James M. S. Hutchison

NUCLEAR MAGNETIC RESONANCE TOMOGRAPHIC IMAGING IN LIVER DISEASE

The non-invasive diagnostic technique of whole-body nuclear magnetic resonance (NMR) imaging was evaluated in 30 patients with established liver disease and 20 patients without liver disease. Comparison with diagnostic ultrasound and radionuclide liver scan shows that NMR easily differentiates malignant tumours from benign cystic lesions and provides useful information in patients with cirrhosis and metastatic deposits. In the demonstration of space-occupying lesions in the liver, NMR is as sensitive as ultrasound and more so than radionuclide liver scans when the metastases are less than 1.5 cm in diameter. In the demonstration of cirrhosis it is more sensitive than both ultrasound and radionuclide liver scan. The specificity of NMR is superior to both ultrasound and radionuclide liver scan, both of which only demonstrate the presence of lesions, whereas NMR tomographic imaging based on the proton spin-lattice time (T1) of tissue accurately indicates the nature of the lesion.

Concomitant magnetic field gradients and their effects on imaging at low magnetic field strengths

Low-field NMR imaging systems which use large amplitude field gradient pulses (e.g., in flow velocity encoding) may be subject to the undesirable effects of concomitant gradients. We demonstrate the effects of these extra gradients, which arise from Maxwells equations, and show that the resultant image phase shifts and amplitude changes are consistent with theory.

Field-Cycled Proton-Electron Double-Resonance Imaging of Free Radicals in Large Aqueous Samples

We have recently published a new method of imaging free radicals in aqueous solutions called proton-electron double-resonance imaging (PEDRI) ( I ). In this technique a conventional proton NMR image is collected while the EPR resonance of a free radical solute is irradiated. If the EPR irradiation has sufficient power, the NMR signal from those protons being relaxed by the paramagnetic solute is enhanced, and the parts of the sample containing free radical exhibit greater intensity in the final image. Unlike EPR imaging (2) the sample size in PEDRI is not constrained by magnetic field gradient requirements. In this Communication we present the first results of an extension of PEDRI which uses magnetic field cycling, allowing larger samples to be imaged with lower levels of applied radiofrequency power. PEDRI is an imaging version of a dynamic nuclear polarization experiment (35). The enhancement of the NMR signal upon irradiation of the EPR of the solute may be written empirically as

Design, construction and use of a large-sample field-cycled PEDRI imager

The design, construction and use of a large-scale field-cycled proton-electron double-resonance imaging (FC-PEDRI) imager is described. The imager is based on a whole-body sized, vertical field, 59 mT permanent magnet. Field cycling is accomplished by the field compensation method, and uses a secondary, resistive magnet with an internal diameter of 52 cm. The magnetic field can be switched from zero to 59 mT or vice versa in 40 ms. It is used with a double-resonance coil assembly (NMR/EPR) comprising a solenoidal NMR transmit/receive coil and a coaxial, external birdcage resonator for EPR irradiation. Experiments to image the distribution of an exogenous nitroxide free radical in anaesthetized rabbits are described.

Three-dimensional NMR imaging using selective excitation

Using numerical solutions of the Bloch (1946) equations under boundary conditions representing the local external magnetic field, and with the aid of a simple model of the nuclear spin process, various effects occurring during selective excitation are described. In particular it is shown that in extending the technique for one to two dimensions there arises the possibility of spurious image response, and that this can be corrected by a suitable combination of signals from a set of three excitation sequences. Since two-dimensional selection activates a column of nuclei, the subsequent application of a lengthwise magnetic field gradient followed by Fourier transformation of the resulting spin signal will directly yield an image line. This provides a technique for full three-dimensional imaging which satisfies moderate data-processing requirements and yet retains (line by line) the advantage of Fourier methods.

A 4.2 K receiver coil and SQUID amplifier used to improve the SNR of low-field magnetic resonance images of the human arm

We describe the design and use of a 48 mm diameter, liquid-helium-cooled MRI receiver coil and DC SQUID pre-amplifier. Comparison of images of a non-conducting room temperature test object collected with the SQUID-based system and those collected with an equivalent-area room-temperature surface coil show that the SQUID system SNR is approximately a factor of four greater, despite a 15 mm vacuum gap between sample and coil in the SQUID case. SQUID images of the lower arm also display improved SNRs over those of the room-temperature coil, this time by a factor of between two and three, and as a result reveal greater anatomical detail. We show that the performance is currently limited by inductively coupled losses from metal components in the imager, but that, by using the same system in a whole-body imager, the SNR of SQUID images of the arm will exceed the room-temperature coils performance by a factor of between 2.8 and 4.5. We believe that these are the first magnetic resonance images of a living sample to have been produced with a SQUID-based receiver.

The limitations of NMR recalled-echo imaging techniques

An echo-planar pulse sequence can encode a single free induction signal with enough information to reconstruct a two- or three-dimensional image. Echo-planar pulse sequences involve the production of a train of echoes (recalled echoes) and the most easily implemented variant achieves this by repeated reversal of one gradient field. Alternate echoes are time reversed and failure to take this into account will result in aliasing. A modified pulse sequence is presented and analyzed using a simple model. The effects of transverse relaxation and magnetic field inhomogeneities are presented in terms of this model. Reversal of sample order in alternate echoes produced by the modified sequence eliminates aliasing associated with the time reversal without an increase in data collection time (or degradation of resolution). Inhomogeneities are likely to limit practically achievable resolution. It is shown that the method is intrinsically less efficient than other Fourier imaging techniques due to the finite gradient reversal time. Practical details of the implementation of the method are discussed and crude phantom images are presented.

Nuclear magnetic resonance imaging and its safety implications: follow-up of 181 patients.

Nuclear Magnetic Resonance (NMR) is a new method for imaging the human body. Its images are displayed in a manner similar to X-ray CT but there the similarity ends. Whilst X-ray CT uses X rays to produce an image of the different coefficients of absorption through a section of the body, NMR depends on the responses of hydrogen protons to a radio-frequency and does not require ionising radiation. The phenomenon of NMR has been known since 1946 (Bloch, 1946; Purcell et al, 1946) and is widely used as an analytical tool in physics and chemistry. In 1971 the concept of using the NMR technique for detecting malignant tissue was introduced (Damadian, 1971) and by 1973 NMR images of phantoms had been made (Lauterbur, 1973).

Improvements in performance time for simultaneous three-dimensional NMR imaging

Although the simultaneous collection of data from the whole of the volume of interest offers some advantages in NMR imaging experiments it is not often used in practice. This is largely because of the long imaging times involved. Two methods of reducing total imaging time for 3D imaging experiments are presented. One involves restriction of the sensitive volume to the volume of interest by selective excitation. The other involves faster information collection which is achieved by repeatedly reforming an echo following a single excitation. Images produced by the latter technique are presented and spatial resolution and signal to noise ratio for both techniques are discussed.

Human whole-body NMR tomographic imaging: normal sections

Abstract In this paper we demonstrate the kind of images that can be produced by a nuclear magnetic resonance (NMR) tomographic section imager from healthy human subjects. Details of the instrumentation and imaging techniques can be found elsewhere (Hutchison et al., 1980; Edelstein et al., 1980), but a brief description of the imaging method is given here. Our imaging machine is based on a four-coil, air-cored electromagnet manufactured by the Oxford Instrument Company. It produces a magnetic field of 0.04 T and a consequent NMR frequency of 1.7 MHz for the hydrogen protons of body tissues. Our machine uses the NMR signals from hydrogen protons in water and in fat. The machine produces two kinds of image. The first uses proton density as the imaging parameter, and the second uses proton spin-lattice relaxation time (T1). T1 in wet tissues depends essentially on the relative proportions of free and bound water within the tissue under examination; tissues with more free water have longer relaxation times (...