Preliminary results of a bone marrow magnetic resonance imaging protocol for patients with high-risk prostate cancer.

OBJECTIVES: To compare the accuracy of a bone marrow magnetic resonance imaging (MRI) protocol in patients at high risk of metastatic disease with radioisotopic bone scans, the standard method for detection of bony metastases in patients with prostate cancer. METHODS: The study group consisted of 19 men with prostate cancer who underwent a bone marrow MRI between November 1993 and February 1996. This protocol images the marrow of the thoracolumbar spine, sacrum, pelvis, and femurs. Indications for MRI included an equivocal bone scan and/or staging of locally advanced or recurrent disease. The findings on MRI and bone scan were compared and the results correlated with the subsequent clinical patient outcome. RESULTS: The bone marrow MRI protocol detected metastatic disease in 1 (7%) of 13 patients with negative bone scans. Four patients had an indeterminate bone scan: 2 had true-positive MRIs, 1 a true-negative MRI, and 1 a false-positive MRI on the basis of subsequent clinical follow-up. Two patients with positive bone scans had true-positive MRIs. CONCLUSIONS: Although not recommended for routine staging, MRI was useful in this study for clarifying an equivocal bone scan. The bone marrow MRI protocol images a high yield volume of the bony skeleton and is fast and economical compared with obtaining many focused MRI scans of these areas separately. These preliminary data suggest that further investigation of its clinical utility for staging locally advanced or recurrent disease is justified.

Simultaneous 3D NMR spectroscopy of proton-decoupled fluorine and phosphorus in human liver during 5-fluorouracil chemotherapy.

Simultaneous acquisition of 1H-decoupled 31P and 19F 3D CSI is demonstrated in the liver of a patient undergoing 5-fluorouracil chemotherapy. Both 31P and 19F shared the same voxel size (64 or 27 ml), bi-level 1H-decoupling and 0.35 s TR. The measurements were done in a 1.5 Tesla clinical imager with three radio-frequency (RF) channels and a triple-tuned surface-coil. The overall MRI and MRS examination time was under 90 min. Simultaneous acquisition of 31P and 19F permits localized study of the influence of hepatic metabolism on the uptake and catabolism of fluoropyrimidine drugs without extra measurement time or higher SAR.

Metabolic characterization of human soft tissue sarcomas in vivo and in vitro using proton-decoupled phosphorus magnetic resonance spectroscopy.

We applied 1H-decoupling and nuclear Overhauser enhancement to obtain well-resolved 31P magnetic resonance spectra accurately localized to 20 soft tissue sarcomas in vivo, using three-dimensional chemical shift imaging. Fifteen spectra had large phosphomonoester signals (21% of total phosphorus) that contained high amounts of phosphoethanolamine (compared to those of phosphocholine) but no signals from glycerophosphoethanolamine, and glycerophosphocholine was detected in only four cases. Prominent nucleoside triphosphates (52% of phosphorus) and low inorganic phosphate (10% of phosphorus) indicated that a large fraction of these 15 sarcomas contained viable cells, and this impression was confirmed histologically in 13 of the sarcomas. High-resolution in vitro 31P spectra of extracts of surgical specimens of four of the sarcomas studied in vivo and six additional sarcomas confirmed the in vivo assignments of metabolites and revealed considerable inter- and intratumoral variations of metabolite concentrations associated with histological variations in the relative amounts of cells and of matrix materials or spontaneous necrosis. Seven sarcomas, all high grade with pleomorphic or round cells rather than spindle cells, contained an unidentified phosphodiester signal in vivo; its absence in the extract spectra indicates that it may be from an abnormally mobile membrane component. We have documented a means to obtain new information about in vivo metabolism in human sarcomas and to provide a basis on which to examine the uses of 31P magnetic resonance spectroscopy in the clinical management of sarcomas.

Molar quantitation of hepatic metabolites in vivo in proton-decoupled, nuclear Overhauser effect enhanced 31P NMR spectra localized by three-dimensional chemical shift imaging.

Proton decoupling and nuclear Overhauser effect (NOE) enhancement significantly improve the signal-to-noise ratio and enhance resolution of metabolites in in vivo 31P MRS. We obtained proton-decoupled, NOE-enhanced, phospholipid-saturated 31P spectra localized to defined regions within the normal liver using three-dimensional chemical shift imaging. Proton-decoupling resulted in the resolution of two major peaks in the phosphomonoester (PME) region, three peaks in the phosphodiester (PDE) region and a diphosphodiester peak. In order to obtain molar quantitation, we measured the NOE of all hepatic phosphorus resonances, and we corrected for saturation effects by measuring hepatic metabolite T1 using the variable nutation angle method with phase-cycled, B1-independent rotation, adiabatic pulses. After corrections for saturation effects, NOE enhancement, B1 variations and point spread effects, the following mean concentrations (mmol/l of liver) (+/-SD) were obtained: [PME1] = 1.2 +/- 0.4, [PME2 + 2,3-DPG] = 1.1 +/- 0.1, [Pi + 2,3-DPG] = 2.8 +/- 0.5, [GPEth] = 2.8 +/- 0.7, [GPChol] = 3.5 +/- 0.6 and [beta-NTP] = 3.8 +/- 0.3. T1 and NOE enhancement were strongly correlated (r = 90), and indicated that the fractional contribution of 1H-31P dipolar relaxation to total 31P relaxation is minimal for NTPs, moderate for PMEs and high for PDEs in liver. Proton-decoupling and NOE enhancement permit one to obtain more information about in vivo metabolism of liver than previously available and should enhance the utility of 31P MRS for the study of hepatic disorders.

Proton magnetic resonance spectroscopy in patients with glial tumors: a multicenter study.

The authors represent a cooperative group of 15 institutions that examined the feasibility of using metabolic features observed in vivo with 1H-magnetic resonance (MR) spectroscopy to characterize brain tumors of the glial type. The institutions provided blinded, centralized MR spectroscopy data processing long with independent central review of MR spectroscopy voxel placement, composition and contamination by brain, histopathological typing using current World Health Organization criteria, and clinical data. Proton 1H-MR spectroscopy was performed using a spin-echo technique to obtain spectra from 8-cc voxels in the tumor and when feasible in the contralateral brain. Eighty-six cases were assessable, 41 of which had contralateral brain spectra. Glial tumors had significantly elevated intensities of choline signals, decreased intensities of creatine signals, and decreased intensities of N-acetylaspartate compared to brain. Choline signal intensities were highest in astrocytomas and anaplastic astrocytomas, and creatine signal intensities were lowest in glioblastomas. However, whether expressed relative to brain or as intratumoral ratios, these metabolic characteristics exhibited large variations within each subtype of glial tumor. The resulting overlaps precluded diagnostic accuracy in the distinction of low-and high-grade tumors. Although the extent of contamination of the 1H-MR spectroscopy voxel by brain had a marked effect on metabolite concentrations and ratios, selection of cases with minimal contamination did not reduce these overlaps. Thus, each type and grade of tumor is a metabolically hetero-geneous group. Lactate occurred infrequently and in all grades. Mobile lipids, on the other hand, occurred in 41% of high-grade tumors with higher mean amounts found in glioblastomas. This result, coupled with the recent demonstration that intratumoral mobile lipids correlate with microscopic tumor cell necrosis, leads to the hypothesis that mobile lipids observed in vivo in 1H-MR spectroscopy may correlate independently with prognosis of individual patients.

Quantitation of 5-fluorouracil catabolism in human liver in vivo by three-dimensional localized 19F magnetic resonance spectroscopy.

The development of clinical applications of 19F magnetic resonance (MR) spectroscopy of 5-fluorouracil (5-FU) has been limited by the inability to localize 19F spectra to specific regions of interest, making it difficult to quantitate drug and metabolite concentrations accurately. To develop methodology for quantitation, we studied the liver of patients receiving rapid bolus i.v. injections of 5-FU. In serial studies, 5-FU disappeared from the liver within 17-26 min, and its catabolite, alpha-fluoro-beta-alanine (FBAL), rose to reach a plateau after 40 min. A high peak level of fluoro-ureido-propionic acid preceded that of FBAL in only one patient, and dihydrofluorouracil was never observed. During the plateau, we obtained MR imaging-directed 19F MR spectra localized using three-dimensional chemical shift imaging. The spin-lattice relaxation time of FBAL in liver, measured using a variable nutation angle method, was 1.6 +/- 0.2 s (mean +/- SD; n = 5). The concentration of FBAL at 60 +/- 10 min after injection was 1.0 +/- 0.2 mm in liver (mean +/- SD; n = 7). This amount represents approximately 20% of the injected dose and 1.4 times the initial hepatic 5-FU concentration. Our approach may permit one to obtain molar concentrations of fluoropyrimidine metabolites simultaneously in hepatic cancers and surrounding liver, and it helps expand pharmacokinetic modeling of fluoropyrimidine catabolism.

MR spectroscopy of musculoskeletal soft-tissue tumors.

Published studies of sarcomas using 31P MRS suffer from technical limitations that include absence of localization to regions of interest, resulting in heavy contamination with signals from muscle, and poor resolution. This review has shown that, in spite of their limitations, many of these studies provide important leads to indicate the directions that need to be taken to further develop clinical and biologic uses of MRS. The uniqueness of the metabolic information available in vivo in a noninvasive manner using MRS provides a major stimulus to pursue these directions. In particular, the potential of 31P MRS to predict treatment sensitivity and resistance in individual cases could lead to a very cost-beneficial clinical use of this procedure. 1H-decoupling and NOE-enhancement, implemented in conjunction with dual-tuned surface coils and accurate localization of 31P MR spectra to regions of interest in three dimensions using CSI, have enabled us to overcome the major technical limitations mentioned earlier, broaden the scope of 31P MRS investigations, and obtain more information about the in vivo metabolic characteristics of soft-tissue sarcomas than has heretofore been available. Our approach, which has been fully implemented in a clinical imager, provides a good technical basis from which to examine potential clinical uses of 31P MRS. In particular, we can now rigorously test the hypotheses, derived from preliminary studies in the literature, that initial metabolic features or early treatment-induced changes in PME predict sensitivity of a sarcoma to that particular treatment. To this end, we at Fox Chase Cancer Center, along with investigators at Duke University, the Institute of Cancer Research/Royal Marsden Hospital, Johns Hopkins University, Memorial Sloan-Kettering Cancer Center, St. Georges Hospital Medical School, the University of California at San Francisco, and Wayne State University have initiated an NCI-sponsored cooperative trial to examine the role of 31P MRS in the clinical management of soft-tissue sarcomas and other selected cancers.

Metabolic characterization of human non-Hodgkin's lymphomas in vivo with the use of proton-decoupled phosphorus magnetic resonance spectroscopy.

Development of biological and clinical uses of in vivo 31P magnetic resonance spectroscopy has been hampered by poor anatomic localization of spectra and poor resolution of overlapping signals within phosphomonoester and phosphodiester regions of the spectrum. We applied 1H-decoupling and nuclear Overhauser enhancement to improve resolution of 31P magnetic resonance spectra accurately localized to 21 non-Hodgkin's lymphomas (NHL) by using three-dimensional chemical shift imaging. All 21 spectra had large phosphomonoester signals (26% of total phosphorus) that contained high amounts of phosphoethanolamine relative to phosphocholine. There were no signals from glycerophosphoethanolamine or glycerophosphocholine but only a broad signal from membrane phospholipids in the phosphodiester region (20% of phosphorus). Prominent nucleoside triphosphates (47% of phosphorus) and low inorganic phosphate (7% of phosphorus) indicate well-perfused tissue with viable cells. Mean intracellular pH was 7.23. These characteristics were similar in all grades and stages of NHL. By analogy with recently reported studies in cell lines in vitro, we hypothesize that the pattern of phospholipid metabolites observed in NHL in vivo is partly a manifestation of sustained activation of phospholipase C or D. The techniques we implemented permitted us to obtain more information about in vivo metabolism of NHL than has heretofore been available. This information is important for the establishment of appropriate experimental models and provides a basis from which to examine potential clinical uses of 31P magnetic resonance spectroscopy.

Synovial thickening detected by MR imaging in osteoarthritis of the knee confirmed by biopsy as synovitis.

Previous studies have established the value of magnetic resonance imaging (MRI) in detecting articular changes characteristic of osteoarthritis (OA) of the knee. We have observed some MRI features in OA of the knee presumably indicating synovial thickening. To determine whether these MR features represent chronic synovial inflammation, we studied the knees of nine patients at the mild end of the spectrum of OA of relatively short duration (89%: < or = 4 yr), who were selected because MRI showed anatomical abnormalities compatible with synovial thickening. The painful knee was examined using conventional and weight-bearing radiographs, MRI, and arthroscopy. MR images suggestive of synovial thickening typically appeared in or near the intercondylar region of the knee, in the infrapatellar fat pad, or in the posterior joint margin. The site of an arthroscopic biopsy of the synovial membrane was guided by MRI to the area thought to represent synovial thickening for each patient knee. Pathological examination of these synovial membrane biopsies showed a mild chronic synovitis, and thus a correspondence with the synovial thickening detected by MRI. Our results suggest that MRI can be used to evaluate the extent of synovitis, observed as synovial thickening, in patients with early OA of the knee.

Factors affecting articular cartilage thickness in osteoarthritis and aging.

OBJECTIVE: To test the hypothesis that both aging and osteoarthritis (OA) contribute to the variance of human articular cartilage thickness of the knee and whether these contributions occur predominantly at weight bearing cartilage sites. METHODS: Thin, sagittal magnetic resonance images (MRI) of both knees were examined in 52 patients with idiopathic OA of the knee of short duration (87%: < or = 4 years) and 40 reference subjects of comparable age and sex distribution. Articular cartilage thickness was measured at the weight bearing and nonweight bearing femoral condylar, tibial plateau, and posterior patellar sites. Multiple regression analyses were performed to examine the influence of OA, age, sex, obesity (body mass index), and bone size on cartilage thickness. RESULTS: Age accounted for a significant linear decrease (p < 0.001) in both lateral and medial weight bearing femoral cartilage thickness of both knees, but failed to account for any significant variance in thickness at other cartilage sites. OA accounted for a significant decrease (p < 0.02) in the thickness of the femoral articular cartilage in the medial and lateral right knee compartments and in the lateral left knee compartment, but not in the medial left knee compartment. Significant cartilage thinning could be detected by MRI in patients with OA, even when the joint space was normal radiographically. CONCLUSION: The asymmetric decrease in the left knee may reflect the effect of mechanical factors. Our data show that articular cartilage thickness decreases at the femoral weight bearing sites both with age and as a consequence of OA and that these contributions can be distinguished from one another.

MR features of osteoarthritis of the knee.

A group of patients with idiopathic osteoarthritis (OA) of the knee was surveyed using weight-bearing radiographs and MR imaging to compare the relative value of these methods in disease evaluation. Fifty-two patients with a clinical and radiological diagnosis of OA of the knee of relatively short duration (87%: < or = 4 yr) were compared to a reference group of 40 age- and sex-comparable subjects with no knee symptoms. All patients had a complete history, physical examination, standard anterior-posterior and lateral weight-bearing radiographs, T1-weighted, and FLASH MR images in both knees. The prevalence of MRI abnormalities was significantly greater in patients with OA of the knee in all radiographic grades (Kellgren and Lawrence) compared to the reference subjects. Significant differences were encountered for synovial thickening (OA, 73%; reference, 0%), synovial fluid (60%; 7%), meniscal degeneration (52%; 7%), osteophytes (67%; 12%), and subchondral bone involvement (65%; 7%), even in the patients at the mild end of the osteoarthritic spectrum, indicating the exquisite sensitivity of MRI compared with weight-bearing radiographs.

Proceedings of a National Cancer Institute workshop: MR spectroscopy and tumor cell biology.

In December 1991, the National Cancer Institute held a workshop to evaluate the role of magnetic resonance (MR) spectroscopy in human cancer biology. The clinical and basic cancer research issues requiring use of MR spectroscopy, the advantages and limitations of MR spectroscopy, and future directions in MR spectroscopy of cancer were discussed. Consensus-building panels were formed on the following four topics: cell membrane biochemistry, tumor therapeutic response or drug resistance, appropriate model systems, and potential clinical applications of MR spectroscopy. The workshop members concluded that large prospective clinical studies as well as in vivo animal and human studies to define prognostic variables should be performed, with correlation between MR spectroscopic results and biochemical and physiologic features. Studies of phospholipid metabolism, the pharmacokinetics of anticancer agents, and effects of new cancer treatments on the tumor vasculature and normal tissues are needed.

Bone marrow necrosis in lymphoma studied by magnetic resonance imaging.

We used magnetic resonance imaging (MRI) to examine the extent of necrosis in the marrow space in three patients with lymphoma who had necrosis in bone marrow biopsies taken from the posterior iliac crest. Two of the patients had multiple focal lesions observed in the spine, pelvis, and femurs. These lesions had MRI characteristics of fluid-filled necrosis within marrow or lymphoma. The third patient exhibited a unique pattern of anatomically extensive marginated lesions otherwise typical of avascular necrosis of bone. These cases provide evidence that MRI can document the extent of necrosis within the marrow space and can differentiate two distinct patterns of necrosis.

In situ assessment of the rat heart during chronic carbon monoxide exposure using nuclear magnetic resonance imaging.

Ventricular hypertrophy induced in male Sprague-Dawley rats by inhalation of 500 ppm carbon monoxide (40% carboxyhemoglobin level) for 0-62 days was assessed by contiguous 2-mm thick axial cardiac cross-sections, using 32 accumulated averages of ungated fast-scan gradient-recalled nuclear magnetic resonance (NMR) images. Following final imaging, the rats were sacrificed and the hematocrit and heart mass were determined. The mean outside diameter of the left ventricle plus interventricular septum (LV + S) showed a strong correlation (r = 0.73, P less than 0.01) with the duration of CO exposure, while the correlation coefficients for the LV + S lumen diameter and wall thickness were marginally significant. The mean pleural space diameter also increased significantly (r = 0.64, P less than 0.05) with the duration of CO exposure. The ratio of the LV + S wall thickness and the lumen radius was 0.53 in the rats exposed to CO for 0-8 days; this value did not change with longer CO exposure. The LV + S outside and lumen diameter showed significant correlations to the hematocrit (r = 0.72, p less than 0.05 and r = 0.66, P less than 0.05, respectively), and the LV + S outside diameter correlated with the increase in the LV + S mass (r = 0.72, P less than 0.05). The results achieved with NMR imaging are consistent with past morphometric studies of CO-induced ventricular hypertrophy, where heart dimensions were determined in relaxed frozen tissue, and corroborate the eccentric nature of CO-induced ventricular hypertrophy.

Lymphomas: MR imaging contrast characteristics with clinical-pathologic correlations.

Mean pixel intensity ratios, contrast parameters, and T1-T2 cross products relative to fat and muscle were derived from T1-weighted and T2-weighted images obtained at 1.5 T in 32 patients with non-Hodgkin lymphomas and 20 with Hodgkin disease. Lymphomas were relatively homogeneous: Only 6% of the lesions had broad, bimodal, or skewed distributions of pixel intensities that could be attributed to intrinsic heterogeneity. On average, lymphomas were hypointense to fat and slightly hyperintense to muscle in T1-weighted images but isointense to fat and hyperintense to muscle in T2-weighted images. Low-, intermediate-, and high-grade non-Hodgkin lymphomas had identical imaging characteristics. The most striking and significant result was the greater brightness of lymphomas with dense fibrosis in T2-weighted images, which explains the trend toward greater brightness of Hodgkin disease than that of non-Hodgkin lymphomas and greater brightness of lymphomas in the mediastinum than of lymphomas in other locations. Neither a contrast parameter nor the T1-T2 cross product added information not evident with the use of simple mean pixel intensity ratios.

Articular cartilage defects of the knee: correlation between magnetic resonance imaging and gross pathology.

Magnetic resonance imaging (MRI) of the knee articular cartilage is possible owing to the contrast provided by different signal intensities of adjacent menisci and subchondral bone. The objective of this study was to determine the accuracy of MRI in quantitatively detecting thinning and focal defects of articular cartilage in vivo. High resolution MRI was performed followed by dissection of the knee within one hour of amputations above the knee of eight patients (62-89 years) with peripheral vascular disease. Articular cartilage was examined for erosions, surface irregularities, and appearance. Mean thicknesses of femoral and tibial articular cartilage sagittal sections from MRI were statistically indistinguishable from matched gross thicknesses. In those joints in which cartilage erosions, thinning, or irregularities were detected by MRI the same defects were apparent by gross examination. Cartilage that appeared normal by MRI had a normal gross appearance by gross examination. Thus high resolution MRI can accurately predict gross articular cartilage appearance and thickness, allowing an objective, quantitative, noninvasive assessment of eroded cartilage.

Magnetic resonance imaging of meniscal degeneration in asymptomatic knees.

Histopathological studies have suggested that spontaneous degeneration of knee menisci predisposes to symptomatic tears. We used magnetic resonance (MR) imaging to study noninvasively 20 patients with documented meniscal tears in one knee but asymptomatic contralateral knees, 18 normal controls, and 15 patients with symptomatic nonmeniscal knee disorders. A scoring system for MR signal changes was developed, and differences between the three groups were tested for significance by a multivariate analysis of covariance. MR signal changes in the menisci begin at around 30 years of age, progress with age, occur in both men and women, and occur in subjects who are inactive as well as those who undergo habitual knee stress exercises. Most subjects with documented meniscal tears in one knee have MR signals in the asymptomatic contralateral knee that reflect a more advanced degree of meniscal degeneration than in age-comparable normal controls or patients with nonmeniscal knee disorders. After adjustment for potential confounding variables, weight, and sex, the mean scores in the asymptomatic knee of patients with meniscal disease are significantly higher than those of normal controls (p = 0.021) and nonmeniscal disease patients (p = 0.019). These results document the occurrence of age-dependent degeneration within knee menisci, and support the hypothesis that a segment of the population has pre-existing meniscal degeneration predisposing them to traumatic or spontaneous meniscal tears.