Mechanisms of contrast enhancement in magnetic resonance imaging

The use of contrast agents has increased the sensitivity and specificity of magnetic resonance imaging (MRI). Contrast in MRI is multifactorial, depending not only on T1 and T2 relaxation rates, but also on flow, proton density and, in gradient-echo sequences, on the angle of the induced field. The use of contrast agents in MRI changes the T1 and T2 relaxation rates, producing increased signal intensity on T1-weighted images or decreased signal intensity on T2-weighted images, or both. All contrast agents produce changes in magnetic susceptibility by enhancing local magnetic fields. These effects are caused by interactions between nuclear and paramagnetic substance magnet moments, which produce accentuated transitions between spin states and cause shortening of T1; the paramagnetic substance causes accentuated local fields, which lead to increased dephasing and thus shortening of T2 or T2* relaxation time. The efficacy of shortening of T1, T2 or T2* relaxation time depends on the distance between the proton nucleus and the electronic field of the paramagnetic compound, the time of their interaction (correlation time) and the paramagnetic concentration. The MRI contrast agents currently in use cause shortening of T1, T2 or T2* relaxation time. Metal chelates (e.g., gadolinium-diethylene triamine penta-acetic acid [Gd-DTPA]) in low concentration cause shortening of T1 relaxation times, and the superparamagnetics (e.g., ferrite) cause shortening of T2 relaxation times.     

Proton relaxation enhancement associated with iodinated contrast agents in MR imaging of the CNS.

PURPOSE: To study the effects of iodinated radiographic contrast agents on proton relaxation in MR imaging. PATIENTS AND METHODS: Two patients were evaluated after the intrathecal administration of an iodinated nonionic contrast agent (Isovue) and five subjects with cranial tumors following the intravenous administration of an iodinated ionic contrast medium (Renografin). RESULTS: Both patients with subarachnoid iodinated contrast media demonstrated a relative reduction in T1 and/or T2 times using a spin-echo sequence, while four of five of the subjects with intracranial tumors (one glioma, one dural metastasis, three meningiomas) and intravenous enhancement revealed a visible MR effect. Confirmation of these in vivo observations was obtained by in vitro measurement of T1 and T2 while varying the concentration of the contrast media in saline. All iodinated contrast media showed progressively reduced relaxation times (T1 and T2) as the concentration of the agent was increased. The largest contributing relaxation mechanism is probably due to the binding and exchange of the surrounding water with the contrast molecules. CONCLUSION: The observed T1/T2 effects suggest that administration of iodinated contrast media in the period immediately prior to MR scanning may be contraindicated in selected cases due to the demonstrated alteration of MR signal intensity that may lead to diagnostic inaccuracies.     

The tissue proton T1 and T2 response to gadolinium DTPA injection in rabbits. A potential renal contrast agent for NMR imaging

Three different doses of gadolinium (Gd) DTPA were administered intravenously to rabbits. Cardiovascular responses and changes in blood T1 and T2 were serially followed for 15 minutes when the animals were sacrificed for in vitro measures of tissue T1 and T2. Gd-DTPA was distributed and excreted like water soluble iodinated contrast agents with large changes in blood, urine, and kidney proton relaxation. An imaging experiment confirmed the efficacy as an NMR contrast agent for renal excretion. At effective doses, no adverse effects were observed and the agent appeared to be much safer than x-ray contrast agents, but with a similar potential for clinical utility.     

T1 rho dispersion imaging and volume-selective T1 rho dispersion weighted NMR spectroscopy

Pulse sequences which permit imaging and volume-selective determination of parameters characterizing the frequency dependence of the spin-lattice relaxation time in the rotating frame, T1 rho, are presented. The contrasts are due to slowly moving macromolecules or paramagnetic contrast agents. In vivo test experiments were carried out with tumorous mice treated with a contrast agent. It is shown that the contrast effect is dramatically enhanced in T1 rho dispersion images compared with images weighted by any of the relaxation times.     

Cooperative T1 and T2 effects on contrast using a new driven inversion spin-echo (DISE) MRI pulse sequence

A pulse sequence is presented for obtaining a single image with combined T1/T2 weighting. T2 relaxation is made to increase intensity, in cooperation with the effect of T1 relaxation, by providing T2 weighting with a 90 degrees-180 degrees-90 degrees driven inversion pulse triplet in an inversion recovery method. Unlike the inversion spin-echo method having a short inversion time (TI), signals in the new driven inversion spin-echo (DISE) method need not be negative and the most T1-sensitive region of the recovery curve can be used. Selecting sensitivity to one relaxation time does not degrade the sensitivity to the other relaxation time. T1 sensitivity is thus extended to longer echo times (TE intervals). T2 sensitivity is extended to longer TI intervals, and the combined T1/T2-weighted technique with intermediate TE and TI has highly cooperative and near-maximal T1 and T2 effects on contrast. Intensity is not multiplicatively degraded by T1 and T2 weighting so that the signal-to-noise of the combined T1/T2-weighted method is high. High intensity and T1 and T2 cooperatively occur for a much wider range of relaxation times, and especially for images heavily weighted to the pathologic intermediate and long T1 and T2 regime.     

Quantitative T(1rho) and adiabatic Carr-Purcell T2 magnetic resonance imaging of human occipital lobe at 4 T.

The feasibility of performing quantitative T(1rho) MRI in human brain at 4 T is shown. T(1rho) values obtained from five volunteers were compared with T2 and adiabatic Carr-Purcell (CP) T2 values. Measured relaxation time constants increased in order from T2, CP-T2, T(1rho) both in white and gray matter, demonstrating differential sensitivities of these methods to dipolar interactions and/or proton exchange and diffusion in local microscopic field gradients, which are so-called dynamic averaging (DA) processes. In occipital lobe, all relaxation time constants were found to be higher in white matter than in gray matter, demonstrating contrast denoted as an "inverse transverse relaxation contrast." This contrast persisted despite changing the delay between refocusing pulses or changing the magnitude of the spin-lock field strength, which suggests that it does not originate from DA, as might be induced by the presence of Fe, but rather is related to dipolar interactions in the brain tissue.     

B0 dependence of the on-resonance longitudinal relaxation time in the rotating frame (T1rho) in protein phantoms and rat brain in vivo.

On-resonance longitudinal relaxation time in the rotating frame (T1rho) has been shown to provide unique information during the early minutes of acute stroke. In the present study, the contributions of the different relaxation mechanisms to on-resonance T1rho relaxation were assessed by determining relaxation rates (R1rho) in both protein phantoms and in rat brain at 2.35, 4.7, and 9.4 T. Similar to transverse relaxation rate (R2), R1rho increased substantially with increasing magnetic field strength (B0). The B0 dependence was more pronounced at weak spin-lock fields. In contrast to R1rho, longitudinal relaxation rate (R1) decreased as a function of increasing B0 field. The present data argue that dipole-dipole interaction forms only one pathway for T1rho relaxation and the contributions from other physicochemical factors need to be considered.     

Studies of tissue NMR relaxation enhancement by manganese. Dose and time dependences

Manganese is a powerful paramagnetic material and potential NMR contrast agent. It drastically affects the NMR properties of solutions and tissues and is less toxic than most other transition elements. It also possesses some unusual and advantageous features; it alters T1 and T2 to different degrees, and it can bind to macromolecules to become even more effective at reducing proton relaxation times. The dose dependence of tissue relaxation rate increases has been measured in mice, and proton relaxation enhancement ratios that describe binding effects have been evaluated. These ratios imply that a tenfold reduction in manganese dose is achievable when the ion binds to intracellular components, and it is demonstrated that such binding effects can be a major factor in the efficacy of contrast enhancement. The effect of manganese on the ratio T1/T2 is dose dependent so that lower doses may be more useful for some imaging techniques. The postmortem time course of relaxation times in organs containing manganese varies between organs and with manganese content, and demonstrates that the relationship between tissue relaxation enhancement and metal content is not a simple correlation with concentration since large variations in T1 and T2 can occur even when metal and water content are fixed.     

A method for T1 rho imaging

The spin lattice relaxation time (T1) is dependent on the strength of the polarizing magnetic field. The relaxation at low field strengths provides information from the processes at macromolecular level. However, the decrease of the polarizing magnetic field decreases the signal-to-noise ratio that determines the resolution of magnetic resonance images. In this report we describe a method for T1 rho imaging. The method possesses the relaxation time contrast of low field strengths with signal-to-noise ratio provided by the higher polarizing field. The relaxation time T1 rho is obtained under spin lock conditions. The spin system relaxes toward thermal equilibrium along the locking field. This process is analogous to the spin lattice relaxation at low field strength and characterized by the time constant T1 rho. T1 rho and T1 rho-dispersion may provide new imaging parameters for noninvasive tissue characterization.     

Contrast and accuracy of relaxation time measurements in acquired and synthesized multislice magnetic resonance images

The effects of interslice spacing, the number of data points and other factors on the accuracy of relaxation time measurements and contrast have been investigated for both acquired and synthesized multislice MR images using experiments and computer simulations.The cross-excitation between adjacent slices in multislice imaging affects both contrast and derived relaxation times. Such measurements also are affected by the T1 and T2 of the materials imaged, the pulse sequence timing parameters, and the number of data points used to estimate the relaxation times. Errors in T1 and T2 may be severe, particularly for slice spacings less than 0.5 slice thickness and for long T1 and T2 materials. Consequently, the difference in signal intensities between two materials with different relaxation times also varies with slice spacing and between acquired and synthetic images, particularly for strongly T1-weighed images.     

Rapid T1 estimation using tagged magnetization-prepared gradient-echo MR imaging.

A technique for estimation of the longitudinal relaxation time of a large homogeneous object with an acquisition time of 4 s or less was developed by combining spatially selective rf tagging pulses with a T1-weighted magnetization-prepared gradient-echo sequence. Multiple 5-mm-wide tagged areas are laid orthogonal to the imaging section of interest. The contrast between each tag and the untagged regions differs because each tag is produced at a different time. The T1 value is determined from the nulling time at which tagged and untagged areas have no contrast.     

Exchange-influenced T2rho contrast in human brain images measured with adiabatic radio frequency pulses.

Transverse relaxation in the rotating frame (T(2rho)) is the dominant relaxation mechanism during an adiabatic Carr-Purcell (CP) spin-echo pulse sequence when no delays are used between pulses in the CP train. The exchange-induced and dipolar interaction contributions (T(2rho,ex) and T(2rho,dd)) depend on the modulation functions of the adiabatic pulses used. In this work adiabatic pulses having different modulation functions were utilized to generate T(2rho) contrast in images of the human occipital lobe at magnetic field of 4 T. T(2rho) time constants were measured using an adiabatic CP pulse sequence followed by an imaging readout. For these measurements, adiabatic full passage pulses of the hyperbolic secant HSn (n = 1 or 4) family having significantly different amplitude-and frequency-modulation functions were used with no time delays between pulses. A dynamic averaging (DA) mechanism (e.g., chemical exchange and diffusion in the locally different magnetic susceptibilities) alone was insufficient to fully describe differences in brain tissue water proton T(2rho) time constants. Measurements of the apparent relaxation time constants (T(2) (dagger)) of brain tissue water as a function of the time between centers of pulses (tau(cp)) at 4 and 7 T permitted separation of the DA contribution from that of dipolar relaxation. The methods presented assess T(2rho) relaxation influenced by DA in tissue and provide a means to generate T(2rho) contrast in MRI.     

Deuterium NMR cerebral imaging in situ

Nuclear magnetic resonance imaging of deuterium is demonstrated in cat brain in vivo and in situ at 4.7 T magnetic field strength. Images were acquired at 4-5% deuterium enrichment, using D2O as a nuclear spin label, with as little as 10-s time resolution. This suggests the potential application of D2O as an exogenous MRI label for quantitative flow imaging or contrast enhancement. The efficient quadrupolar relaxation mechanism of the deuterium nuclide results in a short ca. 250 ms spin-lattice relaxation time (T1). This allows for rapid signal averaging, thus increasing signal-to-noise in the deuterium image. Additionally, three widely dispersed deuterium spin-spin relaxation times (T2) are observed of ca. 10, 40, and 400 ms which provide high compartmental image contrast, thus yielding information of physiological as well as anatomical interest. T2-weighted deuterium cerebral images are presented showing marked tissue relaxation discrimination.     

Contrast mechanisms in MR imaging

This paper is a brief introduction to tissue-specific parameters and the utilization of various MR imaging sequences to display these parameters in order to differentiate normal from pathologic tissue and function. The three dominant tissue-specific parameters discussed are proton density, longitudinal relaxation time T1, and transverse relaxation time T2. For the utilization of gradient-echo sequences, transverse relaxation time T2^* is introduced, more dependent on the environment or tissue interfaces than on the tissue itself. Another tissue-specific parameter is the concentration of macromolecules and their hydration layers as targeted with magnetization transfer imaging. Still another tissue-specific parameter is the chemical environment. Functional parameters that influence the contrast are diffusion, perfusion, flow, or motion. The sequence-related utilization of these tissue-specific parameters start with magnetization preparation as in spectral suppression of fat signal, relaxation-dependent elimination of fat or cerebrospinal fluid (CSF) signal, simple inversion pulse, magnetization transfer saturation, or diffusion weighting. Possible contrast mechanisms for the tissue-specific parameters are discussed for each of the commonly used sequences, whether of spin-echo type or of gradient-echo type, with or without magnetization preparation, conventional single-echo acquisition, or contemporary multiecho acquisition.     

Proton magnetic resonance relaxation behavior of whole muscle with fatty inclusions

This in vitro proton study of spin lattice (T1) and spin spin (T2) relaxation of muscle with storage-fat inclusions demonstrates slow exchange and lack of cross-relaxation between fat and water. Slow exchange causes biphasic T1 relaxation, but T2 relaxation is paradoxically uniphasic due to the nearly equal T2 values for both fractions. By careful dehydration and fat extraction, the relaxation information was deconvolved into water, fat, and protein contributions. The biphasic T1 decay has a short component due to lipid and a long component due to the water-protein combination. The fat content of muscle can be measured from the relative amplitude of the two T1 components or directly from the T2 relaxation time.     

Magnetic resonance imaging of the bone marrow in patients with acute leukemia during and after chemotherapy. Changes in T1 relaxation

Twenty-seven patients with acute leukemia were examined at the time of diagnosis with MR imaging and in vivo T1 relaxation time measurements of the hemopoietic bone marrow. A 1.5 T whole body magnetic resonance scanner was used. Twenty of the patients had follow-up examinations in relation to chemotherapy. Bone marrow biopsies from the posterior iliac crest were obtained within a short time interval of all MR examinations. At the time of diagnosis, T1 relaxation times were increased significantly in all the leukemic patients, compared with 24 age-matched controls. A decrease in T1 relaxation time towards or into the normal range was observed in 10 patients who obtained remission. The T1 relaxation time remained prolonged in 6 patients who failed to obtain remission during chemotherapy. Four patients, who obtained remission with concomitant decrease of T1 values towards or into the normal range, also showed prolongation of T1 relaxation time in relation to leukemic relapse. The results indicate that changes observed in T1 relaxation times of the hemopoietic bone marrow in patients with acute leukemia reflect changes in disease activity, and, that serial measurements of T1 values may provide clinically useful information with the possibility for identification of residual disease in regions inaccessible for biopsy.     

Aerosolized gadolinium-DTPA enhances the magnetic resonance signal of extravascular lung water

Gadolinium diethylenetriamine pentaacetic acid (Gd-DTPA), a contrast agent for magnetic resonance imaging, was administered by either aerosol or intravenous injection to rats. Proton relaxation times in excised lungs and kidneys were then measured. With increasing concentrations of aerosolized Gd-DTPA, the spin-lattice relaxation time (T1) of lungs decreased (enhanced) significantly (P less than .001), an effect that persisted for at least 80 minutes; there was no change in kidney T1. After intravenous injection of Gd-DTPA, lung T1 did not change, but kidney T1 decreased significantly (P less than .001), confirming previous observations of renal clearance. It is concluded that aerosolized Gd-DTPA is a more efficacious method of delivery of paramagnetic contrast agent to the lungs than intravenous injection, and that the lack of systemic effect after aerosolization indicates that enhancement was limited to the extravascular compartment.     

MR imaging using stimulated echoes (STEAM)

The introduction of STEAM (stimulated echo acquisition mode) magnetic resonance (MR) sequences provides access to a variety of MR parameters. T1-weighted and calculated T1 proton MR images of the head of healthy volunteers and a patient with an astrocytoma are presented. MR examinations were performed with a 2.0-T whole-body system. The STEAM T1 method can be used to characterize multiexponential relaxation behavior, to evaluate T1 relaxation times, and to improve the T1 contrast within MR images. Both the measuring time and the spatial resolution are the same as for a conventional image.     

High-speed multislice T1 mapping using inversion-recovery echo-planar imaging

Tissue contrast in MR images is a strong function of spin-lattice (T1) and spin-spin (T2) relaxation times. However, the T1 relaxation time is rarely quantified because of the long scan time required to produce an accurate T1 map of the subject. In a standard 2D FT technique, this procedure may take up to 30 min. Modifications of the echo-planar imaging (EPI) technique which incorporate the principle of inversion recovery (IR) enable multislice T1 maps to be produced in total scan times varying from a few seconds up to a minute. Using IR-EPI, rapid quantification of T1 values may thus lead to better discrimination between tissue types in an acceptable scan time.     

Magnetic resonance imaging of the gastrointestinal tract: investigation of baby milk as a low cost contrast medium.

After the incidental observation of the high signal intensity of the upper GI tract in a nourished baby, we tested eight baby milks; five different fresh commercial milks, one sweetened and condensed and two lyophilized milks in order to compare their ability to contrast MR images. The images were obtained with a 1.5 T magnet whereas the "in vitro" water proton relaxation time (T1 and T2) measurements were carried out at 0.5 T. After having selected the most effective lyophilized product, that was prepared according to the manufacturer's instructions, a group of 23 adult patients, 17 males and 6 females, with a mean age of 55.8 years (range 37 to 71 years) were examined. Thirteen patients had gastric cancer and ten patients had rectal or rectosigmoid junction tumors. The most effective imaging sequence was a spin-echo T1.w. After oral intake of milk a good contrast of the stomach, with sufficient distribution in the duodenum and the very proximal bowel, was achieved in all 13 patients with gastric cancer, as was a good depiction of the rectum and the recto-sigmoid junction after enema achieved in the 10 patients with rectal cancers. Disadvantages of lyophilized milk as a contrast agent are due to partial intestinal absorption, inhomogeneous distribution and irregular intestinal passage, whereas a clear advantage of lyophilized milk as a contrast agent is its good acceptance and palatable, inexpensive and non invasive properties. Because of these limitations lyophilized milk cannot be considered a real oral contrast medium but it can enhance MR imaging of the upper abdomen, and mainly of the lower GI tract in infants and adults.