Skull-base lesions such as meningiomas, clivus chordomas, pituitary lesions, or even primary brain tumors may all be detected on noncontrast MR studies or postcontrast CT scans. However, the combination of contrast enhancement with the multiplanar capacity of MR imaging can often yield invaluable information, especially in preoperative cases. Computerized tomography still remains the best method for detecting bone abnormalities at the skull base. J. Neurosurg. / Volume 73/December, 1990

Clinical Applications: The Spine Postcontrast CT studies of the spine have always been plagued by bone artifact which often masks enhancement. With the development of contrast agents for MR imaging, an exquisite way now exists to evaluate the enhancement properties of the spine and its contents without interference by bone. Enhancement patterns of normal spinal structures and of extradural, intradural extramedullary, and intramedullary disease are reviewed in this section.

Enhancement of Normal Spinal Structures The most marked enhancement is seen in slowflowing vascular structures, especially the epidural and basivertebral plexus 7'82 (Fig. 11 right). However, these enhance inconsistently. The blood-CNS barrier is responsible for visualization of little to no enhancement within the spinal cord and intradural segments of the nerve roots. 6v Once the nerve roots exit the thecal sac, they often enhance and in fact the dorsal root ganglion enhances markedly. Minimal quantitative enhancement of the vertebral bone marrow, ligamentum flavum, and adjacent muscle occur, but these are not generally detectable. Extradural Disease Neoplastic Disease. Unenhanced MR images are excellent at detecting tumor involvement of the vertebrae as well as the epidural space. 9'1~ Contrast MR studies may obscure pathology and thus are not indicated as the initial study of choice. 71 On unenhanced MR studies, vertebral body lesions are detected as lowintensity regions surrounded by the increased signal intensity of normal fat-containing marrow?'l 1Enhancement of spinal lesions with Gd is extremely variable. 7'66'7~ Depending on the degree of enhancement, lesions either remain hypointense to normal marrow or become isointense or hyperintense. The variability of enhancement can occur with different lesions even in the same patient. Gadolinium becomes a liability when the detectable low-intensity lesion enhances to become isointense with the surrounding marrow and is thus obscured (Fig. 11). The same considerations apply to tumor within the epidural space. Depending on the degree of enhancement, tumor may become isointense with either the adjacent vertebral bone marrow or the adjacent epidural fat (Fig. 11). In certain situations, however, Gd may be helpful during MR evaluation of extradural tumors. 7~ These advantages include: 1) characterizing unusual epidural lesions; 2) evaluating regions of more aggressive tumor for biopsy; 3) indicating areas of cord compression; and 4) possibly evaluating tumor response to therapeutic intervention. Contrast enhancement can be useful in more fully characterizing epidural lesions. For example, differentiation of intervertebral disc herniation from epidural tumor adjacent to a narrowed disc on noncontrast MR imaging is occasionally difficult. Since discs 831


FIG. I 1. Magnetic resonance images of metastases from a colon carcinoma. Left: Short-TR/TE sagittal image revealing multiple foci of hypointensity within the vertebral bodies (contrasted by adjacent normal hyperintense fatty marrow) consistent with known metastatic disease. Extension of the tumor at L-5 to the epidural space (arrow) is also appreciated. Center."Long-TR/TE cardiac-gated sagittal image demonstrating hyperintensity of the vertebral body metastases. Right." Gadolinium-DTPA-enhanced short-TR/TE sagittal image displaying varying degrees of enhancement of lesions. Some mildly enhancing lesions remain hypointense to normal marrow. Other lesions, especially at L-5, enhance enough to become isointense to the adjacent marrow. These isointense lesions would not be detected if the contrast-enhanced examination had been the only study performed. The epidural tumor also enhances to a similar extent as the adjacent tumor-infiltrated body of L-5. The enhancing thin line continuing in the anterior epidural space from L-5 superiorly to L-4 (arrowhead) may represent displaced enhancing epidural veins or extension of tumor. (Reprinted from Sze G, Krol G, Zimmerman RD, et al: Gadolinium-DPTA: malignant extradural spinal tumors. Radiology 167.'217-233, 1988, with permission.)

and disc fragments generally do not enhance on immediate postcontrast scans, tumor can be distinguished by its enhancement. 66 The administration of contrast material can also help in MR-guided biopsy. Contrast uptake may often be localized to only a portion of the vertebral body. It is thought that these areas of Gd enhancement indicate regions of more active tumor involvement. In our experience, biopsy of the enhancing portion of the abnormal vertebral body has been more successful than biopsy of the nonenhancing regions. 7~ Cord compression due to diffuse epidural tumor, especially in patients with a narrow spinal canal, is occasionally better delineated on an enhanced rather than unenhanced MR image. 7~ The nonenhancing cord is highlighted against enhancing tumor. However, highquality long-TR images or gradient echo studies may provide comparable information. Contrast enhancement may be useful as an index of tumor response to treatment. 7t Berry, et aL, t noted that prostatic carcinoma metastatic to the skull failed to enhance on MR imaging after radiation therapy. We have noted similar findings in spinal metastases which have been irradiated. Presumably, tumor that is responsive to radiation therapy becomes poorly vascularized and thus does not enhance. We have also noted cases with the reverse condition where tumor which was

832

unresponsive to radiation therapy continues to show enhancement.

Degenerative Changes and Disc Disease. Modic and colleagues4~have noted several types of MR signal intensity changes of vertebral bodies associated with degenerative disc disease. The early stage consists of decreased signal intensity compared to adjacent marrow on short-TR images and increased signal intensity on long-TR images (type 1 degenerative changes). It is believed that these findings are due to edema, inflammation, and fibrovascular changes. Enhancement by Gd often occurs at this stage and is usually mild. 3J'67 One should not mistake this enhancement for tumor or infection. Later stages of degenerative disease result in an increased signal intensity on short-TR images and an isointense to slightly hyperintense signal on long-TR studies (type 2 degenerative changes). This is thought to be due to replacement of the early changes with fatty marrow. No enhancement is seen during this stage. Unenhanced MR images are superb for evaluating disc herniation. Gadolinium is occasionally helpful when one wants to improve the delineation of a disc herniation since the epidural veins will enhance and outline disc protrusions. 7 Degenerative disc, herniated disc, and free disc fragments do not generally enhance on immediate postcontrast scans but can show mild J. Neurosurg. / Volume 73/December, 1990


FIG. 12. Magnetic resonance (MR) images of a case of failed-back syndrome. Upper Left: Short-TR/TE axial image 6 months postsurgery disclosing a large low-intensity region on the left at the operation site. Upper Right: Postcontrast short-TR/TE axial image demonstrating enhancement of only the rim (arrows) of the mass. The remainder of the mass appears as an area of low intensity. The patient underwent repeat surgery and a large disc fragment was found. Lower: Short-TR/TE axial MR images obtained 3 months later, when the patient developed recurrent pain. The precontrast image (left) again reveals a low-intensity mass in the same position as previously noted. The postcontrast image (right) demonstrates homogeneous enhancement of this lesion, consistent with scar formation. The nerve root sleeve passes through the center of the scar. enhancement on delayed images. 31 Occasionally, enhancement is seen due to granulation tissue in the disc space. The "failed-back" syndrome is one example of a situation in which contrast-enhanced MR imaging is the diagnostic modality of choice? ~ Although postcontrast CT and noncontrast MR imaging can often detect abnormal tissue within the spinal canal displacing normal fat in the postoperative back, characterizing the abnormality is more difficult; differentiating scar from disc material is reported to be reliable in 67% to 100% of eases. 3t A blinded prospective study by Hueftle, et al., ~ reported 17 patients with failed-back syndrome who had surgical findings correlated with precontrast and postcontrast MR studies. When contrast MR images were compared to precontrast MR studies, specificity increased from 71% to 100%. Other r e p o r t s 7 have confirmed these findings of improved diil~rentiation of scar versus disc herniation (Fig. 12). On immediate postcontrast studies, disc and disc fragments generally


do not enhance, while scar generally enhances markedly. 7"3~Inflamed nerve roots can also enhance, Infection. Gadolinium enhancement may provide supplementary information for spinal osteomyelitis and discitis. 48 Noncontrast MR images show classic findings of spinal infection consisting of loss of disc-space height, destruction of adjacent endplates, and signal abnormalities involving the disc and adjacent vertebral bone marrow. Gadolinium enhances inflamed tissue and granulation tissue surrounding the infection. Epidural inflammatory changes generally show enhancement. Most importantly, Gd can show extension of infection into the epidural space; even when an abscess does not itself enhance on MR imaging, extension of the disease can still be well demonstrated by enhancement of the inflamed overlying epidural veins.

Intmdural Extramedullary Disease Neoplastic Disease. Unlike MR evaluation of extradural disease, Gd is essential for MR studies evalu833


FIG. 13. Spinal imaging in a case of drop metastases from a cerebral glioblastoma. Upper Left: This noncontrast shortTR/TE sagittal magnetic resonance (MR) image is negative except for poor definition of the conus and proximal nerve roots. In retrospect, very vague nodules may be present in the subarachnoid space. Upper Center and Right: These noncontrast long-TR sagittal MR images are also equivocal. There may be a suggestion of high intensity near the conus. Lower Left: Postcontrast short-TR/TE sagittal MR image showing an enhancing subarachnoid tumor encasing the nonenhancing distal spinal cord, causing the total block seen on the myelogram (lower right). In addition, multiple other drop metastases are clearly seen. Lower Right: Myelogram confirming the presence of multiple nodules and total block at the level of the conus. (Reprinted from Sze G, Abramson A, Krol G, et al. AJNR 9:153-163, 1988, with permission.)

834

ating intradural extramedullary spinal disease (Figs. 13 and 14). Gadolinium-DTPA enhancement helps to distinguish intramedullary lesions from extramedullary disease. Lesions detected on noncontrast MR studies, such as schwannomas and meningiomas, are more conspicuous with enhancement. Lesions invisible on noncontrast MR images strongly enhance and are detectable with Gd. Sze, eta/., 68 evaluated 12 patients with intradural extramedullary disease, most of whom had leptomeningeal spread of tumor. Of the 11 patients who had a myelogram, three had normal or equivocal studies. Nine of the 12 patients had normal or equivocal noncontrast MR images using short- as well as longTR pulse sequences. In contradistinction, contrast MR studies using short-TR pulse sequences demonstrated marked contrast enhancement of lesions in nine patients, often depicting nodules which were 2 to 3 m m in size. The one normal postcontrast MR study was in a patient with a normal myelogram who had leukemia confirmed by cerebrospinal fluid (CSF) studies. There are several reasons why small intradural extramedullary tumors and leptomeningeal tumor are poorly seen on noncontrast MR studies. 68 These include the following. 1) The CSF often has elevated levels of protein while leptomeningeal tumor often has a high water content. Both of these factors result in a decrease in the difference between the T~ and T2 relaxation times of the lesions from those of the surrounding CSF. 2) Meningiomas and neural-sheath tumors are often isotense to the cord or CSF on multiple sequences. While large tumors are usually detected, small lesions may be masked by adjacent cord or CSF. 3) Lesions both in the brain and spinal cord are often associated with edema. It is often the edema and not the lesion which is detected on MR imaging. However, there is usually no edema detectable with intradural extramedullary lesions. 4) Artifact due to many types of motion can


Magnetic resonance imaging contrast agents degrade imaging and decrease lesion detection rate in the spine. Movement and CSF pulsation are constant problems. Since Gd markedly enhances intradural extramedullary lesions, even very small nodules, one can understand why postcontrast MR studies represent the diagnostic modality of choice, 15,66,68,76We prefer shortTR images before and after the administration of Gd. Long-TR images may not be necessary, Enhancement of intradural extramedullary disease is best on the immediate postcontrast study.

Inflammatory Disease. Chronic arachnoiditis can lead to adhesions with resulting clumping of nerve roots and distortion of the thecal sac, These changes can be seen on unenhanced MR studies. Although mild or even moderate enhancement of inflamed and matted nerve roots has been seen, poor enhancement in many cases of severe archnoiditis may also be found. 67"7~For most cases, contrast-enhanced MR imaging will not be more advantageous than noncontrast MR studies. Myelography is probably still the most sensitive imaging modality for detecting early arachnoiditis, while noncontrast MR imaging and myelography are equivalent for moderate to severe arachnoiditis. Enhancement can also occur in infectious conditions involving the subarachnoid space. We have noted Gd enhancement of nerve roots in syphilitic and pyogenic bacterial meningitis. Vascular Disease. There has been a limited amount of material published regarding diagnosis of spinal dural arterial venous fistulas on postcontrast MR studies. ~5'23Enhancement of abnormal medullary veins has sometimes been the only finding on MR images in such cases; however, the "gold standard" at the present time is still myelography and spinal angiography. Gadolinium has been shown to enhance regions of ischemia or infarction within the cord which are secondary to spinal arterial venous malformations. ~523 lntramedullary Disease Neoplastic Disease. Magnetic resonance imaging has revolutionized the evaluation of spinal cord tumors. Most reports indicate that Gd improves characterization and delineation of tumors rather than detection.7,8'46,7276However, Dillon, et al.,15 have recently reported four of 13 cases in which detection occurred only after contrast enhancement. Two of the four patients had recurrent tumor. Contrast-enhanced MR imaging is equal to or better than noncontrast MR studies. Gadolinium is most helpful in cases of focal masses, such as hemangioblastomas and metastases, and helps to separate the solid portion of the tumors from the adjacent edema. Cysts are often associated with primary cord tumors. Gadolinium can help to differentiate the solid tumor from the associated cyst (Figs. 15 and 16). Intramedullary tumors almost always enhance with contrast material on MR imaging. Six studies 7,8,~5'46'72,76 describing a total of 75 patients with spinal cord tumors noted contrast enhancement for all tumors except two, an astrocytoma and a cavernous hemangioma. The J, Neurosurg. / Volume 73/December, 1990

FIG. 14. Magnetic resonance imaging of a surgically proven spinal neurofibroma. Upper: Consecutive noncontrast short-TR/TE sagittal images depicting an apparent widening of the distal cord and conus. Center: Postcontrast short-TR/TE sagittal images demonstrating a well-defined enhancing mass, consistent with a neurofibroma. Lower: Postcontrast short-TR/TE axial image confirming the intradural extramedullary location of this lesion.

835


FIG. 15. Magnetic resonance imaging of breast carcinoma metastasis to the spinal cord. Left and Center: These noncontrast short-TR/TE (left) and long-TR/TE (righO sagittal images of the cervical cord cannot depict the exact location of the lesion. Right: Postcontrast short-TR/TE sagittal image demonstrating a focal enhancing lesion consistent with an intramedullary metastasis. The finding of the focal lesion enabled radiation therapy to be directed at the metastasis itself, sparing the remainder of the cord.

majority of lesions were ependymomas and astrocytomas, with a smaller proportion of hemangioblastomas and metastases. There is usually marked enhancement of a well-circumscribed lesion in those cases of hemangioblastomas and metastases. Heterogeneous or variable enhancement can be seen with ependymomas and astrocytomas, although as a general rule ependymomas enhance more homogeneously. 15.46,72It is interesting to note that, while up to one-half of gliomas of the brain do not enhance, almost all gliomas of the cord enhance regardless of grade. In fact, 54 of 55 spinal cord gliomas reported from six studies 7'8'15'46'72'76enhanced with contrast material. One may extrapolate these data to mean that the probability of the presence of a spinal cord glioma is significantly lessened if no enhancement is seen. If a cyst is seen within the cord, Gd is an excellent means of differentiating a benign syrinx from one associated with a tumor. Non-Neoplastic Disease. Enhancement within the cord on M R imaging is not specific for tumor. Infarction, 15 active multiple sclerosis plaques, 36'72 and transverse myelitis 72 can also result in uptake of contrast 836

material within the spinal cord. In a study of five patients with spinal multiple sclerosis, 36 Gd-DTPA enhancement on M R studies paralleled the degree of symptomatology in the two patients with clinically active disease.

Future Applications Future applications for M R contrast agents are developing in two directions: 1) new uses for an accepted agent (Gd-DTPA), and 2) development of new contrast agents. Dynamic scanning 57 and fat-suppression M R imaging 63 are two areas of research where Gd-DTPA may make a major contribution. Preliminary results show enhancement of lesions by Gd-DTPA with gradient echo fast scanning. This may be the future technique for dynamic scanning and decreasing scanning times. 57 These fast scanning techniques are also currently more easily coupled with the use of three-dimensional Fourier transform imaging. With fat-suppression techniques, one does not need to worry about masking contrast enhancement by fat. Potential uses include skull-base, orbital, and spinal extradural evaluation. 63



FIG. 16. Magnetic resonance imaging of a low-grade cervical astrocytoma. Left: Precontrast short-TR/TE sagittal image disclosing extensive enlargement of the cervical cord. Right: The postcontrast short-TR/TE sagittal image, however, shows enhancement of the nidus of the tumor. While tumor undoubtedly extends beyond the area of enhancement, GdDTPA depicts the region of the most active tumor. Low-grade gliomas of the spinal cord enhance much more frequently than low-grade gliomas of the brain.

iron oxide crystals8~ and Gd-labeled liposomes. 62 One could postulate a situation where these agents would be more useful than Gd-DTPA, such as in distinguishing between a neoplastic versus a traumatic vertebral compression fracture. An area of ongoing investigation is the development of target-specific contrast agents. 34,75 The potential for such agents is immense. In addition to improving detection of lesions, one could also characterize the type of neoplastic process and ultimately treat the disease on a microscopic level. Difficulties, however, have arisen with development of tumor-specific antibody contrast agents. 24 Lipophilic contrast agents which can cross cell m e m branes and the blood-brain barrier could be used to evaluate brain function. 24 One substance, Mn mesoporphyrin, has already been used for enhancing experimental gliomas? 2 These substances may never achieve clinical importance because of their toxicity. In addition to intravenous administration, contrast agents can also be injected intrathecally. Early investigations have already found that intrathecal administration of these contrast agents can enhance the CSF? ~ This approach could be used to evaluate arachnoid cysts, epidermoid tumors, intrathecal adhesions, and CSF leaks. Conclusions

In addition to gadopentetate dimeglumine, a number of other contrast agents incorporating Gd are being investigated in clinical trials including Gd-DOTA, 47 non-ionic Gd-DTPA-bis (methylamide) also known as gadolinium-DTPA-BMA, and the non-ionic gadolinium 1,4,7 tris (carboxymethyl)- 10-(2'hydroxypropyl) 1,4,7,10 tetra-azacyclododecane or gadolinum HPDO3A. The non-ionic substances have the theoretical advantage of improved safety; however, the lack of significant side effects of gadopentetate dimeglumine may make this advantage negligible. The contrast agents used in clinical practice for CT and MR imaging are useful because they do not pass through the blood-brain barrier except when it is damaged. There are new contrast agents on the horizon which have different pharmacokinetics. Intravascular agents have been used to explore blood pool and perfusion properties. Gadolinium-DTPA linked to serum albumin or dextran is retained in the intravascular space and results in enhancement of the entire blood pool. 24'6~ Albumin-(Gd-DTPA) has been used to obtain MR angiography. 41 Paramagnetic and superparamagnetic agents 37 can also be used to evaluate the intravascular space. Perfusion agents have the potential for measuring intravascular volume and for improving evaluation of stroke and vascular lesions. 24 Reticuloendothelial system agents may prove to be useful in specific clinical circumstances involving spinal bone marrow. These agents include superparamagnetic J. Neurosurg. / Volume 73 / December, 1990

Gadopentetate dimeglumine is a safe and efficacious paramagnetic contrast agent. In certain conditions, GdDTPA-enhanced M R imaging is either the study of choice or provides additional helpful information. In the head, contrast MR is useful in the evaluation of intracranial metastases, pituitary microadenomas, acoustic neurinomas, meningeal disease, primary brain tumors, and some skull-base lesions. In spinal imaging, Gd injection is indicated routinely in M R diagnosis of the failed-back syndrome or when intradural and intramedullary lesions are suspected. Future potential directions in the utilization of MR contrast agents include dynamic scanning, M R angiography, intrathecal contrast, cerebral blood flow and volume evaluation, and antibody-labeled contrast agents. References

1. Berry I, Brant-Zawadzki M, Osaki L, et al: Gd-DTPA in clinical MR of the brain. 2. Extraaxial lesions and normal structures. AJNR 7:789-793, 1986 2. Bird CR, Drayer BP, Medina M, et al: Gd-DTPA-enhanced MR imaging in pediatric patients after brain tumor resection. Radiology 169:123-126, 1988 3. Bousquet JG, Saint S, Stark DD, et al: Gd-DOTA: characterization of a new paramagnetic complex. Radiology 166:693-698, 1988 4. Brant-Zawadzki M, Berry I, Osaki L, et al: Gd-DTPA in clinical MR of the brain. 1. Intraaxial lesions. AJNR 7: 781-788, 1986 5. Brasch RC, Bennett HF: Considerations in the choice of contrast media for MR imaging. Radiology 166: 897-899, 1988 (Editorial) 837

R. A. Bronen and G. Sze 6. Breger RK, Papke RA, Pojunas KW, et al: Benign extraaxial tumors: contrast enhancement with Gd-DTPA. Radiology 163:427-429, 1987 7. Breger RK, Williams AL, Daniels DL, et al: Contrast enhancement in spinal MR imaging. AJNR 10:633-637, 1989, and AJR 153:387-391, 1989 8. Bydder G, Brown J, Niendorf HP, et al: Enhancement of cervical intraspinal tumors in MR imaging with intravenous gadolinium-DTPA. J Comput Assist Tomogr 9: 847-851, 1985 9. Cohen MD, Klatte EC, Baehner R, et al: Magnetic resonance imaging of bone marrow disease in children. Radiology 151:715-718, 1984 10. Curati WL, Graif M, Kingsley DPE, et al: Acoustic neuromas: Gd-DTPA enhancement in MR imaging. Radiology 158:447-451, 1986 11. Daffner RH, Lupetin AR, Cash N, et al: MRI in the detection of malignant infiltration of bone marrow. AJR 146:353-358, 1986 12. Daniets DL, Czervionke LF, Millen SJ, et al: MR imaging of facial nerve enhancement in Bell's palsy or after temporal bone surgery. Radiology 171:807-809, 1989 13. Daniels DL, Czervionke LF, Pojunas KW, et al: Facial nerve enhancement in MR imaging. AJNR 8:605-607, 1987 14. Davis PC, Hoffman HC Jr, Malko JA, et al: GadoliniumDTPA and MR imaging of pituitary adenoma: a preliminary report. AJNR 8:817-823, 1987 15. Dillon WP, Norman D, Newton TH, et al: Intradural spinal cord lesions: Gd-DTPA-enhanced MR imaging. Radiology 170:229-237, 1989 16. Doppman JL, Frank KA, Dwyer AJ, et al: Gadolinium DTPA enhanced MR imaging of ACTH secreting adenomas of the pituitary gland. J Comput Assist Tomogr 12:728-735, 1988 17. Dwyer AJ, Frank JA, Doppman JL, et al: Pituitary adenomas in patients with Cushing's disease. Initial experience with Gd-DTPA enhanced MR imaging. Radiology 163:421-426, 1987 18. Elster AD, DiPersio DA: Cranial postoperative site: assessment with contrast-enhanced MR imaging. Radiology 174:93-98, 1990 19. Elster AD, Rieser GD: Gd-DTPA-enhanced cranial MR imaging in children: initial clinical experience and recommendations for its use. AJNR 10:1027-1030, 1989 20. Engelstad BL, Wolf GL: Contrast agents, in Stark DD, Bradley WG Jr (eds): Magnetic Resonance Imaging. St Louis: CV Mosby, 1988, pp 161-179 21. Felix R, Schrrner W, Laniado M, et al: Brain tumors: MR imaging with gadolinium-DTPA. Radiology 156: 681-688, 1985 22. Frank JA, Girton M, Dwyer AJ, et al: Meningeal carcinomatosis in the VX2 rabbit tumor model: detection with Gd-DTPA enhanced MR imaging. Radiology 167: 825-829, 1988 23. Gaenster EHL, Dillon WP, Tsuruda J, et at: Vascular malformations of the spinal cord: MR spectrum and pitfalls. AJNR 10:880, 1989 (Abstract) 24. Gibby WA: MR contrast agents. Overview. Radiol Clin North Am 26:1047-1058, 1988 25. Grossman RI, Braffman BH, Brorson JR, et al: Multiple sclerosis: serial study of gadolinium-enhanced MR imaging. Radiology 169:117-122, 1988 26. Grossman RI, Gonzalez-Scarano F, Atlas SW, et al: Multiple sclerosis. Gadolinium enhancement in MR imaging. Radiology 161:721-725, 1986 27. Haughton VM, Rimm AA, Czervionke LF, et al: Sensitivity of Gd-DTPA enhanced imaging of benign extraax838

ial tumors. Radiology 166:829-833, 1988 28. Healy ME, Hesselink JR, Press GA, et al: Increased detection of intracranial metastasis with intravenous GdDTPA. Radiology 165:619-624, 1987 29. Hesselink JR, Healy ME, Press GA, et al: Benefits of GdDTPA for imaging of intracranial abnormalities. J Cornput Assist Tomogr 12:266-274, 1988 30. Hessetink JR, Press GA: MR contrast enhancement of intracranial lesions with Gd-DTPA. Radiol Clio North Am 26:837-887, 1988 31. Hueftle MG, Modic MT, Ross JS, et al: Lumbar spine: postoperative MR imaging with Gd-DTPA, Radiology 167:817-824, 1988 32. Imakita S, Nishimura T, Yamada N, et al: Magnetic resonance imaging of cerebral infarction: time course of Gd-DTPA enhancement and CT comparison. Neororadiology 30:372-378, 1988 33. Kilgore DP, Breger RK, Daniels DL, et al: Cranial tissues: normal MR appearance after intravenous injection of Gd-DTPA. Radiology 160:757-761, 1986 34. Kornguth SE, Turski PA, Perman WH, et al: Magnetic resonance imaging of gadolinium-labeled monoclonal antibody polymers directed at human T lymphocytes implanted in canine brain. J Neurosnrg 66:898-906, 1987 35. Kuharik MA, Edwards MK, Farlow MR, et al: Gd-enhanced MR imaging of acute and chronic experimental demyelinating lesions. AJNR 9:643-648, 1988 36. Larson EM, Holtas S, Nilsson O: Gd-DTPA-enhanced MR of suspected spinal multiple sclerosis. AJNR 10: 107t-1076, 1989 37. Majumbar S, Zoghbi SS, Gore JC: Regional differences in rat brain displayed by fast MRI with superparamagnetic contrast agents. Magn Reson Imaging 6:611-615, 1988 38. Mathews VP, Kuharik MA, Edwards MK, et al: GdDTPA-enhanced MR imaging experimental bacterial meningitis: evaluation and comparison with CT. AJNR 9:1045-1050, 1988, and AJR 152:131-136, 1989 39. McNamara MT, Brant-Zawadzki M, Berry I, et al: Acute experimental cerebral ischemia: MR enhancement using Gd-DTPA. Radiology 158:701-705, 1986 40. Modic MT, Steinberg PM, Ross JS, et al: Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology 166:193-199, 1988 41. Moseley ME, White DL, Lang SC, et al: Vascular mapping using albumin-(Gd-DTPA), intravascular MR contrast agent, and projection MR imaging. J Comput Assist Tomogr 13:215-221, 1989 42. Nelson JA, Cegnar J, Spence AM, et al: Lipophilic manganese porphyrin crosses blood-brain barrier. Radiology 165 (Sappl):303, 1987 (Abstract) 43. Newton DR, Dillon WP, Normal D, et al: Gd-DTPAenhanced MR imaging of pituitary adenomas. AJNR 10: 949-954, 1989 44. Niendorf HP, Laniado M, Semmler W, et al: Dose administration of gadolinium-DTPA in MR imaging of intracranial tumors. AJNR 8:803-815, 1987 45. Okada Y, Aoki S, Barkovich AJ, et al: Cranial bone marrow in children: assessment of normal development with MR imaging. Radiology 171:161-164, 1989 46. Parizel PM, Baleriaux D, Rodesch G, et al: GadoliniumDTPA-enhanced MR imaging of spinal tumors. AJNR 10:249-258, 1989 47. Parizel PM, Degrise HR, Gheuens J, et al: Gadolinium DOTA enhanced MR imaging of intracranial lesions. J Comput Assist Tomogr 13:378-385, 1989 48. Post MJD, Sze G, Quencer RM, et al: Gadoliniumenhanced MR vertebral osteomyelitis and in spinal infec3". Neurosurg. / Volume 73/December, 1990

Magnetic resonance imaging contrast agents tion. J Comput Assist Tomogr (In press, 1990) 49. Powers TA, Partian CL, Kessler RM, et al: Central nervous system lesions in pediatric patients: Gd-DTPA enhancing MR imaging. Radiology 169:723-726, 1988 50. Robinson JD, Crawford SC, Teresi LM, et al: Extracranial lesions of head and neck. Preliminary experience with Gd-DTPA-enhanced MR imaging. Radiology 172: 165-170, 1989 51. Rosen GM, Griffeth LK, Brown MA, et al: Intrathecal administration of nitroxides as potential contrast agents for MR imaging. Radiology 163:239-243, 1987 52. Runge VM, Clanton JA, Herzer WA, et al: lntravascular contrast agents suitable for magnetic resonance imaging. Radiology 153:171-176, 1984 53. Runge VM, Clanton JA, Lukehart CM, et al: Paramagnetic agents for contrast-enhanced NMR imaging: a review. A JR 141:1209-1215, 1983 54. Runge VM, Clanton JA, Price AC, et al: Evaluation of contrast-enhanced MR imaging in a brain-abscess model. AJNR 6:139-147, 1985 55. Runge VM, Jacobson S, Wood ML, et al: MR imaging of rat brain glioma: Gd-DTPA versus Gd-DOTA. Radiology 166:835-838, 1988 56. Runge VM, Schaible TF, Goldstein HA, et al: Gd-DTPA clinical efficacy. Radiographics 8:147-159, 1988 57. Runge VM, Wood ML, Kaufman D, et al: Gd-DTPA. Future applications with advanced imaging techniques. Radiographics 8:161-179, 1988 58. Russell EJ, Geremia GK, Johnson CE, et al: Multiple cerebral metastases: detectability with Gd-DTPA-enhanced MR imaging. Radiology 165:609-617, 1987 59. Russell EJ, Schaible TF, Dillon W, et al: Multicenter double-blind placebo-controlled study of gadopentetate dimeglumine as an MR contrast agent: evaluation in patients with cerebral lesions. A JR 152:813-823, 1989 60. Saini S, Frankel RB, Stark DD, et al: Magnetism: a primer and review. A JR 150:735-743, 1988 61. Schmidl U, Ogan M, Paajanen H, et al: Albumin labeled with Gd-DTPA as an intravascular, blood pool enhancing agent for MR imaging. Biodistribution and imaging studies. Radiology 162:205-210, 1987 62. Seltzer SE: The rote of liposomes in diagnostic imaging.

Radiology 171:19-21, 1989

63. Simon JH, Szumowski J: Chemical shift imaging with paramagnetic contrast material enhancement for improved lesion depiction. Radiology 171:539-543, 1989 64. Stack JP, Antoun NM, Jenkins JPR, et al: GadoliniumDTPA as a contrast agent in magnetic resonance imaging of the brain. Neuroradiology 30:145-154, 1988 65. Stack JP, Ramsden RT, Antoun NM, et al: Magnetic resonance imaging of acoustic neuromas: the role of gadolinium-DTPA. Br J Radiol 61:800-805, 1988 66. Stimac GK, Porter BA, Olson DO, et al: GadoliniumDTPA enhanced MR imaging of spinal neoplasms: preliminary investigation and comparison with unenhanced spin-echo and STIR sequences. AJNR 9:839-846, 1988 and AJR 151:1185-1192, 1988 67. Sze G: Gadolinium-DTPA in spinal disease. Radiol Clin North Am 26:1009-1024, 1988

J. Neurosurg. / Volume 73 / December, 1990

68. Sze G, Abramson A, Krol G, et al: Gadolinium-DTPA/ dimeglumine in the MR evaluation of intradural extramedullary spinal disease. AJNR 9:153-163, 1988, and AJR 150:911-921, 1988 69. Sze G, Bravo S, Crow G: Post-operative enhancement: gadolinium-DTPA in a clinical series and an animal model. Radiology 169 (Suppl):70, 1988 (Abstract) 70. Sze G, Johnson C: Benign lumbar arachnoiditis: contrastenhanced MR imaging. AJNR 11:763-770, 1990 71. Sze G, Krol G, Zimmerman RD, et al: GadoliniumDPTA: malignant extradural spinal tumors. Radiology 167:217-233, 1988 72. Sze G, Krol G, Zimmerman RD, et al: Intramedullary disease of the spine: diagnosis using gadolinium-DTPA enhanced MR imaging. AJNR 9:847-858, 1988, and AJR 151:1193-1204, 1988 73. Sze G, Milano E, Johnson C, et al: Detection of brain metastases: comparison of contrast-enhanced MR with unenhanced MR and enhanced CT. AJNR 11:785-791, 1990 74. Sze G, Soletsky S, Bronen R, et al: MR imaging of the cranial meninges, with emphasis on contrast enhancement and meningeal carcinomatosis. AJNR 10: 965-975, 1989 75. Unger EC, ToRy WG, Neufeld DM, et al: Magnetic resonance imaging using gadolinium labeled monoclonal antibody. Invest Radiol 20:693-700, 1985 76. Valk J: Gadolinium-DTPA in MR of spinal lesions. AJNR 9:345-350, 1988 77. Virapongse C, Mancuso A, Quisling R: Human brain infarcts Gd-DTPA enhanced MR imaging. Radiology 161:785-794, 1986 78. Vogl T, Bruning R, Schedel H, et al: Paragangliomas of the jugular bulb and carotid body: MR imaging and short sequences Gd-DTPA enhancement. AJNR 10:823-827, 1989 79. Weinmann H J, Brasch RC, Press WR, et al: Characteristics of gadolinium-DTPA complex: a potential NMR contrast agent. A JR 142:619-642, 1984 80. Weissleder R, Stark DD, Engelstad BL, et al: Superparamagnetic iron oxide: pharmacokinetics and toxicity. A JR 152:167-173, 1989 81. West M, Russell EJ, Breit R, et al: Calvarial and skull base metastases: comparison of non-enhanced and GdDTPA-enhanced MR images. Radiology 174:85-91, 1990 82. Williams AL, Czervionke LF, Haughton VM, et al: Quantitative assessment of spinal tissue enhancement with intravenous gadolinium in MR imaging. Radiology 165 (Suppl):78, 1987 (Abstract)

Manuscript received October 31, 1989. Accepted in final form April 19, 1990. Address reprint requests to." Richard A. Bronen, M.D., Department of Diagnostic Radiology, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06510.

839

Non-central nervous system fetal magnetic resonance imaging.

Magnetic resonance imaging observations in primary central nervous system lymphoma.

Application value of magnetic resonance imaging in diagnosing central nervous system lymphoma.

Magnetic resonance contrast agents for liver imaging.

Magnetoliposomes as magnetic resonance imaging contrast agents.

Contrast agents for nuclear magnetic resonance imaging.

Contrast agents in dynamic contrast-enhanced magnetic resonance imaging.

Magnetic resonance imaging in viral and prion diseases of the central nervous system.

Prediction of radiosensitivity in primary central nervous system germ cell tumors using dynamic contrast-enhanced magnetic resonance imaging.

Advances in Magnetic Resonance Imaging Contrast Agents for Biomarker Detection.

[Gadolinium-based contrast agents for magnetic resonance imaging].

Blood pool contrast agents for venous magnetic resonance imaging.

Iron Oxide Nanoparticle Based Contrast Agents for Magnetic Resonance Imaging.

Role of magnetic resonance imaging in the diagnosis of primary central nervous system angiitis.

New calcium-selective smart contrast agents for magnetic resonance imaging.

Paramagnetic oil emulsions as oral magnetic resonance imaging contrast agents.

Gadolinium-enhanced magnetic resonance imaging of the central nervous system in systemic lupus erythematosus.

Central nervous system abnormalities in Fanconi anaemia: patterns and frequency on magnetic resonance imaging.

Fetal Central Nervous System Anomalies Detected by Magnetic Resonance Imaging: A Two-Year Experience.

Magnetic Resonance Imaging Characteristics of Primary Central Nervous System T-cell Lymphoma.

Dynamic contrast-enhanced magnetic resonance imaging: fundamentals and application to the evaluation of the peripheral perfusion.

Correction to "Safety and Efficacy of Gadobutrol for Contrast-enhanced Magnetic Resonance Imaging of the Central Nervous System: Results from a Multicenter, Double-blind, Randomized, Comparator Study".

Toward the Prediction of Water Exchange Rates in Magnetic Resonance Imaging Contrast Agents: A Density Functional Theory Study.

Structural and magnetic characterization of superparamagnetic iron platinum nanoparticle contrast agents for magnetic resonance imaging.