Soft-tissue sarcomas: detection of metabolic heterogeneity with P-31 MR spectroscopy

Regional variations in metabolic parameters derived with multivoxel, localized phosphorus-31 magnetic resonance (MR) spectroscopy in spontaneous human-soft-tissue sarcomas were compared with variations in the same parameters in normal human legs. In addition, multivoxel P-31 MR spectroscopy of transplanted rodent sarcomas and microelectrode measurement of pH and PO2 in several locations within them were performed, and, when appropriate, results were compared with the clinical data. Striking voxel-voxel variability in parameters derived with P-31 spectroscopy was found in tumors, with less marked variability seen in normal legs. In the rodent tumors, spatial variations also were found in pH and PO2 measured by means of microelectrodes. These data are consistent with the results of regional measurements in tumors by means of other methods and suggest that clinical MR spectroscopic studies of tumors may need to consider the variability within malignant lesions.     

Carotid bifurcation: MR imaging. Work in progress

To devise and implement an in-plane magnetic resonance angiography examination of the carotid bifurcation capable of producing high-resolution images, the authors examined 19 normal carotid arteries and 14 patients with angiographically documented disease with two flow-correction techniques: a three-gradient, velocity-refocused technique with spin-echo (SE) and gradient-echo sequences, and a four-gradient velocity- and acceleration-corrected SE technique. With use of three equal gradients in the read direction, velocity-related phase changes were minimized by placing the dephasing gradient after the 180 degree pulse and near the read gradient. Acceleration effects were minimized through the use of short echo times and cardiac gating. Both velocity- and acceleration-produced phase changes were corrected with the four-gradient scheme but at the expense of some limitations in spatial resolution. Both techniques consistently produced satisfactory images of the carotid bifurcation in healthy individuals. However, the results indicate that the present gradient-phase modulation techniques have several drawbacks, including susceptibility to patient motion, overlapping with the jugular vein, and inability to image carotid stenosis accurately due to turbulence.     

MR relaxometry imaging. Work in progress.

Acquisition of relaxation rate dispersion curves from magnetic resonance images was demonstrated on a clinical, whole-body imaging system. Study of the behavior of relaxation rates over a range of field strengths probes the structural environment of imaged hydrogen protons and reveals information about the composition of tissue. The authors determined relaxation rates in extremities and heads of healthy volunteers. The sensitivity of the measurement is sufficient to obtain a distinctive relaxation rate dispersion behavior for different tissues.     

Soft-tissue tumors: MR imaging

We evaluated the suspected soft-tissue masses of 33 patients using magnetic resonance (MR) imaging. Thirty-two masses were defined, of which ten were malignant and 22 were benign. Specific characteristics that would allow us to distinguish the benign lesions from the malignant ones could not be identified on MR imaging. Lesions located within muscles or in the intramuscular septa were best seen with T2-weighted imaging sequences; those located within the subcutaneous fat were best outlined with T1-weighted imaging sequences. Both imaging sequences were necessary to provide the complete representation of the extent of the mass. We performed comparison studies using computed tomography (CT) in 24 of the cases. Of 92 possible comparisons and in four categories MR imaging yielded results that were superior to those obtained by CT scanning in 30 instances, equaled the results obtained by CT scanning in 62 instances, but never yielded results inferior to those obtained by CT scanning. However, MR imaging failed to demonstrate soft-tissue calcification and soft-tissue gas in one case each.     

Gated cardiac MR imaging and P-31 MR spectroscopy in humans at 1.5 T. Work in progress

The authors describe a technique to acquire gated cardiac proton magnetic resonance images and localized hydrogen-1 and phosphorus-31 spectra with a doubly tuned surface coil probe. The probe reduces study time because the need to exchange imaging and spectroscopy coils is eliminated, while at the same time the sensitivity of singly tuned coils is retained. In addition, the probe enhances the ability of the investigator to localize cardiac spectra spatially and temporally during the cardiac cycle. Spectra were acquired from the left ventricular myocardium in five volunteers, and systolic and diastolic gated P-31 spectra were shown.     

Noise-reduced prostatic MR imaging. Work in progress

Measurement-dependent filtering, a nonlinear noise-reduction technique, was used to improve the signal-to-noise ratio of in vivo T2-weighted magnetic resonance images of the prostate gland. In both normal and abnormal prostates, the technique considerably reduced noise in T2-weighted images. The technique may provide more accurate depiction of regions of benign prostatic hyperplasia and carcinoma in the prostate.     

Soft-tissue sarcomas: use of textural patterns in skeletal muscle as a diagnostic feature in postoperative MR imaging.

The authors assessed the value of tissue textural patterns as a diagnostic feature for differentiating nonspecific posttreatment tissue changes from musculoskeletal sarcoma recurrence on magnetic resonance (MR) images. The MR imaging studies of 40 patients who had previously undergone surgery and radiation therapy for soft-tissue sarcomas of the lower extremities were evaluated in a blind fashion. In 31 of the MR imaging studies, T2-weighted images demonstrated diffuse areas of high signal intensity in soft tissues at the operative region. Close examination of the corresponding regions on high-resolution transverse T1-weighted images demonstrated textural features typical of skeletal muscle in 23 patients and the absence of such features in eight. None of the 23 patients with the "texture sign" proved to have macroscopic tumor recurrence at clinical or surgical follow-up. Among the eight patients without recognizable textural features of muscle in the regions suspicious for tumor recurrence, two proved to have recurrent tumor at surgery. Recognition of a texture sign on high-resolution T1-weighted spin-echo images of regions suggestive of tumor recurrence helps improve the diagnostic specificity of follow-up MR examinations in patients who have undergone treatment for soft-tissue sarcomas.     

Comparison of CT, low-field-strength MR imaging, and high-field-strength MR imaging. Work in progress

To assess objectively the sensitivity and specificity of low-field-strength (0.064 T) magnetic resonance (MR) imaging, a prospective blind study of 280 examinations was performed to compare low-field-strength MR imaging with computed tomography (CT) and with high-field-strength (1.5-T) MR imaging of the cranium. The sensitivity (defined as the true-positive rate) with high-field MR imaging was superior to that with low-field MR imaging and CT in helping detect overall abnormalities. Sensitivities were generally similar over a broad range of specific cranial central nervous system diseases. Low-field and high-field MR imaging were equivalent in the blind diagnoses of neoplasms and white matter disease, whereas low-field MR and CT were equivalent in the blind diagnoses of contusion, subdural and epidural hematoma, sinus disease, normality, and abnormality. The specificities with low-field MR imaging and CT were substantially better than those with high-field MR imaging.     

Soft-tissue masses: diagnosis using MR imaging

The MR images of 112 soft-tissue masses of various causes were retrospectively reviewed. Pathologic diagnosis by biopsy was available in 96 cases. Diagnosis in the remaining 16 cases was established by characteristic radiographs, CT scans, and/or arteriograms, in conjunction with appropriate history and clinical follow-up. All masses were evaluated with both T1-weighted, 300-600/20-30 (TR/TE), and T2-weighted, 2000/80-100, images. They were reviewed to determine (1) if these images were sufficiently unique to allow a preoperative diagnosis based exclusively on the MR appearance and (2) if benignity vs malignancy could be predicted on the basis of the analysis of the MR image characteristics of the lesion. Concerning the latter, attention was directed to the margins of the lesions, to the impact of the lesion on the surrounding tissues (edema, infiltration, etc.), and to the intensity and homogeneity of the MR signal of the lesion. MR images were sufficiently characteristic to allow a specific diagnosis in 27 (24%) of the 112 cases (10 lipomas, eight hemangiomas, six pigmented villonodular synovitis, two hematomas, and one arteriovenous malformation). MR was incapable of reliably distinguishing between benign and malignant soft-tissue tumors.     

High-resolution MR imaging using loop-gap resonators. Work in progress

Representative results using loop-gap resonator (LGR) configurations for local high-resolution MR imaging (0.6 mm X 0.6 mm pixel size with 3-mm section thickness) are presented. Four geometries were employed: flat single loops, coplanar (planar pair) loops, coaxial (axial pair) loops, and symmetrically angled (butterfly pair) loops. The LGRs were used as receiver coils with the body coil as transmitter. Butterfly pairs and planar pairs are intrinsically decoupled from a transmitted field of arbitrary orientation, whereas the single loops and axial pairs are decoupled from a field that is linearly polarized and orthogonal to its axis. The coils were evaluated using computer-generated and experimentally determined field distributions and by the imaging of human subjects on a 1.5-T MR system. Matching the region of sensitivity of the coil to the region of anatomic interest is emphasized.     

MR imaging of uterine sarcomas.

OBJECTIVE: The MR imaging appearances of uterine sarcomas are not well described in the literature. We describe the MR imaging features of uterine sarcomas. MATERIALS AND METHODS: MR images from all patients with histologically proven uterine sarcomas scanned between 1993 and 2000 were reviewed. Tumor size, its relationship to the uterus, signal characteristics, and enhancement pattern after IV injection of gadolinium were noted. RESULTS: Twenty-five scans from 22 patients were reviewed. Findings from the scans included 11 leiomyosarcomas, five mixed müllerian tumors, two rhabdosarcomas, and four endometrial stromal sarcomas. Two patterns of disease were observed, including a characteristic large heterogenous pelvic mass (n = 17) and an endometrial mass indistinguishable from endometrial carcinoma (n = 8). On T2-weighted images, the large masses were characteristically of low or intermediate background signal intensity with pockets of very high T2 signal. The areas of high T2 signal corresponded to cystic necrosis in the tumor. Pockets of high T1-weighted signal corresponded to hemorrhage. Gadolinium enhancement was present in the solid components of all tumors. This pattern was observed in all recurrent sarcomas. Some correlation was shown between the histologic subtypes and the MR imaging appearances. CONCLUSION: Uterine sarcomas show two patterns on MR imaging. The most common presentation is a large heterogenous mass. However, sarcomas can mimic endometrial carcinoma.     

Clinical and experimental vasogenic edema: in vivo sodium MR imaging. Work in progress

To investigate the sodium magnetic resonance (MR) imaging characteristics of acute vasogenic edema, an experimental canine model was developed. Vasogenic edema was produced in the hemisphere of the dogs by the intraarterial infusion of hypertonic mannitol (25%). This solution opens the blood-brain barrier, allowing the influx of water, electrolytes, and proteins into the brain. The main advantage of this model over the established "cold injury" model is the lack of associated brain necrosis. Two patients with chronic vasogenic edema secondary to well-circumscribed meningiomas also underwent MR imaging. The sodium signal was markedly elevated in both clinical and experimental studies of vasogenic edema fluid compared with signal in healthy brain tissue. Extracellular sodium associated with vasogenic edema displayed MR imaging characteristics similar to that of sodium in serum. There was a trend toward a shortened T2 in edema fluid secondary to the presence of serum macromolecules.     

Association of heterotopic gray matter with seizures: MR imaging. Work in progress

Heterotopic gray matter, which previously had been associated with severe congenital malformations of the brain and developmental delay, was found without these associated conditions. The authors found ten cases of heterotopic gray matter on magnetic resonance (MR) images. The lesions had a signal intensity that was isointense compared with that of gray matter on T1, spin-density, and T2-weighted images. Nine of the ten cases were associated with a seizure disorder. The tenth case, discovered during a workup for metastatic lung disease, was confirmed with pathologic studies. Heterotopic gray matter is the presence of cortical neurons in an abnormal location, which may be periventricular (nodular) or within the white matter (laminar). A knowledge of heterotopic gray matter and its association with seizures may prevent the misinterpretation of findings on MR images.