Carsten Thomsen

MR diffusion imaging of human intracranial tumours

Abstract We used MRI for in vivo measurement of brain water self-diffusion in patients with intracranial tumours. The study included 28 patients (12 with high-grade and 3 with low-grade gliomas, 7 with metastases, 5 with meningiomas and 1 with a cerebral abscess). Apparent diffusion coefficients (ADC) were calculated in a single axial slice through the tumours; the sequence was sensitive to diffusion along the cephalocaudal axis. Our main finding was that ADC in contrast-enhancing areas within cerebral metastases was statistically significantly higher than ADC in contrast-enhancing areas in high-grade gliomas (P≤ 0.05). Furthermore, the ADC in oedema surrounding metastases were statistically significantly higher the ADC in oedema around high-grade gliomas (P≤ 0.02). The ADC in patients with meningiomas did not differ significantly from those seen with high-grade gliomas or cerebral metastases. The highest ADC were found within cystic or necrotic tumour areas. In one patient with a cerebral abscess, suspected of having a high-grade glioma, the ADC was similar to that in high-grade gliomas. The finding of higher ADC in cerebral metastases than in high-grade gliomas may be helpful in trying to distinguish between these tumours preoperatively; it suggests increased free extracellular and/or intracellular water fraction in cerebral metastases. The method seems to hold potential for further noninvasive characterisation of intracranial tumours.

Pulsatile brain movement and associated hydrodynamics studied by magnetic resonance phase imaging

SummaryBrain tissue movements were studied in axial, sagittal and coronal planes in 15 healthy volunteers, using a gated spin echo MRI sequence. All movements had characteristics different from those of perfusion and diffusion. The highest velocities occurred during systole in the basal ganglia (maximum 1.0 mm/s) and brain stem (maximum 1.5 mm/s). The movements were directed caudally, medially and posteriorly in the basal ganglia, and caudally-anteriorly in the pons. Caudad and anterior motion increased towards the foramen magnum and towards the midline. The resultant movement occurred in a funnelshaped fashion as if the brain were pulled by the spinal cord. This may be explained by venting of brain and cerebrospinal fluid (CSF) through the tentorial notch and foramen magnum. The intracranial volume is assumed to be always constant by the Monro-Kellie doctrine. The intracranial dynamics can be viewed as an interplay between the spatial requirements of four main components: arterial blood, capillary blood (brain volume), venous blood and CSF. These components could be characterized, and the expansion of the arteries and the brain differentiated, by applying the Monro-Kellie doctrine to every moment of the cardiac cycle. The arterial expansion causes a remoulding of the brain that enables its piston-like action. The arterial expansion creates the prerequisites for the expansion of the brain by venting CSF to the spinal canal. The expansion of the brain is, in turn, responsible for compression of the ventricular system and hence for the intraventricular flow of CSF.

Valve area and cardiac output in aortic stenosis: Quantification by magnetic resonance velocity mapping☆

Valve area and cardiac output were determined with magnetic resonance (MR) velocity mapping in 12 patients with aortic stenosis. Heart catheterization, Doppler echocardiography, and indicator dilution were performed for comparison. Left ventricle could be catheterized in only nine patients; in these cases, MR measured a mean valve area of 1.2 cm2 compared with 0.9 cm2 derived from catheterization data, with a mean difference of 0.2 cm2 between the 2 methods. The limits of agreement were [0.0, +0.5] cm2, less in patients with an important degree of concomitant regurgitation. In the whole material, MR measured a mean area of 1.1 cm2 compared with 1.2 cm2 derived from Doppler echocardiography data, with a mean difference of 0.1 cm2 and [-0.5, +0.6] cm2 as limits of agreement. In 11 patients the cardiac output was quantified by MR to a mean of 4.9 L/min and by indicator dilution to 5.0 L/min, with a mean difference of 0.2 L/min, and [-0.6, +0.8] L/min as limits of agreement. In addition, MR offers the major advance of simultaneous quantification of regurgitant volume in cases of concomitant regurgitation. In conclusion, because the two important prognostic determinants in aortic stenosis--the valvular area and the cardiac output--may be quantified, MR has potential to become a clinical tool in assessment of severity in aortic stenosis.

A method for MR quantification of flow velocities in blood and CSF using interleaved gradient-echo pulse sequences

The aim of this study was to establish a rapid method for in vivo quantification of a large range of flow velocities using phase information. A basic gradient-echo sequence was constructed, in which flow was encoded along the slice selection direction by variation of the amplitude of a bipolar gradient without changes in sequence timings. The influence of field inhomogeneities and eddy currents was studied in a 1.5 T interleaved sequences for calibration and in vivo flow determination were constructed, and flow information was obtained by pairwise subtraction of velocity-encoded from velocity non-encoded phase images. Calibration was performed in a nongated mode using flow phantoms, and the results were compared with theoretically calculated encoding efficiencies. In vivo flow was studied in healthy volunteers in three different areas using cardiac gating; central blood flow in the great thoracic vessels, peripheral blood flow in the popliteal vessels, and flow of cerebrospinal fluid (CSF) in the cerebral aqueduct. The results show good agreement with results obtained with other techniques. The proposed method for flow determination was shown to be rapid and flexible, and we thus conclude that it seems well suited for routine clinical MR examinations.

24-Hour lithium concentration in human brain studied by Li-7 magnetic resonance spectroscopy

Brain and serum lithium concentrations were measured every second hour during a 24-hr period following lithium intake, and again 48-hr later in two normal subjects in steady state lithium treatment receiving lithium carbonate (Priadel Synthelabo) once every evening. The brain-lithium concentration was measured by 7Li magnetic resonance spectroscopy (MRS). The brain lithium level was found to undulate in a peak-trough pattern that followed the serum lithium profile, although in an attenuated form. The brain/serum lithium concentration ratio varied considerably during the 48-hr period, ranging from 0.5 to 1.3, but the ratio was independent of the serum-lithium concentration. The median half-life for lithium was 28 hr in the brain, and 16 hr in serum. The brain lithium concentration in the morning was about 75% of the clinically relevant standard 12-hr serum lithium concentration. The finding that brain lithium undulates during the day means that MRS measurements of brain lithium can only be compared if carried out under standard conditions that include a fixed interval following lithium intake and an identical treatment regimen.

Assessment of CSF dynamics and venous flow in the superior sagittal sinus by MRI in idiopathic intracranial hypertension: a preliminary study.

A velocity-sensitive magnetic resonance imaging (MRI) phase-mapping method was used for noninvasive study of cerebrospinal fluid (CSF) flow in the cerebral aqueduct, for indirect calculation of supra-tentorial CSF production, and for measurement of blood flow in the superoor sagittal sinus (SSS). We examined 12 patients with idiopathic intracranial hypertension (HH; pseudotumour cerebri), and 10 healthy volunteers. The peak caudal and rostal CSF flow in the aqueduct during the cardiac cycle did not differ significantly between the patients and the volunteers. A significant correlation was found between the CSF volume flow amplitude and the resistance to cerebrospinal fluid out-flow in the patients (p<0.05). The calculated mean supratentorial CSF production rate was 0.79 ml/min in the patients and 0.70 ml/min in the controls, but this difference was not statistically significant. However, the MRI measurements suggested CSF hypersecretion in three patients, whereas increased transependymal passage of CSF could have been the cause of negative calculated CSF production rates in two others. A tendency towards lower mean blood flow in the SSS (mean 345 ml/min) in the patients than in the controls (mean 457 ml/min) was found, and in two patients showed very low values. We showed that MRI phase-mapping may be used to study the relative importance of the pathophysiological factors thought to play a role in the development of IIH.

Improved energy kinetics following high protein diet in McArdle's syndrome. A 31P magnetic resonance spectroscopy study

A patient with McArdles syndrome was examined using bicycle ergometry and 31P NMR spectroscopy during exercise. The patients working capacity was approximately half the expected capacity of controls. Muscle energy kinetics improved significantly during intravenous glucose infusion and after 6 weeks of high protein diet. During intravenous infusion of amino acids, no changes in working capacity could be detected. No decrease was seen in intracellular muscle pH during aerobic exercise. A significant decrease in muscle pH during aerobic exercise was detected in all controls.

In vivo measurements of T1 relaxation times of 31P-metabolites in human skeletal muscle

The T1 relaxation times were estimated for 31P-metabolites in human skeletal muscle. Five healthy volunteers were examined in a 1.5 Tesla wholebody imaging system using an inversion recovery pulse sequence. The calculated T1 relaxation times ranged from 5.517 sec for phosphocreatine to 3.603 sec for -ATP.

Twelve-hour brain lithium concentration in lithium maintenance treatment of manic-depressive disorder: daily versus alternate-day dosing schedule.

The 12-h brain lithium concentration was measured by lithium-7 magnetic resonance spectroscopy in ten manic-depressive patients receiving daily or alternate-day lithium carbonate treatment. The median dose of lithium carbonate was 800 mg in the daily treatment group and 1200 mg in the alternate-day group. Median 12-h serum lithium concentration in the two groups was 0.86 mmoll−1 and 0.55 mmoll−1, respectively, while the corresponding concentration in brain was 0.67 mmoll−1 and 0.52 mmoll−1, respectively. The 12-h brain lithium concentration was independent of lithium dosing schedule (multiple linear regression), but correlated significantly with the 12-h serum lithium concentration (P=0.003; B=0.53, 95% c.l. 0.24–0.82; β=0.83). Thus at identical 12-h serum lithium concentrations the 12-h brain lithium concentration is similar with both treatment regimes. As the risk of manic-depressive relapse during alternate-day lithium treatment is in our experience 3-fold greater than with daily treatment (at similar mean 12-h serum lithium concentration), the findings suggest that the difference in the prophylactic efficacy of the two dosing schedules is unrelated to differences in the 12-h brain lithium concentration.

Use of an enhanced gradient system for diffusion MR imaging with motion-artifact reduction

Purpose: A spin-echo diffusion-sensitized pulse sequence using high gradients (23 mT/m) is introduced. Material and Methods: In order to minimize motion artefacts, velocity-compensating gradients, ECG-triggering and post-processing with phase correction and raw data averaging using navigator echoes was performed. The in vitro ratio of diffusion coefficients for water and acetone was determined and the water self-diffusion coefficient at different temperatures was evaluated. The pulse sequence was tested in 7 healthy volunteers and in 2 tumour patients with astrocytomas of grades I—II and III—IV. Both single-slice and multi-slice techniques were used. Results: The incorporation of phase correction clearly improved the quality of both diffusion-encoded images and the calculated diffusion maps. Mean values of the diffusion coefficients in vivo were for CSF 2.66×10−9 m2/s and for white and grey matter 0.69×10−9 m2/s and 0.87×10−9 m2/s, respectively. Conclusion: Velocity-compensating gradients in combination with a high gradient strength were shown to be useful for in vivo diffusion MR imaging.