D. S. Hickey

MR imaging of the intervertebral disc: a quantitative study

The T1 and T2 relaxation times and the proton density of the nucleus pulposus have been measured in 107 normal and 18 surgically proven degenerate intervertebral discs. Data from total saturation recovery and spin echo sequences have been utilised in a robust multi-point method and relaxation times and proton density calculated. The results show that both the T1 and T2 values of the normal nucleus pulposus decrease with age. There was no significant correlation between proton density and age in normal discs. At all ages there was a highly significant difference between the T1 values of normal and degenerate discs. With T2 a highly significant difference in the younger age groups reduced to no distinction in the seventh decade. The observed change in the T1 and T2 values of the nucleus is in agreement with the reduction of water content known to occur with age. Our results indicate that quantitative MR imaging may assist in the diagnosis of intervertebral disc degeneration.

Analysis of magnetic resonance images from normal and degenerate lumbar intervertebral discs.

Degenerate discs can be identified quantitatively by measurement of magnetic resonance (MR) relaxation times. MR images have been recorded from 16-year-old and 82-year-old cadaveric L3-4 discs at the highest resolution attainable with a Picker International MR Imaging System operating at 0.26 Tesla. By recording images with a series of spin-echo and/or saturation-recovery sequences of differing time intervals, the values for sample magnetization, M∞ and the T1 and T2 relaxation times, have been calculated from each pixel in the MR image. The distribution of M∞ values shows the relative degrees of hydration in different regions of the disc while the corresponding T1 and T2 values are sensitive to the chemical environment of the water molecules. Images from cadaveric discs allowed the nucleus pulposus and annulus fibrosus to be distinguished clearly, and the laminated structure of the annulus could be seen. Loss of water from the nucleus during aging was demonstrated by a reduction and change in the distribution of the M∞ values for an 82-year-old disc, as compared with a 16-year-old disc. Values of T1 and T2 indicated a difference in the chemical environment of water molecules in the nucleus pulposus and annulus fibrosus; the extent of this difference was much greater for younger than for older discs. High-resolution MR images from discs of living subjects showed almost as much detail as those from experimental specimens, but in the latter, the laminated structure of the annulus was resolved. Distributions of M∞, T1 and T2 values were calculated and the results were comparable with those from cadaveric discs. Confirmed at surgery, degenerate discs in patients showed greatly reduced values and modified distributions of M∞ T1 and T2 for the nucleus pulposus, corresponding to a loss of water molecules and a change in their chemical environment. Differences between normal and degenerate discs represent an extreme form of the changes that occur during aging.

A method for the clinical measurement of relaxation times in magnetic resonance imaging

A method for the determination of relaxation times in clinical magnetic resonance images is described. Three components are measured: the spin-lattice (T1) and spin-spin (T2) relaxation times and the proton density (M infinity). These components are separated in the algorithm to give increased tissue discrimination. Multiple data points are used to minimise error and increase reproducibility. Errors that arise in imaging data because of the short sequence repetition periods are considered and a technique for their reduction described. Clinical results obtained using the method are reviewed. These results demonstrate the clinical utility of the technique.

Quantitative tissue characterisation in pancreatic disease using magnetic resonance imaging

Twenty-nine patients, 27 of whom had either inflammatory disease of the pancreas or pancreatic tumour, were studied by magnetic resonance imaging (MRI) and computed tomography (CT). Six healthy volunteers were studied by MRI alone. The pancreatic T1 and T2 relaxation times were calculated using a multipoint iterative method with data from seven total saturation recovery and six spin echo sequences. Magnetic resonance imaging can demonstrate the normal pancreas and a variety of pathological processes greater than 1-2 cm in size, but with less spatial resolution than CT. The relaxation-time results indicated no significant discrimination between chronic pancreatitis and pancreatic tumour. A significant elevation in the relaxation times was observed, however, in those patients with calcific chronic pancreatitis compared with the non-calcific chronic pancreatitic group and normal controls, suggesting a different pathophysiology for the two subgroups of chronic pancreatitis. The active phase of acute pancreatitis was associated with significantly elevated relaxation times, which returned to normal levels during the resolved phase of the disease. Associated extrapancreatic fluid collections were characterised by their very long relaxation times. The problems associated with spatial resolution, respiratory motion and lack of quantitative tissue characterisation suggest that MRI of the pancreas, using present methods, is unlikely to contribute to the overall management of patients with exocrine pancreatic disease.

Thorecolumlbar fascia can increase the efficiency of the erector spinae muscles

A model is developed which demonstrates that the thoracolumbar fascia can increase the axial stress generated by active contraction of the erector spinae muscles by restricting their radial expansion (bulging). The radial stress in the muscles was calculated from their radii, the radius of the abdominal cavity, and the intra-abdominal pressure generated during lifting. Radii were measured from magnetic resonance images and values of intra-abdominal pressure obtained from the literature. The resulting radial stress was used to calculate the increase in axial stress developed as a result of restricting radial expansion. Values for the axial stress generated by contraction of an unrestricted muscle, which are also required for this calculation, were obtained from the literature. The results indicate that restriction of radial expansion of the erector spinae muscles by the thoracolumbar fascia may increase the stress generated during their contraction by up to about 30%. This would lead to a proportional increase in the extensor moment exerted by these muscles.

DETERMINATION OF COLLAGEN FIBRIL ORIENTATION IN THE CARTILAGE OF VERTEBRAL END PLATE

X-Ray diffraction and scanning electron microscopy show that collagen fibrils tend to be oriented parallel to the surface of the vertebral end plates in pig, rat and man. It is suggested that this orientation allows the end plate to withstand the pressure exerted on it by the nucleus pulposus. Scanning electron microscopy shows that lamellae occur in human end plates. Collagen fibrils also tend to be preferentially aligned in the direction of the ventral-dorsal axis.

Increased potential for tissue discrimination in quantitative magnetic resonance imaging

A method has been developed that substantially increases quantitative tissue discrimination of pathology in in-vivo magnetic resonance (MR) imaging. Results from a study of carcinoma and benign hypertrophy of the prostate show the potential for classification of tissues from individual patients and demonstrate increased class separation. This appears to be the firstapplication of pattern recognition techniques to relaxation-time data from in-vivo MR images. Multiple-point relaxation-time maps were classified using both statistical measures and texture descriptors derived from spatial grey level dependency matrices. These parameters were combined in a new pattern-recognition method based upon the Karhunen- Loeve expansion.

Computer simulation of the slice profile in magnetic-resonance imaging

A computer simulation has been developed of the slice profile generated by the most commonly used pulse sequences in clinical imaging. This simulation enables a full description of the magnetisation vector to be made at any position across the slice at any time, both during and after the application of a pulse, to any desired accuracy. Presented here is an analysis of the most common pulse sequences: free-induction decay, inversion recovery, spin echo, saturation recovery and Carr-Purcell. The authors show preliminary results of the variation of slice profile with the key interpulse decay times for the first four of these sequences. Saturation recovery is shown to be the best choice of sequence for multiple-point T1 measurement as it presents the best compromise between slice profile artefacts and imaging time.

Computer methods for high resolution relaxation time imaging

Methods for calculating multiple point T1 relaxation images are described and compared. A robust line fitting method is presented and its relevance to high resolution relaxation imaging discussed. Selective removal of data points is demonstrated to reduce systematic errors in linear models. A non-linear least squares iterative method is also developed and implemented. The results include simulated data, clinical images and a phantom study. Several of the methods are substantial improvements on existing linear techniques. The increased quality and consistency of the calculated images make them particularly appropriate for automated pattern recognition.

Accuracy and precision in the measurement of relaxation times from nuclear magnetic resonance images.

We refer to the paper by Johnson et al (1987) which describes a method based on the minimum number (three) of scan images for calculation. The authors state that “A multiple scan examination is impractical with patients because it can take several hours”. Previously published data in this journal (Jenkins et al, 1985; Hickey et al, 1986) and elsewhere have drawn attention to the need for a rigorous multiple-point approach to obtain precision and reproducibility of both T1 and T2 relaxation times in the realistic clinical situation, where the character of diseased tissues is unpredictable.