Comparison between high-field-strength MR imaging and CT for screening of hepatic metastases: a receiver operating characteristic analysis.

The diagnostic performance of high-field-strength magnetic resonance (MR) imaging (1.5 T) for detection of liver metastases was compared with that of computed tomography (CT). All patients (n = 52) underwent preoperative screening for metastases by means of MR imaging with T1-weighted, proton-density-weighted, and T2-weighted pulse sequences and CT scanning with unenhanced, incremental dynamic bolus-enhanced, and delayed contrast medium-enhanced techniques. Diagnostic performance was evaluated by means of receiver operating characteristic analysis in which 800 images (400 with and 400 without lesions) and five readers (4,000 observations) were used; images were obtained from patients (n = 39) in whom the same anatomic levels were available for all MR imaging and CT studies. Direct comparison between the best MR imaging technique (T2-weighted spin-echo imaging [repetition time, 2,000 msec; echo time, 70 msec]) and the best CT technique (incremental dynamic bolus CT) showed a strong trend of superiority of T2-weighted MR imaging over incremental dynamic bolus CT. No highly statistically significant difference (P greater than or equal to .01), however, was found between these two techniques.     

Cryptic structural lesions in refractory partial epilepsy: MR imaging and CT studies

Results of contrast material-enhanced computed tomography (CT) and T2-weighted spin-echo magnetic resonance (MR) imaging were correlated with pathologic findings in 25 patients treated surgically for refractory partial epilepsy. Of 12 lesions present, ten (83%) were detected by MR imaging and seven (58%) by CT scanning. Of nine low-grade gliomas, eight were detected by MR imaging and four by CT scanning. One posttraumatic scar and one case of temporal lobe atrophy were better demonstrated by MR imaging. A small, thrombosed arteriovenous malformation was the only lesion detected by CT scanning but not by MR imaging. No lesions were detected in 13 patients with mild gliosis and one patient with a 1.2-cm grade 1 astrocytoma. Although more sensitive than CT for detection of structural lesions in patients with refractory partial epilepsy, MR imaging resulted in a 25% false-negative diagnostic rate when a repetition time of 2,000 msec and echo time of 60 msec were used. Multi-echo imaging with at least one long echo time may be needed to increase the sensitivity of MR imaging in these patients.     

Urinary bladder MR imaging. Part II. Neoplasm

The potential of magnetic resonance (MR) imaging for the evaluation and staging of bladder tumors was analyzed in 15 patients (11 cases of transitional cell carcinoma, two adenocarcinomas, one leiomyosarcoma, and one leiomyoma). Neoplasms were characterized by size, site, and growth pattern, and the accuracy of the staging was compared with the results of computed tomography and pathologic study. Malignancies were accurately detected and staged by MR imaging in 12 of 14 patients (85%). Tumor site and degree of bladder distention did not adversely affect detection; tumors greater than 1.5 cm were detected easily. In situ carcinoma (stage Tis) was not detected on MR images. Imaging in both sagittal and transverse planes was needed for optimal bladder evaluation. Bladder carcinoma was best displayed with a short echo delay time (TE) of 28 msec and repetition (TR) times of 1.0-2.0 sec: TR = 1.0 gave 34% contrast and TR = 2.0 gave 36% contrast between tumor and surrounding urine. Bladder-wall invasion by tumor was best evaluated with long TR (2.0 sec) and long TE (56 msec) (82% contrast). For assessing tumor extension into perivesical fat, short TR (0.5 sec) and TE (28 msec) were optimal (58% contrast). MR imaging offers an increased sensitivity for tumor detection and promises to greatly improve the staging of bladder neoplasms.     

Optimal pulse sequence for imaging hepatic metastases

For magnetic resonance (MR) imaging studies in which the diagnosis is dependent on image contrast, it is essential that an optimized imaging technique be used. Using detection of hepatic metastases as an example, the authors describe a rational strategy for optimizing MR imaging technique. First, for a single patient with proved hepatic metastases, a variety of imaging sequences is discussed and evaluated, leading to characterization of the patient's hepatic tissues. Then the characteristics of the tissues of a representative patient population are presented. These are used to determine two optimal pulse sequences that maximize the achievable signal difference-to-noise ratio achievable in a fixed imaging time. The recommended imaging sequence for detection of hepatic metastases at 0.15 T is either a three-dimensional volume spin-echo (SE) sequence with echo time (TE) = 12 msec and repetition time (TR) = 184 msec or a multisection inversion recovery sequence with TE = 22 msec, inversion time = 250 msec, and TR = 1,375 msec. The variation of this optimum pulse sequence with field strength is also presented.     

Magnetic resonance imaging of multiple sclerosis: a study of pulse-technique efficacy

Forty-two patients with the clinical diagnosis of multiple sclerosis were examined by proton magnetic resonance imaging (MRI) at 0.5 T. An extensive protocol was used to facilitate a comparison of the efficacy of different pulse techniques. Results were also compared in 39 cases with high-resolution x-ray computed tomography (CT). MRI revealed characteristic abnormalities in each case, whereas CT was positive in only 15 of 33 patients. Milder grades 1 and 2 disease were usually undetected by CT, and in all cases, the abnormalities noted on MRI were much more extensive than on CT. Cerebral abnormalities were best shown with the T2-weighted spin-echo sequence (TE/TR = 120/1000); brainstem lesions were best defined on the inversion-recovery sequence (TE/TI/TR = 30/400/1250). Increasing TE to 120 msec and TR to 2000 msec heightened the contrast between normal and abnormal white matter. However, the signal intensity of cerebrospinal fluid with this pulse technique obscured some abnormalities.     

Hyperfunctioning and nonhyperfunctioning benign adrenal cortical lesions: characterization and comparison with MR imaging

The authors evaluated the potential of magnetic resonance (MR) imaging at 0.35 T to permit differentiation of nine hyperfunctioning adrenal cortical lesions from 21 nonhyperfunctioning adrenal cortical adenomas. Both qualitative data (visual assessment) and quantitative data (signal intensity ratios, T1, and T2) were used for tissue characterization. With a 2,000/56-100 sequence (repetition time msec/echo time msec), the majority of lesions were visually isointense to liver. Of 34 quantitative measures, only lesion-liver and lesion-kidney intensity ratios at 2,000/150 showed statistically significant differences among nonhyperfunctioning adenomas, aldosterone-producing lesions, and corticosteroid-producing lesions; however, the authors question the significance of these differences because of the abundant noise associated with the 2,000/150 sequence. The results suggest that nonhyperfunctioning adrenal cortical adenomas cannot be distinguished from benign hyperfunctioning cortical lesions with use of MR imaging at 0.35 T.     

MR imaging of the dilated biliary tract

The magnetic resonance (MR) examinations of 18 patients with dilated bile ducts were reviewed retrospectively to determine the capability of MR to demonstrate biliary dilatation, assess MR appearance of the dilated biliary tract using spin-echo techniques, and define the optimal MR imaging parameters (repetition time [TR] and echo time [TE]) for its demonstration. On images with short TR (0.5 sec) and TE (28 msec), the dilated intrahepatic and intrapancreatic bile ducts usually had lower signal intensity compared with the surrounding liver or pancreas; on images with long TR (2.0 sec) and TE (56 msec), they had higher signal intensity. Because of the observed variation in percentage of contrast between dilated bile ducts and surrounding liver and pancreas, two imaging sequences are recommended to obtain reliable demonstration of dilated intrahepatic and intrapancreatic bile ducts. The dilated common bile duct at the level of the hepatic hilus is best seen with a short TR and TE.     

MR evaluation of Chiari I malformations at 0.15 T

Twelve patients with known or presumed Chiari I malformations and two with clinical diagnoses of multiple sclerosis were examined by magnetic resonance (MR) imaging. MR confirmed or established the diagnosis of Chiari I malformation in all 14 cases. The spin-echo technique with a short time to echo (TE = 40 msec) and a short time to recover (TR = 1000 msec) provided optimum imaging of tonsillar position, hydromyelia cavities, and cervicomedullary "kinking." Long TE (greater than 80 msec) and TR (greater than 2000 msec) increase the signal intensity of cerebrospinal fluid and may obscure the pathology. Sagittal, transaxial, and coronal images provided complementary data; sagittal and coronal views best imaged the abnormal spinal cord and tonsils, but slitlike cavities were best seen on transaxial images. Cervicomedullary kinking was found in 10 (71%) of 14 patients and in 90% of the hydromyelic patients. This high incidence suggests that in other radiologic techniques tonsillar herniation masks the kinking. Symptoms of the Chiari I malformation overlap those of demyelinating diseases and brain tumors. Our early experience suggests MR is the preferred noninvasive procedure for identifying Chiari I malformation. Moreover, the ability to portray the variable cavity morphology of hydromyelia directly offers the potential for improved shunt placement.     

Superparamagnetic iron oxide: enhanced detection of focal splenic tumors with MR imaging

Superparamagnetic iron oxide (AMI-25), a reticuloendothelial cell-specific contrast agent for magnetic resonance (MR) imaging, was evaluated for its ability to permit detection of splenic metastases in 18 patients. Superparamagnetic iron oxide, at a dose of 30 mumol of iron per kilogram, decreased the signal intensity of spleen from 19.5 +/- 4.8 to 3.1 +/- 2.2 (spin-echo sequence, repetition time msec/echo time msec = 1,500/42; P less than .05), without changing the signal intensity of tumor. As a result, the tumor-spleen contrast-to-noise ratio increased from 0.2 (tumor isointense relative to spleen) to 18.0 (tumor hyperintense relative to spleen). As a consequence of increased contrast, splenic tumors were detected in four of 18 patients (45 individual lesions; P less than .05), whereas nonenhanced MR imaging permitted detection of splenic lesions in only two of 18 patients (four individual lesions). Maximum tumor-spleen contrast was achieved within 60 minutes after intravenous administration. These initial clinical results indicate that MR imaging with superparamagnetic iron oxide may offer improved accuracy in the diagnosis of splenic tumors.     

Cervical spine: MR imaging with a partial flip angle, gradient- refocused pulse sequence. Part II. Spinal cord disease

A magnetic resonance imaging pulse sequence (GRASS) with a short repetition time (TR), short echo time (TE), partial flip angle, and gradient refocused echo was prospectively evaluated for the detection of cervical cord disease that caused minimal or no cord enlargement in eight patients. Sagittal T2-weighted, cerebrospinal fluid (CSF)-gated images and sagittal and axial GRASS images were obtained in all patients. The following GRASS parameters were manipulated to determine their effect on signal-to-noise ratio (S/N) and contrast: flip angle (4 degrees-18 degrees), TR (22-50 msec), and TE (12.5-25 msec). Flip angle had the greatest effect on S/N and contrast. There were no differences between axial and sagittal imaging for the spinal cord or lesion. However, because the signal intensity of CSF did differ on sagittal and axial images and because this influenced the conspicuity of lesions, there was a difference in the useful flip angle range for axial and sagittal imaging. No one set of imaging parameters was clearly superior, and in all patients, the gated image was superior to the sagittal GRASS image in lesion detection. GRASS images should be used in the axial plane primarily to confirm spinal cord disease detected on sagittal CSF-gated images. For this, a balanced approach is suggested (TR = 40 msec, TE = 20 msec, with flip angles of 4 degrees-6 degrees for sagittal and 6 degrees-8 degrees for axial imaging).     

Proton magnetic resonance spectroscopic imaging reveals differences in spinocerebellar ataxia types 2 and 6

The objective of this study was to investigate cerebellar metabolism in patients with autosomal dominant cerebellar ataxia type 1 (ADCA-I) carrying two distinct mutations of spinocerebellar ataxia (SCA). Non-invasive image-guided proton magnetic resonance spectroscopy imaging (1H-MRSI) was performed in 4 patients with SCA2, and 3 patients carrying the SCA6 mutation. For MRSI, we employed a spin-echo sequence (TR = 1500 msec, TE = 135 msec, slice thickness = 15 mm, FOV = 240 mm) and a stimulated-echo sequence (TR = 1500 msec, TE = 20 msec, slice thickness = 15 mm, FOV = 240 mm). Measures included the peak integral ratios of neuronal and glial markers [N-acetylaspartate (NA) to creatine (Cr), choline-containing compounds (CHO) to Cr, and lactate (LAC) to Cr]. We found NA:Cr ratios were significantly lower in patients with SCA2 (40.4% lower) compared to patients carrying the SCA6 mutation. CHO:Cr ratios differed between the two mutations using short echo time (30.8% lower in SCA2), but not when applying long echo time 1H-MRSI. Measurements using long echo time revealed LAC peaks in all SCA2 patients. 1H-MRSI revealed metabolic differences between SCA2 and SCA6 patients. NA:Cr ratios were significantly lower in patients with the SCA2 mutation compared to the SCA6 mutation, and LAC signals were obtained in the cerebella of SCA2 patients. In addition, CHO:Cr ratios showed different behavior using short and long TE, indicating differences in relaxation times of choline compounds in SCA2.     

Adrenal gland: MR imaging

The authors investigated the utility of magnetic resonance (MR) imaging in identifying the normal adrenal gland in 100 patients as well as in distinguishing adrenal adenomas (n = 12) from malignant neoplasms (n = 14). The left adrenal gland was seen in 99 of 100 cases and the right in 91 of 100 cases. The adrenals were most easily seen with T1-weighted spin-echo pulse sequences. The ratio of the intensity of the adrenal mass to that of fat at 2,100/90 (repetition time msec/echo time msec) was most helpful in distinguishing adrenal adenomas from malignant neoplasms. In contrast to other studies, the adrenal mass/liver intensity ratios were not helpful. All ten lesions with adrenal mass/fat ratios at 2,100/90 of 0.8 or greater were malignant, whereas all eight adrenal masses with a ratio less than 0.6 were adrenal adenomas. However, eight (31%) of the masses (four adenomas and four malignant neoplasms) had ratios between 0.6 and 0.8. Although MR imaging has considerable potential in characterizing adrenal masses, larger studies are needed to determine its true sensitivity and specificity.     

Coagulative interstitial laser-induced thermotherapy of benign prostatic hyperplasia: online imaging with a T2-weighted fast spin-echo MR sequence--experience in six patients

PURPOSE: To determine if hypointense lesions clearly outline on T2-weighted fast spin-echo (SE) magnetic resonance (MR) images obtained during coagulative interstitial laser-induced thermotherapy (LITT) of a prostate with benign hyperplasia. MATERIALS AND METHODS: In six patients with benign prostatic hyperplasia (BPH), 12 LITT treatments were followed online with repetitive axial T2-weighted fast SE imaging (repetition time, 3,700 msec; echo time, 138 msec; acquisition time, 19 seconds). Development, time course, correlation with interstitial tissue temperature, and diameters of hypointense lesions around the laser diffusor tip were investigated. Lesion diameters on T2-weighted images acquired during LITT were compared with diameters of final lesions on T2-weighted images and unperfused lesions on enhanced T1-weighted SE images obtained at the end of therapy. RESULTS: Hypointense lesions developed within 20-40 seconds of LITT. Average correlation coefficients between interstitial temperature development and signal intensity development were 0.92 during LITT and 0.90 after LITT. Regression slopes were significantly steeper during LITT (0.67% signal intensity change per degree Celsius) than after LITT (0.47% per degree Celsius; P = .038). Lesions remained visible after LITT for all procedures. Average maximum diameters of lesions were 1-3 mm larger during LITT than after LITT (P = .0006-.019). CONCLUSION: Repetitive T2-weighted fast SE MR imaging during interstitial coagulative LITT of BPH demonstrates the development of permanent hypointense prostate lesions. However, posttherapeutic lesion diameters tend to be overestimated during LITT.     

Hepatocellular carcinoma and cavernous hemangioma: differentiation with MR imaging. Efficacy of T2 values at 0.35 and 1.5 T

Seventy-two patients with hepatocellular carcinoma (hepatoma) and 56 with hemangioma were studied with magnetic resonance (MR) imaging at 0.35 or 1.5 T to evaluate the efficacy of T2 values in differential diagnosis. T2 values were calculated with the two-point method. The mean T2 values of hepatoma and hemangioma were 58.9 msec +/- 8.9 and 101.6 msec +/- 25.8 at 0.35 T and 49.1 msec +/- 9.8 and 85.3 msec +/- 21.2 at 1.5 T. The difference in the T2 values for hepatoma and hemangioma was statistically significant (P less than .001) at both 0.35 and 1.5 T. Fifty-three of 56 lesions (94.6%) at 0.35 T and 86 of 102 lesions (84.3%) at 1.5 T were correctly classified when the T2 borderline between hepatoma and hemangioma was set at 80 msec. All misdiagnosed lesions were hemangioma, and all but one were smaller than 2 cm. However, over 90% of lesions smaller than 2 cm were correctly diagnosed when 70 msec at 0.35 T and 60 msec at 1.5 T were used as borderline T2 values. MR imaging with T2 measurement was very useful for differentiating between hepatoma and hemangioma (including small lesions) at 1.5 T as well as at 0.35 T.     

Segmented turboFLASH: method for breath-hold MR imaging of the liver with flexible contrast

A method called segmented turboFLASH imaging allows high-resolution, multisection, short-inversion-time (TI) inversion-recovery (STIR), T1- or T2-weighted magnetic resonance (MR) studies of the liver to be completed within a breath-hold interval. The method was applied in a phantom and in 19 patients with hepatic lesions. Sequence comparisons were performed among segmented turboFLASH, single-shot turboFLASH, T1-weighted gradient-echo with ultrashort echo time, and T2-weighted spin-echo (SE) techniques. Signal from fat and liver could be nulled with the segmented turboFLASH method, with TIs of 10 and 300 msec, respectively; signal from these tissues could not be eliminated with the single-shot approach. Signal-difference-to-noise ratios and contrast for the best segmented sequences were comparable with those of the best T2-weighted SE and T1-weighted gradient-echo techniques. It is concluded that it is feasible to obtain breath-hold images with arbitrary tissue contrast by means of segmented turboFLASH imaging. The method may prove helpful for the detection and characterization of hepatic lesions and will likely have applications to other anatomic regions such as the chest and pelvis.     

MR imaging of spinal epidural sepsis

Four patients with spinal epidural sepsis were evaluated with MR imaging. The lesions were best visualized with spin-echo techniques with long repetition (2000 msec) and long echo (80-100 msec) times. Sagittal and axial images were equally important in defining the extent of the lesions. Comparison with available contrast-enhanced CT scans showed that MR was more definitive in the early demonstration of the abscesses. This early recognition influenced the management greatly and improved the clinical outcome significantly. The findings in our four cases support previous reports that MR is superior to other imaging methods for early recognition and anatomic localization of infectious diseases in patients suspected of having either spinal osteomyelitis or spinal epidural sepsis.     

Comprehensive MR imaging of acute gynecologic diseases.

Rapid advances in techniques of magnetic resonance (MR) imaging have enabled diagnosis of acute gynecologic conditions, which are characterized by sudden onset of lower abdominal pain, fever, genital bleeding, intraperitoneal bleeding, or symptoms of shock. The chemical-selective fat-suppression technique not only helps establish the characteristics of lesions that contain fat components but also increases the conspicuity of inflammatory lesions. When a T2-weighted image is obtained with a very long effective echo time (>250 msec), even a small amount of ascites can be easily identified and the contrast between urine and complex fluid becomes more conspicuous. T2*-weighted images are useful for identification of hemorrhagic lesions by demonstrating deoxyhemoglobin and hemosiderin. Contrast material-enhanced dynamic subtraction MR imaging performed with a three-dimensional fast field-echo sequence and a rapid bolus injection of gadopentetate dimeglumine allows evaluation of lesion vascularity and the anatomic relationship between pelvic vessels and a lesion and allows identification of the bleeding point by demonstrating extravasation of contrast material. To optimize the MR imaging examination, attention should be given to the parameters of each pulse sequence and proper combination of the sequences.     

Urinary bladder MR imaging. Part I. Normal and benign conditions

The normal urinary bladder and several benign entities of the bladder were examined in 50 patients by magnetic resonance (MR) imaging. Specific features assessed included appearance of the bladder wall, optimal repetition (TR) and echo delay (TE) parameters for bladder-wall demonstration, and differentiation among various benign abnormalities, including bladder-wall hypertrophy, inflammation, and mucosal congestion, on MR images. The bladder wall in the 30 healthy subjects was best displayed using a TR = 2 sec, TE = 56 msec image, which gave 60% contrast between the bladder wall and urine and 48% contrast between the bladder wall and fat. Demonstration of bladder-wall hypertrophy required similar imaging; bladder distention was necessary to demonstrate the thickness of the bladder wall. Congestion and inflammation were best demonstrated on TR = 2 sec, TE = 56 msec images, which gave 45% contrast. Normal and/or hypertrophic bladder wall were distinguished from inflammation and congestion on the basis of signal intensity variations and/or T1 and T2 relaxation parameters.     

Breast and axillary tissue MR imaging: correlation of signal intensities and relaxation times with pathologic findings

We tested a variety of inversion-recovery (IR) and spin-echo (SE) sequences by imaging the breast masses of 22 patients before surgery and 23 tissue specimens with magnetic resonance (MR) imaging at 0.6 T to determine the most effective pulse sequences to evaluate breast disease. An SE pulse sequence using a long repetition time (TR) of 1,600 msec and a long echo time (TE) of 90 msec was found to be the most sensitive in depicting carcinoma in the excised tissue specimens, with all of the carcinomas (n = 15) demonstrating irregular areas of higher signal intensity (SI) than that of the adjacent fat. However, only five of 11 breast carcinomas present in the preoperative patients produced a higher SI than that produced by fat on the same T2-weighted sequence. Five of the remaining six carcinomas in the preoperative patients appeared as localized distortions of fibroductular architecture on both T2-weighted SE and IR sequences. In axillary tissue specimens, both metastatic carcinoma and hyperplastic lymph nodes produced a high SI on T2-weighted SE sequences. However, metastatic carcinoma had a significantly longer T2 relaxation time than did hyperplastic lymph nodes.     

In vivo fluorine-19 MR imaging: Relaxation enhancement with Gd-DTPA

The lack of a naturally occurring background signal from fluorine in magnetic resonance (MR) imaging makes fluorinated compounds potentially attractive candidates for tissue-specific MR contrast agents. Problems associated with the in vivo use of fluorinated compounds are toxicity, which limits the amount of agent that can be used; multiple resonance lines; and an excessively long T1, which leads to long sequence TRs and consequently long imaging times. Many fluorinated agents also possess complex MR spectra that result in chemical shift artifacts if not corrected. The authors demonstrate the use of an extracellular fluorinated agent with a single MR peak for selective imaging of a brain abscess in an animal model and show that the image signal per unit of acquisition time can be enhanced through the use of a T1 relaxation agent, gadolinium diethylenetriamine-pentaacetic acid (DTPA). Trifluoromethylsulfonate was administered at a fluorine-19 dose of 4 mmol/kg, and fluorine images of the induced abscess were acquired before and after the injection of a standard dose of Gd-DTPA (0.1 mmol/kg); non—section-selected projection images were used. Typical imaging times were less than 5 minutes. The signal enhancement factor achieved was approximately four (4.0 ± 0.8) with use of a 500/12 (TR msec/TE msec) spinecho sequence.