Magnetic resonance imaging in acute head injury

Clinical Radiology (1988) 39, 131-139

Magnetic Resonance Imaging in Acute Head Injury D. M. HADLEY, G. M. TEASDALE, A. JENKINS, B. CONDON, P. MACPHERSON, J. PATTERSON and J. O. ROWAN*

Institute of Neurological Sciences, Southern General Hospital and *Regional Department of Clinical Physics, Glasgow

Using cardiorespiratory monitoring and support equipment compatible with a low field (0.15 T) system, magnetic resonance imaging (MRI) of patients suffering acute head injuries proved to be both feasible and safe. An abnormality was demonstrated by magnetic resonance imaging in 46 of 50 patients examined within 7 days of head injury using T2 weighted (SE2200/80) and T1 weighted (IR2000/600/40) multislice sequences. In contrast, computed tomography (CT) demonstrated abnormalities in only 31 of the 50 patients. Intracranial extracerebral space-occupying collections of blood were well shown by magnetic resonance imaging which provided especially clear definition in the posterior fossa, subtemporal and subfrontal regions. Magnetic resonance imaging was more sensitive to cerebral abnormalities associated with traumatic unconsciousness and detected parenchymal lesions both in patients in coma and in those who had lost consciousness for only a few minutes. Lesions seen with MRI but not with CT included non-haemorrhagic contusions and abnormalities thought to reflect shearing injuries of white matter and intracerebral vessels. Magnetic resonance imaging is an effective alternative to CT; the additional information it can provide should be valuable in increasing the understanding of the early effects and late consequences of a head injury.

It was rapidly recognised that the evaluation of acute head injuries was revolutionised by the development of X-ray computed tomography (CT). In contrast, information on the use of magnetic resonance imaging (MRI) in the first week following a head injury is still limited 8 years after the first volunteer head images were produced (Hounsfield, 1980). Only a few individual cases have been reported (Sipponen et al., 1983; Snow et al., 1986; Zimmerman et al., 1986) and a comprehensive series has not been published. This delay reflects the problems in providing adequate cardiorespiratory monitoring and support for acutely ill patients in a strong magnetic field, and possibly also the impression that acute haemorrhagic lesions are not adequately visualised at 0.15 Tesla (DeLa Paz et al., 1984; Zimmerman et al., 1985; Gomori et al., 1985; Snow et al., 1986). We have found it feasible and safe to use magnetic resonance imaging in patients with recent head injuries and report the findings; in particular MRI of primary traumatic brain damage which we compare with computed tomography.

Address for correspondence: Dr D. M. Hadley, MRI Unit, Roundhouse, Institute of NeurologicalSciences, Southern General Hospital, Govan Road, GlasgowG51 4TF.

PATIENTS AND METHODS

Fifty patients who had sustained head injuries of sufficient severity to merit transfer from a primary hospital to the regional neurosurgical department were studied. The time between injury and transfer varied, but in each patient MRI was first performed within 7 days of injury (Table 1). Data recorded included the patient's age, the cause of their injury and its effects upon the conscious level, both immediately and as it evolved up to the time of the investigation. The presence or absence of focal neurological signs, major extracranial injuries, and the results of plain radiographs of the skull were noted. Computed tomography was carried out in all patients within 12 h of magnetic resonance imaging. Management of the patients included monitoring the blood pressure, blood gases, electrolytes, and correction of any haemodynamic or biochemical abnormalities. Seven patients underwent operation for the evacuation of a traumatic intracranial haematoma, three patients before magnetic resonance imaging and four patients afterwards.

Magnetic Resonance Imaging Technique

Patients were imaged in a Picker 'Vista 1100', 0.15 Tesla resistive magnetic resonance system operating at 6.38 MHz. A 2D Fourier transform collection mode with two averages was used. The data was acquired on a 128x256 matrix interpolated to a 256x256 display. From an initial 2 cm thick spin echo TR200/TE40 pilot image in the sagittal or coronal plane, slice positions were determined for 16 slice T2 weighted spin echo TR2200/TES0, and 8 slice T1 weighted inversion recovery TR2000/TI600/TE40 sets of axial images each 8 mm thick. Acquisition times were 20 s, 9.2 min and 8.7 min respectively. In selected cases additional sequences and orientations were used as thought appropriate by the radiologist but the examinations including computer processing time were completed in an average of 34 min. A patient monitoring system was devised to assist observation by the technician and the radiologist or anaesthetist (Fig. 1). The blood pressure and heart rate were recorded automatically at 1-min intervals with a pneumatic arm cuff 'Dynamap' (Critikon). The heart rate was shown continuously by means of an optical capillary pulse meter'S & W' (Medicoteknik A/S) fitted with an earthed and screened cable. An electrocardiogram (ECG) trace was obtained by utilising the Picker fibreoptic cardiac gating leads. Although the ECG was degraded during the fast gradient pulses, it was useful during the transfer and positioning of the patient and during computer-processing. Gases and suction reached the patient by nonferromagnetic tubing from cylinders

132 Table 1 - Time elapsed b e t w e e n imaging

CLINICAL RADIOLOGY injury and magnetic resonance

Days

No. of patients

0-1 2-3 4-7

12" 18 20

*Includes four patients imaged at less than 6 h.

transaxial sections were obtained. The examination took 20 min. If a large, surgically correctable lesion was found in a deteriorating patient, the number of slices was reduced to shorten the examination. Contrast enhancement was not performed. When necessary patients were ventilated, and E C G and end tidal CO 2 were monitored with conventional apparatus.

RESULTS Patients Studied Fifty patients were examined. Their ages ranged from 3-72 years with a mean of 31.6 years. In 21 patients the injury was due to a road traffic accident, and in 23 patients to an assault or fall related to alcohol ingestion. In 12 patients imaging was carried out on the day of the injury, and the shortest time interval was less than 3 h. At the time of the magnetic resonance investigation, the majority of patients had an altered conscious level and 12 patients were in coma. In 11, coma had been present from time of injury and in only one had there been a preceding lucid interval. Ten patients had a major focal neurological deficit characterised by a lateralised motor defect. Twenty-nine patients had a fracture of the vault of the skull demonstrated by plain radiography. Fig. 1 - An intubated and ventilated patient prepared for imaging. Heart rate, blood pressure, ECG and end tidal C02 are being monitored.

or wall units situated outside the 5 Gauss line. Mechanical ventilation via an extended Bain-Spoerel circuit was driven by a 'Penlon Nuffield ventilator' (Intermed) fitted with a 6 m tube. During the examination, end tidal CO 2 could be measured using a 'Datex Normocap' CO 2 and O 2 monitor (Instrumentation OY) fitted with an extended sampling tube. This allowed the patients' respiratory state to be observed, and also acted as a disconnection alarm. Patients were transferred from bed or stretcher with a 'Surgilift' (American Hamilton) modified to fit beneath the patient couch of the imaging machine. The head-rest and patient pallet were protected from seepage of blood and other body fluids by the use of disposable plasticlined absorbent pads. Great care was exercised in moving the patient in and out of the imager because drip tubes, drains, catheters and monitoring wires could readily be caught up in the patient loading systems. The magnetic resonance examinations were performed with the approval of the Research Ethics Committee of the Institute of Neurological Sciences, following the guidelines proposed by the National Radiological Protection Board (NRPB, 1983). Informed consent was obtained from the patient if conscious or from the next of kin if the patient was comatose or confused. Parental consent was obtained for those less than 18 years old.

Computed Tomography Technique For the first 22 patientsin the study an EMI 1010 head unit was used to obtain ten, 1 cm thick transaxial scans taking approximately 15 min. Subsequent patients were imaged in a Philips Tomoscan 310. Slice position was selected from a 'scanogram' and 14 or 15, 6 mm thick,

Magnetic Resonance Magnetic resonance imaging was abnormal in 46 patients (Table 2). Both intra and extracerebral lesions were demonstrated.

Intracerebral. Abnormalities were noted in the cortical grey matter of 44 patients, in isolation in 14 and in association with other white matter lesions in 30 patients (Table 3). Lesions in the deep white matter of the corpus callosum were noted in nine patients and in 10 patients there was a lesion in the region of the basal ganglia. A lesion was detected in the brain stem in only one patient. All patients who had sustained an extracerebral lesion (subdural haemorrhage) had evidence of intracerebral injuries (Table 4). Table 2 - Radiological findings in patients with acute head injury

Abnormality

Magnetic resonance imaging No. of patients

Computed tomography No. of patients

Cortical Subcortical white matter Deep white matter Corpus callosum Basal ganglia Brain stem Subarachnoid haemorrhage Subdural haemorrhage

44 30 18 9 10 1 11 14

23 23 2 1 2 0 4 9

Table 3 - Distribution of abnormalities associated with cortical lesions on magnetic resonance imaging

Site

No. of patients

Subeortical white matter Deep white matter Subcortical+deep white matter Deep white matter+basal ganglia Subcortical+deep white matter+basal ganglia

13 1 6 1 9

MRIINACUTEHEADINJURY Table 4 - Distribution of abnormalities associated with subdural haematomas on magnetic resonance imaging

Site Cortex Cortex+subcorticalwhite mattter Cortex+subcortical+deepwhite matter Cortex+subcortical+deepwhite matter+basalganglia

No. of patients 4 5 3 2

Extracerebral. Fourteen patients had a subdural haematoma and 11 had evidence of subarachnoid haemorrhage. In all but one patient the subdural haematoma was associated with a skull fracture. Extradural haematomas were not seen in any patients imaged in this series. Space-occupying lesions were detected in each of the three patients examined before operation. The four patients with normal magnetic resonance appearances were examined between 12 h and 3 days after injury. Computed Tomography Computed tomography was abnormal in 31 patients (Table 2). Intracerebral haemorrhagic lesions were found in 25 patients, cortical contusions in 23 and subcortical white matter lesions in 23 patients. There was a haemorrhage in the region of the basal ganglia in two patients and a lesion in the corpus callosum in only one patient. Computed tomography did not show a brain stem lesion in any patient in this series. Extracerebral lesions were noted in 11 patients, nine had a subdural haematoma and four had subarachnoid haemorrhage. There was not a significant difference (test of equal proportions P>0.3) between the proportions of patients with an abnormal examination on the EMI 1010 or Philips Tomoscan.

133

effects of acute haemorrhage. The heavily T1 weighted inversion recovery TR2000/TI600/TE40 sequence emphasised the contrast between grey and white matter, identified parenchymal lesions with increased or decreased T1 values, clearly displayed the ventricles and cisterns and was sensitive to subtle space-occupying lesions causing minor anatomical distortions. Faster spin echo sequences with a short repeat time and a short echo time (e.g. TR500/TE20) have been used by others to give T1 weighted images (Gomori el .al., 1985; Snow el al., 1986; Zimmerman el al., 1986). These produce little contrast either between grey matter and white matter or between lesions and normal brain. We use these only when images with higher resolution (256 x 256 with four averages) were appropriate. Magnetic Resonance Imaging of Traumatic Brain Damage

Superficial Abnormalities Lesions in the cerebral cortex extending to the subjacent white matter were found in 44 patients; they occurred particularly in the frontal (Fig. 2) and temporal lobes (Fig. 3a), in sites characteristic of the cortical contusions recognised at autopsy in fatal cases (Adams el al., 1986). They varied from 4-50 mm in size with larger lesions having a wide cortical base, involving several gyri (Fig. 4). Within small lesions T1 and T2 signals were uniformly prolonged (Fig. 5) whereas in larger lesions often there was a central area with T2 relaxation times ,shorter than either the surrounding 'halo' or normal brain tissues (Fig. 2, 3a). We believe that the central (short T2 moderately short T1) area reflects acute haemorrhage within a surrounding region of ischaemia and oedema. The corresponding CT appearance is of a high

DISCUSSION Feasibility of Magnetic Resonance Imaging after Head Injury

We were able to achieve safe and effective imaging of patients with recent head injuries. We made use o f equipment that is generally available and with this were able to monitor continuously respiratory and circulatory events and so avoid changes likely to exacerbate the patient's condition. Similar monitoring has been used effectively in the imaging of acutely ill patients by others (Roth et al., 1985; Weston et al., 1985; Nixon el al., 1986). Technique of Magnetic Resonance Imaging in Acute Head Injury

We found both T2 weighted spin echo sequences and T1 weighted inversion recovery sequences valuable. The spin echo TR2200/TE80 sufficiently distinguished grey matter and white matter anatomical landmarks, with cerebral spinal fluid approximately isointense with grey matter; therefore, lesions which produced oedema and prolonged relaxation times showed maximum contrast with normal structures. This sequence also detected the shortening of the T2 signal due to the susceptibility variations caused by the paramagnetic

Fig.2 - A 15-year-oldpatientin coma16h aftera 7-ftfall.T2weighted (SE2200/80) image showstypicalhyper-intensebilateral, subfrontal cortical contusions containing hypo-intensepetechial haemorrhage (arrow).

134

CLINICAL RADIOLOGY

(a)

(b) Fig. 3 - A 62-year-old man found in coma following an assault and imaged 12 h later (a) The T2 weighted (SE2200/80) image shows a well demarcated hyper-intense mid temporal contusion with a hypointense centre representing spots of haemorrhage, There is an associated thin subdural haemorrhage. (b) Computed tomography shows an ill-defined haemorrhagic contusion in mid temporal lobe (L+40, WIO0).

attenuation haemorrhage and surrounding low attenuation tissue (Fig. 3b). The site of the uniform small lesions appeared normal on CT. Such areas may reflect changes in blood flow and water content or may contain unrecognised haemorrhage. Cortical lesions of both types were seen in patients

Fig. 4 - A 20-year-old patient in coma after a road traffic accident. A T2 weighted (SE2200/80) image at 20 h showing multiple extensive hyper-intense cortical contusions: Right frontal, left temporal-occipital and right temporal.

Fig. 5 - A 31-year-old patient in coma as a result of falling down steps. A T2 weighted (SE2200/80) image at 30 h showing small paramedial gyral contusions with uniform prolongation of relaxation time. Computed tomography was normal.

imaged as soon as 3 h after injury. T h e y were extremely common in patients who had persisting depression of consciousness but were also found in a half of those who had not lost consciousness even briefly after the injury.

MRI IN ACUTE HEAD INJURY

135

These were found surprisingly frequently and occurred in locations considered to be particularly susceptible to the shearing forces of a head injury. Thus, these

lesions were in the white matter immediately subjacent to apparently normal cortex (Fig. 6a), in the central white matter (Fig. 6b), the corpus callosum (Fig. 7a, b) and in the region of the basal ganglia (Fig. 8). Some lesions appeared as a uniform zone of prolonged relaxa-

(a)

(a)

Deeper Lesions Within the Brain

(b)

t

Fig. 6 - A 19-year-old patient who had been involved in a road traffic accident, had eye opening and flexing to painful stimuli, but no verbal response. CT was normal. T2 weighted (SE2200/80) images. (a) A hyper-intense lesion in white matter is immediately subjacent to normal cortex (arrow). (b) A hyper-intense lesion in the central white matter is shown (arrow).

(b) Fig. 7 - A 20-year-old student who was injured in a motorcycle accident. CT was normal. The T2 weighted (SE2200/80) images at 24 h: (a) The axial image shows focal hyper-intense lesion in the corpus callosum (arrow). (b) The coronal image confirms the position of the corpus callosal lesion (arrow).

136

CLINICAL RADIOLOGY

Secondary Space-Occupying Lesions

Fig. 8 ~- A 12-year-old pedestrian who had been hit by a car, was in coma with no verbal response. A T2 weighted (SE2200/80) image showing hyper-intense lesion at the grey-white matter junction in the basal ganglia region.

tion time, whereas in others there was a clear inner and outer zone, with the latter having the longer relaxation time. The subcortical lesions probably correspond with the lesions termed 'gliding contusions', by pathologists (Adams et al., 1986). It is tempting to relate the lesions in the central white matter and corpus callosum to the shearing lesions of the white matter that occur in severe primary injury (Adams et al., 1982). Certainly some of these lesions occurred at sites in which haemorrhagic lesions can occasionally be seen with CT and where tears are found at post-mortem. Such CT and post-mortem appearances are found almost exclusively in patients who have sustained immediate deep and persisting coma. This was not so with all patients with such lesions on MRI. Magnetic resonance imaging therefore, appears to be able to detect lesions associated with lesser degrees of axonal injury than CT or macroscopic postmortem examination. The MRI finding of lesions in patients with short periods of unconsciousness is consistent with recent post-mortem evidence. Pilz (1983) has shown that axohal injury can be produced by relatively mild injury, with histological abnormalities after unconsciousness of as short a duration as 5 min. It is also possible that some of the parenchymal lesions result from damage to perforating vessels rather than the neural tissue (Macpherson et al., 1986). Further studies using susceptibility mapping, modified partial saturation sequences (Bydder et al., 1986; Edelman et al., 1986) and the contrast agent gadolinium DTPA may clarify the nature of the different patterns of these deep and cortical lesions seen after head injury.

During this pilot study patients requiring urgent surgery usually underwent CT alone. The proportion of patients studied by MRI who had a surgically significant space-occupying lesion (16%) was therefore less than the overall incidence in our Institute (25%). Nevertheless, it is clear that MRI is an effective method of detecting significant traumatic intracranial space-occupying lesions. In the first week after injury traumatic intracerebral haematoma was characterised by a hypo-intense region on the spin echo image surrounded by a hyper-intense periphery (Fig. 9a). In inversion recovery images taken from the first to the fourth day (Fig. 9b), the haematoma showed a central area iso-intense with grey matter surrounded by a hypo-intense rim; thereafter, increasing hyper-intensity developed either in the whole area of haemorrhage or in the peripheral ring. On the third and fourth day, the area of haemorrhage usually gave a hyper-intense signal compared with grey matter; by the fifth to seventh days (Fig. 9c) it was hyper-intense even compared with white matter. These appearances in vivo reflect the interaction of several factors: the intra- and extra-cellular occurrence of at least four paramagnetic species: deoxyhaemoglobin, methaemoglobin, free Fe z+ and haemosiderin. There are also changes as a result of the changing consistency of the lesion with clot formation and its subsequent liquefaction (Gomori et al., 1985; Bydder et al., 1986). These findings in patients are in contrast with studies in vitro of haemorrhage (Jenkins et al., 1985), which showed recently clotted blood to have longer relaxation times than either white matter or grey matter, with an iso-intense signal only after three to five days. Coronal MRI identified developing herniation in patients with significant haematomas. In one patient this was associated with a focal prolongation of relaxation time in the medial part of the displaced temporal lobe, suggestive of early temporal lobe oedema, possibly caused by ischaemia (Fig. 10). In another patient with a large posterior fossa haematoma, the magnetic resonance images showed displacement of the fourth ventricle and marked dilatation of the lateral ventricles with prolonged periventricular relaxation times suggestive of acute hydrocephalus. By contrast it is not clear if MRI can be used to indicate raised intracranial pressure in patients without a space-occupying lesion. Seven patients in this study had an intracranial pressure over 20 mmHg at the time of the scan but in each the signal corresponding to the cerebral spinal fluid in the third ventricle could be detected on MRI. In such patients CT would be expected to show an obliteration of the third ventricle or basal cisterns (Teasdale et al., 1984).

Comparison of Computed Tomography and Magnetic Resonance Imaging Magnetic resonance imaging was clearly more sensitive than CT in detecting traumatic brain damage. The greater number of cortical lesions displayed reflects both the greater relative change in magnetic resonance relaxation times compared with X-ray attenuation (Chakeres and Bryan, 1986) and the lack of artefact from adjacent bone. There was an even more striking

137

MRI IN ACUTE HEAD INJURY

(a)

(c)

Fig. 9 - A 16-year-old girl who had been in coma for 4 h following a fall from her horse: (a) 10 h post trauma, a T2 weighted (SE2200/80) image shows hypo-intense bifrontal haematomas surrounded by hyper-intense oedema. (b) 30 h post trauma, a T1 weighted (IR2000/600/40) image shows the haemorrhage iso-intense with grey matter surrounded by hypo-intense oedema. (c) 7 days post trauma, the T1 weighted (IR2000/600/40) image shows the haemorrhage hyper-intense compared with grey and white matter. (b)

discrepancy between CT and MRI in the detection of parenchymal lesions. MRI showed many more lesions than CT and detected lesions even in patients with relatively brief and mild impairment of consciousness. The additional lesions demonstrated by MRI are also likely to be due to shearing, but in some patients they may be a consequence of forces with a predilection for vascular rather than axonal injury.

Computed Tomography or Magnetic Resonance for Head Injury Imaging? Computed tomography has improved the management of head injuries (Teasdale et al., 1982) and when available will remain the preferred investigation. If CT is not available MRI can be effective in patients with acute head injuries and was used as the first investiga-

138

CLINICAL RADIOLOGY

However, diagnostic images can be obtained without the practical management problems associated with high field strengths. Further work is needed to clarify the nature of the parenchymal lesions demonstrated by MRI and to determine how the additional information it provides contributes to the management of patients. Magnetic resonance imaging may provide much new information about the factors responsible for the sequelae of head injury. Studies of the evolution of lesions demonstrated by MRI and of the findings in disabled survivors are needed. Acknowledgements. The Glasgow Magnetic Resonance Imaging Unit was established with support from the Medical Research Council, the Scottish Home and Health Department, the Scottish Hospital Endowment Research Trust, the Greater Glasgow Health Board and the University of Glasgow. REFERENCES

Fig. 1 0 - A 21-year-old steel erector fell 30 ft. He was in coma, ventilated, and had an unreacting left pupil. A T2 weighted (SE2200/80) image at 72 h shows a large temporal lobe hypo-intense haemorrhagic contusion surrounded by hyper-intense oedema causing a mass effect with mid-line shift and a focal hyper-intense crescentic region in the medial aspect of the opposite temporal lobe (arrow) indicating tentorial herniation.

tion in three patients in this series. When compared with the overall time needed for resuscitation, transport and transfer of such patients, the difference in examination time between CT and MRI is small. Currently MRI more often requires sedation or anaesthesia but the 'pilot' facility of MRI can identify major space-occupying lesions within seconds and multislice acquisition techniques permit a comprehensive examination within minutes. Faster examination times using low flip angles are likely to be available on commercial imagers in the future. Although high field imaging (1 to 2 Tesla) would produce an inherently stronger signal without averaging and hence faster scans, the susceptibility and T2 shortening effects of acute haemorrhage described by Gomori et al. (1985) can also be demonstrated at middle range (0.5 Tesla) (Edelman et al., 1986) and in our study at low (0.15 Tesla) field strengths (Figs 2, 9a, 10). In trauma cases many of the supposed advantages of high field imaging are outweighed by the practical difficulties in managing the critically-ill patient within the magnet. These can be largely overcome in low field imagers with minor modifications to existing monitoring and support equipment. At ultra low (0.02 Tesla) field strengths the further advantage of increasing T1 contrast, allowing acute haematomas to be differentiated from normal brain or neoplasms (Sepponen et al., 1987), is partly offset by the increased time required to collect the data for images of comparable quality to high field systems.

Adams, JH, Graham, DI, Murray, LS & Scott, G (1982). Diffuse axonal injury due to non-missile head injury in hi, roans. An analysis of 45 cases. Annals of Neurology, 12, 557-563. Adams, JH, Doyle, D, Graham, DI, Lawrence, AE & McLellan, DR (1986). Gliding contusions in non-missile head injury in humans. Archives of Pathology and Laboratory Medicine, 110, 485-488. Bydder, GM, Pennock, JM & Young, IR (1986). Magnetic resonance imaging of intracerebral haematoma at low field (0.15T). Proceed-

ings of the European Society of" Magnetic Resonance in Medicine and Biology, 3, 30-31. Chakeres, DW & Bryan, RN (1986). Acute subarachnoid haemorrhage: In vitro comparison of magnetic resonance and computed tomography. American Journal of Neuroradiology, 7, 223-228. DeLa Paz, RL, New, PFJ, Buonanno, FS, Kistler, JP, Oot, RF, Rosen, BR et al. (1984). NMR imaging of intracranial haemorrhage. Journal of Computer Assisted Tomography, 8, 599-607. Edelman, RR, Johnson, K, Buxton, R, Shoukimas, G, Rosen, BR, Davis, K et al. (1986). MR of haemorrhage: a new approach. American Journal of Neuroradiology, 7, 751-756. Gomori, JM, Grossman, RI, Goldberg, HI, Zimmerman, RA & Bilaniuk, LT (1985). Intracranial haematomas: imaging by highfield MR. Radiology, 157, 87-93. Hounsfield, GN (1980). Computed medical imaging. Journal of Computer Assisted Tomography, 4, 665-674. Jenkins, A, Condon, B, Patterson, J, Hadley, DM & Teasdale, GM (1985). Acute intracerebral haemorrhage: use of in vitro MR studies to guide clinical imaging. Proceedings of the European Society of Magnetic Resonance in Medicine and Biology, 2, 278282. Macpherson, P, Teasdale, E, Dhaker, S, Allerdyce, G & Galbraith, S (1986), The significance of traumatic haematoma in the region of the basal ganglia. Journal of Neurology, Neurosurgery and Psychiatry, 49, 29-34. Nixon, C, Hirsch, NP, Ormerod, IEC & Johnson, G (1986). Nuclear magnetic resonance its implications for the anaesthetist. Anaesthesia, 41, 131-1.37. National Radiological Protection Board (NRPB, 1983). Revised guidance on acceptable limits of exposure during nuclear magnetic resonance clinical imaging. British Journal of Radiology, 56, 974977. Pilz, P (1983). Axonal injury in head injury. Acta Neurochirurgica, Supplement 32, 119-123. Roth, JL, Nugent, M, Gray, JE, Julsrud, PR, Berquist, TH, Sill, JC et al. (1985). Patient monitoring during magnetic resonance imaging. Anaesthesiology, 62, 80-83. Sepponen, RE, Tanttu, JI, Souranta, H & Suramo, I (1987). Visualization of intracerebral haemotomas with T1, T2, T1.p and Tlp dispersion imaging at 0.02T. Magnetic Resonance Imaging, 5, Supp 1, 85-86. Sipponen, JT, Sepponen, RE & Sivula, A (1983). Nuclear magnetic resonance (NMR) imaging of intracerebral haemorrhage in acute and resolving phases. Journal of Computer Assisted Tomography, 7, 954-959. Snow, RB, Zimmerman, RD, Gandy, SE & Deck, DF (1986). Com-

MRI IN ACUTE HEAD INJURY parison of magnetic resonance imaging and computed tomography in the evaluation of head injury. Neurosurgery, 18, 45-52. Teasdale, E, Cardoso, E, Galbraith, S & Teasdale, G (1984). CT scan in severe diffuse head injury: physiological and clinical correlations. Journal of Neurology, Neurosurgery and Psychiatry, 47,600603. Teasdale, G, Galbraith, S, Murray, L, Ward, P, Gentleman, D & McKean, M (1982). Managment of traumatic intracranial haematoma. British Medical Journal, 285, 1695-1697. Weston, G, Strunin, L & Amundson, GM (1985). Imaging for

139

anaesthetists: a review of the methods and anaesthetic implications of diagnostic imaging techniques. Canadian Anaesthetists' Society Journal, 32, 552-561. Zimmerman, RA, Bilaniuk, LT, Grossman, RI, Levine, RS, Lynch, R, Goldberg, HI et al. (1985). Resistive NMR of intracranial haematomas. Neuroradiology, 27, 16-20. Zimmerman, RA, Bilaniuk, LT, Hackney, DB, Goldberg, HI & Grossman, RI (1986). Head injury: early results of comparing CT and high-field MR. American Journal of Neuroradiology, 1,757764.

Book Reviews Imaging in Hepatobiliary Disease. By J. Dooley, R. Dick, M. Viamonte Jr and S. Sherlock. Blackwell Scientific, Oxford 1987, 245 pp.

Computed Tomography in Trauma. By Barry D. Toombs and Carl M. Sandier. W. B. Saunders, Philadelphia, 1987. 224 pp. 250 figs.

£49.50. This is a very good new practical book on imaging in hepatobiliary disease. As the authors state in their preface, it is essentially an atlas with guidelines for the use of new current techniques. The quality of the illustrations is excellent throughout with clear annotations. The book is divided into two parts. In the longer first part, common and rare conditions are described and various authors discuss the current 'state of the art' including the controversial issues as to which technique is the best. This is difficult as the scene is constantly shifting. They give clear guidelines as to which diagnostic techniques are most effective and least costly (a most important consideration today). The diagnostic chapters each consider different clinical aspects of liver and biliary disease such as jaundice in adults, jaundice in the paediatric age patient, palpable liver in adults, liver masses in children, fever associated with hepatobiliary disease and portal hypertension as well as liver trauma and a palpable spleen. The smaller second part of the book covers interventional procedures used in hepatobiliary disease. The subjects covered include the different types of liver biopsy, aspiration cytology, drainage of abscesses and cysts. Percutaneous trans-hepatic techniques such as sclerosis of varices, embolisation of the hepatic artery and drainage of the biliary tract with insertion of stents are covered. There is a short chapter on the endoscopic approach to the biliary tract and another on the removal of retained scaling. The techniques are very clearly described including the indications, possible complications and materials needed. I can thoroughly recommend this very practical book on hepatobiliary disease which has come mainly from the Royal Free Hospital with contributions particularly from North America. It should be found on the bookshelves of any Radiology Department where such procedures are carried out. H. B. Nunnerley

In the UK trauma accounts for more deaths than cancer and at the recent meeting of the British Association for Science, it was reported that approximately a quarter of the deaths from trauma resulted from improper and delayed diagnoses and the lack of a properly organised trauma service including the availability of computed tomography (CT). This excellent short textbook clearly defines the role of CT in trauma. The author's experience is drawn from the Trauma Surgery Service at the University of Texas Medical School with 1300 admissions in 1984, many of whom had multisystem injuries. The book begins with chapters on the mechanism and epidemiology of trauma and the initial evaluation of trauma patients. Subsequent separate chapters discuss acute thoracic, abdominal and pelvic trauma and their complications where the important role of CT is emphasised; at the same time there is a useful and practical correlation with other established imaging modalities. High resolution CT of spinal trauma is particularly well covered as is eranio-facial trauma supplemented by 3-dimensional reconstruction images. Apart from the cranio-facial region, skull trauma is largely omitted; this is not to underestimate the role of CT, as the authors point out, but to emphasise its importance in other areas. The book is extremely well written, easy to read with a large number of excellent figures, it is well presented and good value for money. CT scanning has a major impact on the care of patients suffering from multisystem trauma. In the UK, not only are we short of CT scanners but the services for major trauma require urgent review. Hopefully, books such as this will provide the necessary stimulus. It is highly recommended for radiologists and surgeons involved with major trauma. J. K. Davidson