Magnetic resonance imaging in pediatric hematology/oncology. Part I. Basic technology.

Magnetic resonance imaging (MRI) is a promising new diagnostic modality that is well suited for the evaluation of children with hematological or oncological diagnosis. The side effects of ionizing radiation are avoided, the tomographic pathological anatomy in three orthogonal planes can be obtained, and differences between normal and abnormal tissues are often present. In order to present our preliminary clinical experience with MRI in pediatric hematology and oncology, the historical background of MRI, the technique, and possibilities for tissue characterization are reviewed.

Magnetic resonance imaging in pediatric hematology/oncology. Part II. Illustrative cases and assessment of technique.

Preliminary clinical experience with magnetic resonance imaging (MRI) in 28 pediatric patients with 20 different hematological diseases, benign tumors, or malignant neoplasms is presented. The clinical results are presented in the form of case presentations that are discussed in the context of alternative diagnostic imaging modalities. Also discussed are the known biological effects, or lack thereof, the need for sedation, the effect of motion, the effect of MRI on foreign metallic objects, the financial considerations, and the trends for the future of MRI. This imaging modality has many unique merits. Present difficulties should be overcome by future innovations, making MRI even more efficacious for the diagnosis of blood diseases and cancer in children.

Assessment of severity of experimental pulmonary edema with magnetic resonance imaging. Effect of relaxation enhancement by Gd-DTPA.

This study was done to demonstrate the capability of magnetic resonance imaging (MRI), using a spin-echo technique to detect permeability pulmonary edema in vivo, to correlate the signal intensity to the water content of the lung, and to assess the influence of administration of gadolinium-DTPA upon this relationship. Pulmonary edema was induced in 28 rats by intravenous injection of oleic acid (0.05-0.1 cc/kg). This was detected in all animals on images obtained with a repetition time (TR) of 2.0 seconds and echo times (TE) of 28 or 56 msec as increased signal intensity. Compared with control animals, the intensity was increased primarily in the dependent and peripheral portions of the lung. There was a significant linear relationship between the mean signal intensity, obtained from the area of both lungs within one slice, and the water content of both lungs (r = .94). Intravenous administration of gadolinium (Gd)-DTPA, 5 minutes prior to imaging, produced an easily observable increase in signal intensity on images of short TR (0.5 second) in the edematous lung tissue. However, these values did not exceed the intensities obtained without Gd-DTPA, if a long TR (2.0 seconds) was used for imaging. Thus, MRI has the capability for quantitating the severity of oleic acid-induced pulmonary edema. Gd-DTPA distributes rapidly into permeability pulmonary edema. It allows improved sensitivity at shorter pulse sequence repetition times and thus may diminish imaging time.

Measurement of regional myocardial blood flow in dogs by ultrafast CT.

We measured blood flow within each of eight segments of the left ventricular myocardium in dogs by an Ultrafast CT scanner. The results were compared with flow determined by radiolabeled microspheres. Computed tomography (CT) flow was measured by an intravenous injection of nonionic contrast agent done simultaneously with the left atrial injection of microspheres. We calculated flow from the CT data by obtaining CT number versus time curves for regions of interest in the myocardium and by using a formula that related flow to both the time and value of the peak enhancement. Measurements were obtained in five dogs at rest and during hyperperfusion induced by chromonar. Based on 169 regional measurements, the Ultrafast CT and microsphere-determined flows correlated moderately (r = 0.68) over a range of 0.4 to 8 mL/min/g. However, when the data were divided into resting and hyperperfusion (ie, 20 to 30 minutes after the injection of the chromonar) states, a significant (P less than .001) increase in regional flow was determined from the CT measurements. The conclusion was that Ultrafast CT can distinguish between low and high myocardial flow states in dogs and has considerable potential for evaluating coronary flow reserve.

Diradical nitroxyl spin label contrast agents for magnetic resonance imaging. A comparison of relaxation effectiveness.

The proton relaxation enhancement characteristics of seven potential MRI contrast agents containing two nitroxyl spin labels per molecule (diradicals) were compared with eight similar agents with only one spin label per molecule (monoradicals). Diradical nitroxyls were evaluated to test the hypothesis that multiple paramagnetic centers in one molecule will result in stronger proton relaxation enhancement characteristics, allowing effective contrast enhancement at lower molar concentrations and thus a reduced osmotic load and greater safety. The acute toxicity of these agents is believed to be largely related to osmotic load. Five of seven diradical nitroxyls tested had spin-lattice relaxivities that were substantially greater than all eight of the monoradicals tested. The spin-spin relaxation properties of these agents and other pertinent characteristics are favorable for contrast enhancement. The results indicate that diradical nitroxyl spin labels may be used advantageously for the design of safer, more effective MRI contrast agents.

Hepatobiliary magnetic resonance contrast agents assessed by gadolinium-153 scintigraphy.

A simple method to test new gadolinium complexes potentially useful as enhancement agents for magnetic resonance imaging was developed. Healthy rats underwent scintigraphy with two potential hepatobiliary agents, diethyl IDA and diisopropyl IDA complexed with gadolinium-153. Control products included 153Gd DTPA, 153GdCl3 and technetium-99m diethyl IDA. As shown scintigraphically, 153Gd IDA complexes were partially excreted by urinary and hepatobiliary excretion early after administration. These findings paralleled significant reduction in 1H T1 values of excised livers. However, these agents exhibited prolonged 153Gd whole-body retention. The prolonged tissue distribution of 153Gd activity in animals given 153Gd diethyl IDA did not differ significantly from that observed in animals given GdCl3, and could be attributed to chemical instability or reticuloendothelial uptake. The scintigraphic method permits screening of gadolinium complexes in animals by showing mass balance, kinetics, distribution, and effective stability. Biologic effects of tracer or pharmacologic levels can be compared with those of carrier-free and carrier-added pharmaceuticals.

Magnetic resonance imaging of acute myocardial infarction using a nitroxyl spin label (PCA).

The effects of an intravenously administered nitroxyl spin label (PCA) on the magnetic resonance imaging (MRI) appearance and relaxation times of acute canine myocardial infarctions were studied. Twenty-four hours after ligation of the left anterior descending coronary artery (LAD), animals were either sacrificed immediately (three dogs) or injected with 3.0 mmol/kg of PCA prior to sacrifice (six dogs). The PCA group dogs were sacrificed at either 5 minutes postinjection (three dogs) or 15 minutes postinjection (three dogs). Magnetic resonance imaging (0.35 T) using spin-echo techniques demonstrated high signal intensity in the infarct relative to normal myocardium in all three groups. In the control group, the T1 and T2 relaxation times were longer in infarcted compared with normal myocardium, but only the measure in T2 reached statistical significance (P less than .05). PCA produced infarct-avid T1 shortening in the six dogs that received it. Contrast in the group sacrificed at 15 minutes postcontrast administration was greater than that in the control group due to T1 shortening in the infarct. Thus, PCA produces differential effects on normal and infarcted myocardium. Between 5 and 15 minutes after IV administration, it causes greater changes in the infarct due to prolonged retention in this region.

Use of the magnetic resonance contrast agent gadodiamide in the central nervous system. Results of a multicenter trial.

To investigate the safety and efficacy of the low-osmolar, nonionic contrast agent, gadodiamide injection (Omniscan, Sanofi Winthrop Pharmaceuticals, New York, NY), for magnetic resonance imaging (MRI) of the head and spine, a multicenter study involving 439 patients was done at 15 centers as part of a Phase II/III clinical trial. Unenhanced MRI scans were obtained after which the patients were injected with 0.1 mmol/kg gadodiamide, and the MRI was repeated. The patients' vital signs were monitored, and laboratory studies were conducted. Neurologic status was examined before and after the study. The images were evaluated for contrast enhancement. No patient had any significant adverse event or serious change in clinical status. Abnormalities were found in 80% (351) of all patients studied, and it was found that, in 75% (266) of these, the postgadodiamide injection images were improved or facilitated visualization of lesions compared with preinjection images. The investigators believe that, based on the results of this study, gadodiamide injection is safe and effective for imaging the head and spine. They suggest that future studies further assess and compare the safety parameters of gadodiamide injection with those of other nonionic and ionic gadolinium ligands.

Cardiovascular applications of magnetic resonance imaging.

Magnetic resonance imaging (MRI) is a completely noninvasive modality that has shown significant promise for the evaluation of the cardiovascular system. Our imaging technique employed electrocardiographic (ECG) gating, which resulted in well-resolved images of the cardiac structures. Patients and animals with a variety of cardiovascular abnormalities were also assessed with this technique; the abnormalities included acute and remote myocardial infarctions and their sequelae, atherosclerotic plaques, hypertrophic cardiomyopathy, pericardial diseases, and aneurysms. The diagnostic utility of MRI includes direct tissue characterization, and such utility may be further extended by the use of paramagnetic contrast media. In addition, metabolic imaging of elements other than hydrogen may further increase the clinical potential of MRI for assessment of the cardiovascular system.

Enhanced MRI of tumors utilizing a new nitroxyl spin label contrast agent.

Nitroxyl spin labels have been shown to be effective in vivo contrast agents for magnetic resonance imaging (MRI) of the central nervous system, myocardium, and urinary tract. A new pyrrolidine nitroxyl contrast agent (PCA) with better resistance to in vivo metabolic inactivation than previously tested agents was studied for its potential to enhance subcutaneous neoplasms in an animal model. Twenty-two contrast enhancement trials were performed on a total of 15 animals 4-6 weeks after implantation with human renal adenocarcinoma. Spin echo imaging was performed using a .35 T animal imager before and after intravenous administration of PCA in doses ranging from 0.5 to 3mM/kg. The intensity of tumor tissue in the images increased an average of 35% in animals receiving a dose of 3 mM/kg. The average enhancement with smaller doses was proportionately less. Tumor intensity reached a maximum within 15 min of injection. The average intensity difference between tumor and adjacent skeletal muscle more than doubled following administration of 3 mM/kg of PCA. Well-perfused tumor tissue was more intensely enhanced than adjacent poorly perfused and necrotic tissue.

Magnetic resonance imaging of chronic myocardial infarcts in man.

To evaluate the magnetic resonance imaging (MRI) features of chronic myocardial infarction (MI), 22 patients and several normal volunteers were studied with a 0.35-T cryogenic imaging system. The MIs were 9 months to 16 years old. The patients also had either left ventriculography (17 patients) or two-dimensional echocardiography (17 patients). At least one abnormality indicative of prior infarction was demonstrated on MRI in 20 of the 22 patients. Wall thinning was seen in 20 patients; in six of these, the thinning resulted in aneurysm formation. The other 14 patients had sufficient residual wall thickness to permit measurement of T2 relaxation times and MR signal intensity in the infarcted region. Ten of these 14 patients demonstrated low intensity and shortened T2 of the thinned segments (mean T2 = 28.7 msec) compared to adjacent normal myocardium (mean T2 = 45.4 msec) and to the myocardium of volunteers (mean T2 = 41.3 msec). The percentage of difference in intensity between thinned and normal myocardium was greater on 56-msec-TE images (98.2%) than on 28-msec-TE images (46.1%). In the other four patients, no difference in intensity of the myocardium was perceptible in the thinned region of the myocardial wall. Thus MRI shows regional wall thinning at the site of prior MI. In some patients, the chronic infarct is characterized as decreased spin-echo signal intensity and shortened T2 consistent with replacement of myocardium by fibrous scar.

Acute myocardial infarction: MR evaluation in 29 patients.

This study evaluated the ability of MR to identify and characterize the region of myocardial infarction in humans. Twenty-nine patients, all with ECG and enzyme rises consistent with an acute myocardial infarction, were studied by MR 3-17 days from the onset of acute chest pain. Four patients were excluded because of inability to acquire adequate MR studies. For comparison, 20 normal subjects were studied who also had gated MR examinations. The site of infarction was visualized in 23 patients as an area of high signal intensity in relation to the normal myocardium, a contrast that increased on the second-echo image. The regions of abnormal signal intensity corresponded to the anatomic site of infarction as defined by the ECG changes. The mean T2 relaxation time of the infarcted myocardium (79 +/- 22 msec) was significantly prolonged in comparison with the mean T2 (43.9 +/- 9 msec) of normal myocardium (p less than .01). The mean percentage of contrast (intensity difference) between normal and infarcted myocardium was much greater on the second-echo images (65.6 +/- 34.0%) than the first-echo images (27.5 +/- 18.7%). In the normal subjects there was no difference in T2 between the anterolateral (40.3 +/- 5.7 msec) and septal (39.5 +/- 7.4 msec) regions, and percentages of contrast between these two regions of myocardium on the first-echo (9.1 +/- 7.4%) and second-echo (15.0 +/- 13.3%) images were similar. Thus, MR can be used to directly visualize acute infarcts. However, it has several pitfalls, including the necessity to differentiate signal from slowly flowing blood in the ventricle, from increased signal from a region of infarction and artifactual variation of signal intensity in the myocardium due to respiratory motion or residual cardiac motion.

Early MR follow-up of partial hepatectomy.

Twenty-seven patients were studied with MRI between 3 and 40 days following partial liver resection. Twenty-four patients had undergone major hepatectomy (three to six segments) and three had undergone minor hepatectomy (tumorectomy, one; bisegmentectomy, two). Indications for surgery were as follows: metastases (n = 16), hepatocellular carcinoma (n = 5), hemangioma (n = 3), focal nodular hyperplasia (n = 2), and cholangiocarcinoma (n = 1). A total of 36 MR examinations were performed using a 1.5 T superconducting unit. Three patients were studied three times and three patients were studied twice. The MR images were evaluated to detect and to characterize liver parenchymal abnormalities and intraabdominal fluid or blood collections as well as to assess vascular and/or graft patency. The MR images showed hepatic ischemia in two cases and allowed differentiation between intraabdominal hemorrhagic (n = 30 and nonhemorrhagic (n = 4) fluid collections. Gradient echo images allowed assessment of polytetrafluoroethylene graft patency as well as demonstration of iliac vein (one case) and portal vein (one case) thrombosis. The presence of surgical clips at the resection margins did not affect image quality.

Spin-echo and STIR MR imaging of sports-related muscle injuries at 1.5 T.

Seventy patients with clinically diagnosed athletic muscle injuries of varying severity were studied with MR imaging at 1.5 T. Twenty underwent follow-up MR studies. In all cases, SE T1-weighted and double-echo T2-weighted pulse sequences were used. These were supplemented by short T1 inversion recovery (STIR) sequence in 36 cases. Muscle injuries were more readily seen with STIR images than with SE T2-weighted images. In both initial assessment and follow-up of tears, the use of the STIR technique allowed the greatest lesion/muscle contrast. Short TR, short TE SE images provided anatomic detail and were an adjunct to T2-dependent SE images in the evaluation of organized hematomas (11 cases). Follow-up MR studies in 20 patients at variable time intervals allowed demonstration of regression of the tear in 11 cases, fibrous scar formation in 5 cases, and recurrence of the tear in 4 cases. Evolution of hematomas into scar and into cyst was demonstrated in three and two cases, respectively. Owing to the additive effect of T1 and T2 mechanisms, the STIR sequence is well suited for initial evaluation and can replace T2-weighted images in the follow-up of muscle trauma.

MR imaging of the thyroid.

The thyroid gland was evaluated with MR imaging in six normal subjects and 32 patients with thyroid disease. The purpose was to evaluate signal characteristics of normal and diseased thyroid tissue; determine the contrast between normal and diseased tissue on T1- and T2-weighted images; compare relaxation times of normal thyroid, adenomas, and carcinoma; and assess the capability of MR for showing the extent of large thyroid masses. Adenomas and carcinomas were frequently isointense with normal thyroid tissue on T1-weighted images but had markedly higher intensity on T2-weighted images. The mean T1 (1202 +/- 717 msec) and T2 (118 +/- 48 msec) relaxation times of adenomas were markedly longer than the T1 (721 +/- 97 msec) and T2 (59 +/- 10 msec) times of normal thyroid tissue. Likewise, the T1 and T2 values of carcinomas were markedly prolonged compared with normal thyroid but the values overlapped with those of the adenomas. Sagittal and coronal images effectively depicted the extent of large goiters, adenomas, and carcinomas and indicated extension below the cervicothoracic junction. The marked prolongation of relaxation times associated with thyroid disease causes excellent contrast of lesions with normal thyroid and surrounding structures. The large field of view possible with coronal and sagittal images is useful for assessing extensive thyroid masses. These attributes indicate the potential clinical utility of MR for evaluating thyroid disease.

Occlusive and reperfused myocardial infarcts: effect of Gd-DTPA on ECG-gated MR imaging.

In vivo magnetic resonance (MR) imaging was performed to determine the effect of gadolinium-DTPA on MR intensity and relaxation times of occlusive and reperfused acute myocardial infarcts. In 18 dogs the left anterior descending (LAD) coronary artery was ligated. Five hours after LAD artery occlusion, six dogs in group 1 received intravenous Gd-DTPA, 0.5 mmol/kg, and 6 dogs in group 2 received 0.1 mmol/kg. The myocardia of the remaining six dogs (group 3) were reperfused after 1 hour of coronary artery occlusion; these dogs received 0.1 mmol/kg Gd-DTPA intravenously 5 hours later. MR imaging was performed before (control) and 5 minutes after Gd-DTPA administration. Intravenous administration of 0.1 mmol/kg Gd-DTPA significantly improved contrast between infarcted and normal myocardium on T1-weighted spin echo images in group 3. In all groups, T2-weighted precontrast images provided contrast between infarcted and normal myocardium equivalent or better than that provided by T1-weighted postcontrast images. Thus, Gd-DTPA did not improve contrast between ischemically injured and normal myocardium in the early hours after coronary artery occlusion.

Early-phase myocardial infarction: evaluation by MR imaging.

In vivo gated magnetic resonance (MR) imaging was performed in 12 dogs immediately after occlusion of the left anterior descending coronary artery and serially up to 5 hours and again between 4 and 14 days. This was done to evaluate the appearance of acute myocardial infarcts and to determine how soon after coronary artery occlusion MR imaging can demonstrate the site of acute myocardial ischemia. In nine dogs with postmortem evidence of myocardial infarction, regional increase of signal intensity of the myocardium was present by 3 hours after coronary artery occlusion and conformed to the site of myocardial infarct found at autopsy. The signal intensity on T2-weighted images of the infarcted myocardium was significantly greater than that of normal myocardium at 3, 4, and 5 hours after occlusion. The T2 (spin-spin) relaxation time was significantly prolonged in the region of myocardial infarct at 3, 4, and 5 hours postocclusion compared with normal myocardium. Myocardial wall thinning and increased intracavitary flow signal were found in six dogs with comparable pre- and postocclusion images in late systole.

Magnetic resonance imaging with respiratory gating: techniques and advantages.

Respiratory motion is an important problem in magnetic resonance imaging (MRI) of the thorax and upper abdomen. This study assessed several approaches for practical respiratory gating. Methods of acquiring respiratory signals, gated sequencing methods, duration of examination, strategies for reducing examination time, diagnostic quality of gated images, and the influence of respiratory gating on relaxation time measurements were evaluated. Of three different devices for acquiring the respiratory signal, a belt containing a displacement transducer placed around the upper abdomen was found to be most effective and practical. Two pulse-gating modes were implemented, as well as a method for combining cardiac and respiratory gating. Gating methods were tested using phantoms and human volunteers. A spin-conditioned mode of respiratory gating was found to be superior to a more simply implemented triggered mode in which spin-echo (SE) sequencing was interrupted. The time penalty for respiratory gating is technique-dependent. Gated studies with uncontrolled tidal breathing took two to four times longer than nongated studies. When the time between respirations was voluntarily prolonged, gated studies could be only 30%-50% longer than nongated. The standard deviation of relaxation-time measurements for organs that are displaced during respirations was substantially reduced by respiratory gating. Gating acquisition without spin-conditioning gating. Respiratory gating is a practical and useful technique for improving the contrast and spatial resolution of SE images of the upper abdomen and chest. SE images produced with short repetition times were particularly improved by respiratory gating.