A proposed model of cerebrospinal fluid circulation: observations with radionuclide cisternography.

PURPOSE: To determine the mechanisms of the tracer distribution at radionuclide cisternography (RC). METHODS: Ten patients with venous vasculitis were studied with RC. Flow phantom studies were performed mimicking cerebrospinal fluid (CSF) circulation with and without a main outlet comparable to the pacchionian granulations. RESULTS: Nine of the 10 patients had normal findings on RC images, including a maximum uptake over the vertex at 24 hours. In all patients, a second maximum occurred in the lumbosacral area. The flow phantom studies showed no tracer accumulation at an open outlet corresponding to the pacchionian granulations. On the contrary, a maximum arose without such an outlet. A maximum always arose at the closed dead ends of the phantom, including the lumbosacral area. CONCLUSION: The commonly accepted flow model for CSF circulation needs to be revised. The pattern of the normal RC cannot be explained by a bulk flow transport of the tracer to an outlet at the pacchionian granulations but rather by a primary mixing caused by pulsatile flow with a secondary dilution by newly formed CSF from the ventricular system. We suggest that the main absorption of the CSF is through the central nervous system to the blood.

A spinal thecal sac constriction model supports the theory that induced pressure gradients in the cord cause edema and cyst formation.

OBJECTIVE: Spinal cord cysts are a devastating condition that occur secondary to obstructions of the spinal canal, which may be caused by congenital malformations, trauma, spinal canal stenosis, tumors, meningitis, or arachnoiditis. A hypothesis that could explain how spinal cord cysts form in these situations has been presented recently. Therefore, a novel spinal thecal sac constriction model was implemented to test various aspects of this hypothesis. METHODS: Thecal sac constriction was achieved by subjecting rats to an extradural silk ligature at the T8 spinal cord level. Rats with complete spinal cord transection served as a second model for comparison. The animals underwent high-resolution magnetic resonance imaging and histological analysis. RESULTS: Thecal sac constriction caused edema cranial and caudal to the ligation within 3 weeks, and cysts developed after 8 to 13 weeks. In contrast, cysts in rats with spinal cord transection were located predominantly in the cranial spinal cord. Histological sections of spinal cords confirmed the magnetic resonance imaging results. CONCLUSION: Magnetic resonance imaging provided the specific advantage of enabling characterization of events as they occurred repeatedly over time in the spinal cords of individual living animals. The spinal thecal sac constriction model proved useful for investigation of features of the cerebrospinal fluid pulse pressure theory. Edema and cyst distributions were in accordance with this theory. We conclude that induced intramedullary pressure gradients originating from the cerebrospinal fluid pulse pressure may underlie cyst formation in the vicinity of spinal canal obstructions and that cysts are preceded by edema.

A theoretical study of amplitude modulation and time shifting in quantitative MR measurements of motion in brain tissue.

MR imaging pulse sequences can be made sensitive to motion by adding gradients with different strengths at different time intervals. In the well-known phase mapping method, such velocity encoding gradients are used to obtain phase information linear to the velocity of the studied object in the direction of the gradient. When very low velocities are studied, a long duration velocity-encoded gradient is required to obtain sufficient velocity sensitivity. In such cases, variation in the object velocity during the execution of the sequence may hamper the accuracy of the method. In this study, we have made a computer simulation of the performance of a phase mapping method sequence (TE = 46 msec) designed for quantitative studies of motion in brain tissue. Using a Gaussian-shaped velocity input function, the time shifting and the amplitude modulation properties of the sequence was studied for various values of the duration, defined as the full width of tenth of maximum (FWTM), of the input function. The time shift corresponded well to the center of the 180 degrees RF pulse, and the amplitude modulation was seen to decrease with increasing time duration of the velocity input function. Applied on in vivo data, where an approximately gaussian-shaped brain motion velocity pattern was assumed to have a duration of 150 msec, the amplitude modulation of the sequence was estimated to 2%.

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.

Cerebrospinal fluid circulation and associated intracranial dynamics. A radiologic investigation using MR imaging and radionuclide cisternography.

UNLABELLED: AIMS OF THE PRESENT INVESTIGATION: Observations made in a preliminary study of pulsatile cerebrospinal fluid (CSF) and brain motions using MR imaging called for a reconsideration of the CSF flow model currently accepted. The following questions were addressed: 1) The nature of the CSF-circulation, e.g., the magnitude and pattern of pulsatile and bulk flow; 2) The driving forces of the CSF circulation and assessment of the role of associated hemodynamics and brain motions; 3) The major routes for the absorption of CSF. MATERIAL AND METHODS: CSF flow and associated hemodynamics were studied using gated MR imaging, in 26 healthy volunteers, 5 patients with communicating hydrocephalus and 10 with benign intracranial hypertension. Radionuclide cisternography was performed in 10 individuals with venous vasculitis. RESULTS AND CONCLUSIONS: 1) The CSF-circulation is propelled by a pulsating flow, which causes an effective mixing. This flow is produced by the alternating pressure gradient, which is a consequence of the systolic expansion of the intracranial arteries causing expulsion of CSF into the compliant and contractable spinal subarachnoid space. 2) No bulk flow is necessary to explain the transport of tracers in the subarachnoid space. 3) The main absorption of the CSF is not through the Pacchionian granulations, but a major part of the CSF transportation to the blood-stream is likely to occur via the paravascular and extracellular spaces of the central nervous system. 4) The intracranial dynamics may be regarded as the result of an interplay between the demands for space by the four components of the intracranial content, i.e. the arterial blood, brain volume, venous blood and the CSF. This interaction is shown to have a time offset within the cerebral hemispheres in a fronto-occipital direction during the cardiac cycle (the fronto-occipital "volume wave"). 5) The outflow from the cranial cavity to the cervical subarachnoid space (SAS) is dependent in size and timing on the intracranial arterial expansion during systole. Similarly, the outflow from the aqueduct mirrors the brain expansion. The brain expansion is typically very small as evident from the minute aqueductal flow observed in healthy individuals. This expansion occurs simultaneously with an inflow of CSF and will be directed inwards towards the ventricular system. The brain expansion is of decisive importance for the formation of the normal transcerebral pressure gradient. 6) The instantaneous increase of flow in the superior sagittal sinus at the beginning of the systole reflects a direct pressure transmission via the SAS from the expanding arteries to the cerebral veins. It is contended that this early increase in venous pressure together with the volume wave is most likely an important prerequisite for sustaining normal intracranial pressure (ICP) and normal cerebral blood flow. This counter pressure should be reduced in hydrocephalus due to the decreased arterial expansion and could explain the reduced blood flow as well as an increased transmantle pressure gradient causing the ventricular dilatation. An increased pressure in the venous system is likely to be the cause of increases in ICP, including the increased pressure observed in benign intracranial hypertension (BIH).

MRI of intracranial arteries in nitroglycerin induced cluster headache attacks.

Eight patients with episodic cluster headache, five in active episode, three out of episode, were investigated as to diameters of intracranial arteries before and after nitroglycerin (NG) administration. The diameter of all intracranial carotids were increased about 10 minutes after NG, although more in the patients in episode than in patients out of episode. The dilatation remained for the next 60 minutes in the patients who did not get a cluster headache attack. There was a normalization of the diameters of the internal carotid arteries compared to the initial values, at maximum pain in all patients who got a cluster headache attack. Similar changes were also found in the basilar arteries. The findings support the hypothesis of a constriction of intracranial arteries at maximum pain in cluster headache attacks to stop the pain.

Cerebrospinal fluid pressure and venous pressure in "dynamite headache" and cluster headache attacks.

Six patients with episodic cluster headache were investigated as to blood pressure, heart rate, cerebrospinal fluid pressure (Pcsf) and frontal vein pressure (Pvf) during five nitroglycerin (NG) provoked attacks and one spontaneous attack. In a seventh studied patient the NG failed to provoke an attack. The earlier reported decrease of systolic blood pressure and increase of diastolic blood pressure and heart rate after NG administration were also found in these patients. The "dynamite headache" was related to the start and duration of an increase of the cerebrospinal fluid pressure. There was no relationship between the start or the maximum pain of the cluster headache attack and changes in Pcsf or Pvf. On breathing oxygen during a cluster headache attack, there was a decrease of Pcsf but in some patients a temporary increase of Pvf was observed, which possibly indicates that oxygen simultaneously attains constriction of arteries and veins.

Spread of blood in cerebrospinal fluid following craniotomy simulates spinal metastases.

Postoperative myelography with water-soluble contrast media was performed in 36 children with a diagnosis of posterior cranial fossa tumour. The myelograms were normal in 15. In 5 an intramedullary tumour was present and 3 of these had in addition subarachnoid changes as evidence of tumour spread. The remaining 16 patient had subarachnoid changes of a different character, mainly located in the posterior thoracic region and similar to those seen after subarachnoid haemorrhage. It is suggested that they represent adhesions caused by blood from the operation. The blood is assumed to be distributed by the large cerebrospinal fluid pulsations to the cervical and thoracic regions. It is important to recognise and differentiate subarachnoid changes due to tumour and to postoperative adhesions to avoid unnecessary radiotherapy to the spinal cord.

A new view on the CSF-circulation with the potential for pharmacological treatment of childhood hydrocephalus.

A new model of the cerebrospinal fluid (CSF) circulation is proposed, implying that the main absorption of CSF occurs through the brain capillaries. This model is based on recent observations of CSF dynamics using radionuclide cisternography and cardiac gated magnetic resonance imaging. Magnetic resonance imaging of communicating hydrocephalus has demonstrated a highly significant decrease of CSF flow through the foramen magnum, which is explained by decreased expansion of the intracranial arteries. This invariable finding in combination with the new view of the CSF-circulation makes a hemodynamic pathogenesis of hydrocephalus very probable. Communicating hydrocephalus may be caused by any process that restricts the arterial pulsations and is therefore termed restricted arterial pulsation hydrocephalus. In obstructive hydrocephalus, the ventricular dilatation leads to a compression of the cortical veins and consequently is termed venous congestion hydrocephalus. Based on these considerations, a new concept of pharmacological treatment of hydrocephalus is proposed by using a selective venous constrictor.

Periorbital venous vasculitis, intracranial hypertension and empty sella.

Sixteen patients with chronic periorbital venous vasculitis without nerve involvement and 9 patients with active episodic cluster headache were studied as to cerebrospinal fluid (CSF) pressure. Eighty-one percent of the patients with chronic and 33% with episodic symptoms had pathologically increased CSF pressure. Magnetic resonance imaging of the brains in the chronic group showed empty sella in 60%, cerebral atrophy in 21% and white matter lesions with high signal intensity on T2 weighted sequences in 29%. Abnormal obesity was found in 31% of the patients with chronic periorbital venous vasculitis under 60 years of age. Venous vasculitis is suggested as a cause of intracranial hypertension, empty sella, and endocrinologic dysfunctions.

On the pulsatile nature of intracranial and spinal CSF-circulation demonstrated by MR imaging.

Cerebrospinal fluid (CSF) flow was studied in 24 healthy volunteers using gated MR phase imaging. The subarachnoid space (SAS) was divided into 5 compartments depending on the magnitude of the pulsatile CSF flows: a high velocity compartment in the area of the brain stem and spinal cord, 2 slow ones at the upper and lower extremes of the SAS, and finally 2 intermediate velocity compartments in between. The main pulsatile spinal flow channel had a meandering pattern. The extraventricular CSF-circulation can be explained by pulsatile CSF flow without the necessity of assuming existence of a net flow. A successive time offset during the cardiac cycle has been found in the fronto-occipital direction of the interplay between the arterial expansion, brain expansion, volume changes of the CSF spaces and of the veins. It is proposed to name this time offset the intracranial "volume wave" (VoW).

Is there a relationship between obesity and intracranial hypertension?

BACKGROUND: Intracranial hypertension and obesity have been reported in recent studies of patients with periorbital venous vasculitis. These findings indicate that obese patients should be investigated for signs of inflammation in serum and lumbar cerebrospinal fluid (CSF) pressure. PATIENTS AND METHODS: Twenty obese females, aged 27-68 years participated in the study of associated symptoms, signs of inflammation in serum, intracranial hypertension and magnetic resonance imaging of the brain (MR). Twenty randomly selected age- and sex-matched females were also investigated for associated symptoms and MR as controls. RESULTS: There were no statistically significant differences in associated symptoms and diseases except for infertility (P < 0.05) between the two groups. The values for orosomucoid, haptoglobin, IgG, IgM and tests for rheumatic and antinuclear factors were significantly increased in the obese group compared with normal values at the hospital. The lumbar CSF pressure was increased above 20 cm water in 79% and above 25 cm water in 42% in the obese patients. MR showed that the subarachnoidal space in the obese patients were significantly smaller than in the controls. CONCLUSIONS: Signs of inflammation in serum, intracranial hypertension and decreased subarachnoidal space were statistically significantly more common in patients with obesity, than in controls.

Changes in cross-sectional area in human exercising and non-exercising skeletal muscles.

This research was performed to study how the cross-sectional area (CSA) changes in the skeletal muscles of exercising (E-leg) and contralateral non-exercising (N-leg) legs and to evaluate to what extent changes in CSA mirror changes in blood flow or extravascular water displacement. Seven healthy volunteers performed plantar flexion exercise at three different exercise intensities for 10 min each. Six plantar flexions followed by a 2-s rest in between allowed repeated measurement of the blood flow to the lower limbs by duplex ultrasonography in the popliteal artery and CSA by magnetic resonance imaging. The CSA was measured using manual planimetry at rest and after 3 and 9 min of the exercise periods. The CSA increased in the E-leg by 4.5% and decreased in the N-leg by -2.4%, from rest to highest exercise intensity. Post-exercise imaging of the E-leg showed a bi-phasic recovery of CSA with a rapid phase followed by a slower phase while the blood flow very rapidly returned almost to basal. The time course of the post-exercise decrease indicated that about 50% of the increase in CSA at the highest exercise intensity might have been a result of extravascular water displacement and 50% of an increase in the vasculature volume related to the flow increase. The CSA reduction in N-leg seems to have been related to vasoconstriction, probably mainly of the capacitance vessels since blood flow was not reduced.

Gamma knife surgery in acoustic tumours.

Presentation of the experiences with 254 acoustic neurinomas, treated at the Karolinska Gamma Knife Center from 1969 to 1991, with a minimum follow-up of 12 months. Early loss of contrast enhancement on CT or MRI was seen in 70%. Unilateral tumours showed size decrease in 55%, no change in 33%, and increase in 12%. NF 2 tumours had decrease in 33%, no change in 43%, and increase in 24%. Some degree of facial weakness was seen after 17% of treatments, but always with later improvement of function. The incidence of trigeminal neuropathy was 19%. Preservation of hearing was 77%. Gamma knife treatment is as efficient as microsurgery, but without risk of infection, bleeding or CSF leak. It requires no hospitalisation. The patient can go back to work after a few days. It therefore should be offered as an alternative to every acoustic neurinoma patient.

Skeletal muscle perfusion during exercise using Gd-DTPA bolus detection.

The study was performed to evaluate if skeletal muscle perfusion can be determined during exercise using an IV bolus injection of Gd-DTPA. A fast spoiled gradient echo sequence (T1 weighted) was used with intermittent imaging during one-legged plantar flexion at different workloads. Between repetitive flexions, a 2-sec rest allowed magnetic resonance imaging (MRI) of the lower legs and measurements of the blood flow in the popliteal artery by ultrasonography for subsequent calculation of muscle perfusion. Maximal signal intensity, upslope and downslope of the bolus, mean transit time, and integrated curve area were measured within regions of interest bilaterally. The skeletal muscle perfusion estimated by ultrasonography increased in the exercising leg from 4 ml x 100 g(-1) x min(-1) at rest to 38 ml at low, 86 ml at medium, and 110 ml x 100 g(-1) x min(-1) at high workload. The SImax increased from 1.38 +/- 0.12 to 1.58 +/- 0.15 and the negative slope of the peak nonsignificantly from - 2.38 +/- 1.75 to - 12.05 +/- 9. 71. All obtained MRI parameters could visually separate the muscles into exercising, nonexercising, and presumably low active muscles. It is concluded that the signal intensity curve using a fast spoiled gradient echo sequence did not overall quantitatively mirror the perfusion, evaluated as the blood flow measured by ultrasonography. However, the signal intensity seemed to follow the blood flow velocity within a limited range of 15-60 cm x sec(-1), corresponding to 35-90 ml x 100 g(-1) x min(-1). Nonetheless, it might be useful when studying ischemia or endothelial dysfunction in skeletal muscles during exercise.

MR spectroscopy shows reduced frontal lobe concentrations of N-acetyl aspartate in patients with juvenile myoclonic epilepsy.

PURPOSE: Neuropsychological studies suggest frontal lobe dysfunctions in patients with juvenile myoclonic epilepsy (JME). In this study we investigated whether an underlying mechanism could be a regional neuronal damage not visible with structural magnetic resonance (MR), but detectable with magnetic resonance spectroscopy (MRS). METHODS: The study included 15 patients with JME and 10 matched healthy controls. Quantitative single voxel MRS was conducted at 1.5 Tesla by using a STEAM sequence (TR/TE/TM = 6,000/30/13.7 ms). The voxels were placed over the right cerebellum, right thalamus, and the prefrontal and occipital cortex. The quantitation included fitting of transmitter gain, and correction for partial volume of cerebrovascular fluid. LC-Model was used for estimation of the absolute concentrations of total N-acetyl aspartate (NAA), cholines, total creatine, and myoinositol. RESULTS: Patients with JME had significantly reduced prefrontal concentrations of NAA in relation to controls (9.1 +/- 1.0 vs. 10.2 +/- 0.8 mM; p = 0.031 after Bonferroni correction). The other regions showed normal NAA values, as did the other metabolites. CONCLUSIONS: The observed reduction in NAA levels suggests a prefrontal neuronal lesion in patients with JME.

SUNCT may be another manifestation of orbital venous vasculitis.

A patient with more than 20 years of SUNCT, i.e., long lasting periods with frequent attacks of intense orbital pain with a duration of about one minute, associated with ipsilateral conjunctival injection, lacrimation, rhinorrhea and facial sweating is described. Some attacks were possibly related to increased cerebral blood flow but could also be triggered from the oral area. Orbital phlebography showed pathologic changes on the side of the pain, changes which were normalized when these attacks ceased to appear. Due to these findings in conjunction with serum evidence of inflammation, associated systemic symptoms and susceptibility to steroids and azathioprine, venous vasculitis is suggested to be the cause of SUNCT in this patient. Carbamazepine and sumatriptan decreased the frequency, intensity and duration of attacks, although not completely.

Reliability and reproducibility of brain tissue volumetry from segmented MR scans.

Reliable measurement of different tissue volumes in the living brain is of great importance for human brain research. In this article, we report on the inter- and intraoperator reliability and scan-rescan reproducibility of segmented intracranial tissue volumes from MR images using the image analysis software suite BRAINS. The absolute data of tissue volume measurements are also presented. The reliability and consistency of the measurements of the segmented volumes were excellent. The segmentation is robust and rapid and the volume measurements are plausible and suitable for quantitative studies in clinical brain research.

Magnetic resonance imaging and echocardiography in assessment of ventricular function in atrially corrected transposition of the great arteries.

OBJECTIVE: Patients with atrial correction of transposition of the great arteries (TGA) may develop right ventricular (RV) failure with time. A reliable non-invasive method for assessment of ventricular function is therefore needed. To evaluate the accuracy of echocardiography in assessment of ventricular volumes and function in these patients we compared echocardiography with magnetic resonance imaging (MRI) in 10 patients late after the Mustard and Senning procedures. DESIGN: Prospective echocardiography and MRI examinations were performed on the same day. Two different echocardiography technicians examined all patients. All echocardiography and MRI examinations were performed at the university hospital outpatient clinic and MRI department respectively. Ten patients, age 14.0+/-2.9 years, who had been operated on with atrial correction of TGA at 8 (2-60) months of age (median and range) were examined. Echocardiography RV and left ventricular (LV) end-systolic volumes (ESVs), end-diastolic volumes (EDVs), stroke volumes (SVs) and ejection fractions (EFs) were calculated, using the modified Simpson method, and compared with the same measurements obtained from MRI. RESULTS: For RV function there was good agreement between echocardiography- and MRI-derived measurements. Both echocardiography and MRI revealed reduced RV function with EFs of 42.6+/-9.1% and 46.4+/-7.2% respectively. For RV volumes there were no significant differences between echocardiography and MRI. LV function was significantly overestimated by echocardiography (EF with echocardiography = 72.7+/-4.4% vs. MRI = 50.5+/-7.6%) while all LV volumes were greatly underestimated. Echocardiography measurements of volumes in repeated examinations by different technicians showed large variations, 13-50%, for different variables in individual patients. CONCLUSION: Echocardiography can provide clinically important information concerning RV function in follow-up of patients late after atrial correction of TGA. It has limited value in assessment of LV function in these patients. Volume measurements by echocardiography are, however, highly user-dependent and interobserver variation is high. MRI may accordingly serve as an important reference method in individual patients.

Pulsatile brain movement and associated hydrodynamics studied by magnetic resonance phase imaging. The Monro-Kellie doctrine revisited.

Brain 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 funnel-shaped 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 re-moulding 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.