Red cerebral veins and the cerebral steal syndrome. Evidence from fluorescein angiography and microregional blood flow by radioisotopes during excision of an angioma

Recent fluorescence imaging technology applications of indocyanine green in general thoracic surgery

Thoracic surgeons perform a wide variety of cancer operations, which are often associated with high morbidity and mortality. Thus, thoracic surgery involves many special challenges that require innovative solutions. The increased utilization of minimally invasive practices, poor overall cancer survival, and significant morbidity of critical operations remain key obstacles to overcome. Fluorescence imaging technology (FIT), involving the implementation of fluorescent dyes and imaging systems, is currently used as an adjunct for general thoracic surgery in many situations and includes sentinel lymph node mapping, pulmonary intersegmental plane identification, pulmonary nodule identification, pulmonary bullous lesion detection, evaluation of the anastomotic perfusion after tracheal surgery, and thoracic duct imaging for postoperative chylothorax. This technology enhances the surgeon's ability to perform operations, and has specific advantages. We review some of the key studies that demonstrate the applications of FIT in the field of general thoracic surgery, focusing on the use of indocyanine green.

Application of Fluorescein Fluorescence in Vascular Neurosurgery

Background: Fluorescein sodium (FNa) is a fluorescent drug with a long history of use for assessing retinal blood flow in ophthalmology; however, its application in vascular neurosurgery is only now gaining popularity. This review summarizes the current knowledge about using FNa videoangiography in vascular neurosurgery.

Methods: We performed a literature review on the usage of FNa for fluorescent videoangiography procedures in neurosurgery. We analyzed methods of injection, dosages of FNa, visualizing platforms, and interpretation of FNa videoangiography. We also reviewed practical applications of FNa videoangiography during various vascular neurosurgeries.

Results: FNa videoangiography can be performed with intraarterial (intracarotid) or intravenous dye injections. Both methods provide excellent resolution with enhanced fluorescence that shows intravascular blood flow on top of visible surrounding anatomy, and both allow simultaneous purposeful microsurgical manipulations. Although it is invasive, an intracarotid FNa injection results in faster contrast appearance and higher-intensity fluorescence and requires a lower dose per injection (reported range, 1–50 mg) compared with peripheral intravenous FNa injection (reported range, 75–2,000 mg or 1–1.5 mg/kg body weight). Four optical excitation/detection tools for FNa videoangiography have been successfully used: conventional xenon-light operating microscope with a special filter set, pencil-type light-emitting diode probe with a filter set, laser-illumination operating microscope, and an endoscope with a filter set. FNa videoangiography was reported to be feasible and useful in various clinical scenarios, such as examining the feeders and drainers in arteriovenous malformation surgery, checking the patency of a microvascular anastomosis, and assessing blood flow during aneurysm clipping. FNa videoangiography can be repeated during the same procedure and used along with indocyanine green (ICG) videoangiography.

Conclusions: Compared with ICG videoangiography, FNa videoangiography has the advantages of enabling real-time inspection and better visualization at deep locations; however, thick vessel walls limit visualization of FNa in larger vessels. FNa videoangiography is a useful tool in multiple neurovascular scenarios and merits further studies to establish its clinical value.

Optical Principles of Fluorescence-Guided Brain Tumor Surgery: A Practical Primer for the Neurosurgeon

Fluorescence-guided surgery is a rapidly growing field that has produced some of the most important innovations in surgical oncology in the past decade. These intraoperative imaging technologies provide information distinguishing tumor tissue from normal tissue in real time as the surgery proceeds and without disruption of the workflow. Many of these fluorescent tracers target unique molecular or cellular features of tumors, which offers the opportunity for identifying pathology with high precision to help surgeons achieve their primary objective of a maximal safe resection. As novel fluorophores and fluorescent probes emerge from preclinical development, a practical understanding of the principles of fluorescence remains critical for evaluating the clinical utility of these agents and identifying opportunities for further innovation. In this review, we provide an "in-text glossary" of the fundamental principles of fluorescence with examples of direct applications to fluorescence-guided brain surgery. We offer a detailed discussion of the various advantages and limitations of the most commonly used intraoperative imaging agents, including 5-aminolevulinic acid, indocyanine green, and fluorescein, with a particular focus on the photophysical properties of these specific agents as they provide a framework through which to understand the new agents that are entering clinical trials. To this end, we conclude with a survey of the fluorescent properties of novel agents that are currently undergoing or will soon enter clinical trials for the intraoperative imaging of brain tumors.

Intraoperative fluorescence imaging in thoracic surgery

Intraoperative fluorescence imaging (IFI) can improve real-time identification of cancer cells during an operation. Phase I clinical trials in thoracic surgery have demonstrated that IFI with second window indocyanine green (TumorGlow^® ) can identify subcentimeter pulmonary nodules, anterior mediastinal masses, and mesothelioma, while the use of a folate receptor-targeted near-infrared agent, OTL38, can improve the specificity for diagnosing tumors with folate receptor expression. Here, we review the existing preclinical and clinical data on IFI in thoracic surgery.

Intraoperative Imaging for Vascular Lesions

Neurovascular surgery is a broad and challenging, yet exciting field within neurologic surgery. The neurovascular surgeon must be meticulous; because the brain and spinal cord are unforgiving to ischemic insults. Along with the pressures of this demanding subspecialty comes the potential to help patients recover from potentially devastating pathology to go on and lead normal, healthy lives. Several intraoperative imaging modalities are available to help maximize treatment success while reducing risk. This article reviews each of these modalities, including digital subtraction angiography, fluorescence angiography, Doppler ultrasonography, laser Doppler, laser speckle contrast imaging, neuronavigation, and neuroendoscopy.

Intraoperative fluorescein staining for benign brain tumors

OBJECTIVE

Successful use of high-dose fluorescein-sodium (20mg/kg) with a standard light microscope for resection of high-grade gliomas, meningiomas, hemangioblastoma and metastases was reported. The principle of brain tumor staining by fluorescein-sodium (Fl-Na) consists in the accumulation of fluorescein in brain tumors with impaired blood-brain barrier. The aim of our study was to investigate for the first time the usefulness of high-dose fluorescein in patients operated on for benign neuroepithelial brain tumors (grade I WHO tumors) with contrast enhancement on magnetic resonance imaging.

METHODS

Our study included 11 patients operated on for benign neuroepithelial primary brain tumors with contrast enhancement on magnetic resonance imaging (MRI): pilocytic astrocytomas (5 patients), dysembrioplastic neuroepithelial tumors (4) and gangliogliomas grade I (2). In all cases, Fl-Na was injected intravenously (20mg/kg) just after the craniotomy using a peripheral venous line. The dural opening was performed 10min later. Microsurgical tumor resection using conventional neurosurgical microscope guided by the fluorescein staining was performed.

RESULTS

Complete resection of the yellow-green stained tissue was achieved in 10 patients confirmed by postoperative control MRI study. Subtotal resection of the colored tissue was achieved in one case with fourth ventricle pilocytic astrocytoma because of the involvement of the medial eminence and functional constraints discovered during intraoperative neuromonitoring. Three patients have had a postoperative volume of resection greater than the tumor volume because of the planed perilesionectomy by our epilepsy surgery team. Surrounding tissue not stained by Fl-Na was obtained in these 3 cases. The histopathological examination did not find tumor tissue in the perilesional Fl-Na negative tissue. On the other hand, all 11 Fl-Na positive specimens presented signs of tumor involvement. We did not observe complications related to the use of high dose Fl-Na.

CONCLUSIONS

High doses intravenous Fl-Na seems to be a useful intraoperative technique for delineation of benign neuroepithelial brain tumors with contrast enhancement. Further larger studies may reveal the real value of high doses Fl-Na as intraoperative method for increasing the extent of resection in these particular indications.

Development of and Clinical Experience with a Simple Device for Performing Intraoperative Fluorescein Fluorescence Cerebral Angiography: Technical Notes

To perform intraoperative fluorescence angiography (FAG) under a microscope without an integrated FAG function with reasonable cost and sufficient quality for evaluation, we made a small and easy to use device for fluorescein FAG (FAG filter). We investigated the practical use of this FAG filter during aneurysm surgery, revascularization surgery, and brain tumor surgery. The FAG filter consists of two types of filters: an excitatory filter and a barrier filter. The excitatory filter excludes all wavelengths except for blue light and the barrier filter passes long waves except for blue light. By adding this FAG filter to a microscope without an integrated FAG function, light from the microscope illuminating the surgical field becomes blue, which is blocked by the barrier filter. We put the FAG filter on the objective lens of the operating microscope correctly and fluorescein sodium was injected intravenously or intra-arterially. Fluorescence (green light) from vessels in the surgical field and the dyed tumor were clearly observed through the microscope and recorded by a memory device. This method was easy and could be performed in a short time (about 10 seconds). Blood flow of small vessels deep in the surgical field could be observed. Blood flow stagnation could be evaluated. However, images from this method were inferior to those obtained by currently commercially available microscopes with an integrated FAG function. In brain tumor surgery, a stained tumor on the brain surface could be observed using this method. FAG could be performed with a microscope without an integrated FAG function easily with only this FAG filter.

Critical cerebral perfusion pressure at high intracranial pressure measured by induced cerebrovascular and intracranial pressure reactivity

OBJECTIVES

The lower limit of cerebral blood flow autoregulation is the critical cerebral perfusion pressure at which cerebral blood flow begins to fall. It is important that cerebral perfusion pressure be maintained above this level to ensure adequate cerebral blood flow, especially in patients with high intracranial pressure. However, the critical cerebral perfusion pressure of 50 mm Hg, obtained by decreasing mean arterial pressure, differs from the value of 30 mm Hg, obtained by increasing intracranial pressure, which we previously showed was due to microvascular shunt flow maintenance of a falsely high cerebral blood flow. The present study shows that the critical cerebral perfusion pressure, measured by increasing intracranial pressure to decrease cerebral perfusion pressure, is inaccurate but accurately determined by dopamine-induced dynamic intracranial pressure reactivity and cerebrovascular reactivity.

DESIGN

Cerebral perfusion pressure was decreased either by increasing intracranial pressure or decreasing mean arterial pressure and the critical cerebral perfusion pressure by both methods compared. Cortical Doppler flux, intracranial pressure, and mean arterial pressure were monitored throughout the study. At each cerebral perfusion pressure, we measured microvascular RBC flow velocity, blood-brain barrier integrity (transcapillary dye extravasation), and tissue oxygenation (reduced nicotinamide adenine dinucleotide) in the cerebral cortex of rats using in vivo two-photon laser scanning microscopy.

SETTING

University laboratory.

SUBJECTS

Male Sprague-Dawley rats.

INTERVENTIONS

At each cerebral perfusion pressure, dopamine-induced arterial pressure transients (~10 mm Hg, ~45 s duration) were used to measure induced intracranial pressure reactivity (Δ intracranial pressure/Δ mean arterial pressure) and induced cerebrovascular reactivity (Δ cerebral blood flow/Δ mean arterial pressure).

MEASUREMENTS AND MAIN RESULTS

At a normal cerebral perfusion pressure of 70 mm Hg, 10 mm Hg mean arterial pressure pulses had no effect on intracranial pressure or cerebral blood flow (induced intracranial pressure reactivity = -0.03 ± 0.07 and induced cerebrovascular reactivity = -0.02 ± 0.09), reflecting intact autoregulation. Decreasing cerebral perfusion pressure to 50 mm Hg by increasing intracranial pressure increased induced intracranial pressure reactivity and induced cerebrovascular reactivity to 0.24 ± 0.09 and 0.31 ± 0.13, respectively, reflecting impaired autoregulation (p < 0.05). By static cerebral blood flow, the first significant decrease in cerebral blood flow occurred at a cerebral perfusion pressure of 30 mm Hg (0.71 ± 0.08, p < 0.05).

CONCLUSIONS

Critical cerebral perfusion pressure of 50 mm Hg was accurately determined by induced intracranial pressure reactivity and induced cerebrovascular reactivity, whereas the static method failed.

Neurosurgical confocal endomicroscopy: A review of contrast agents, confocal systems, and future imaging modalities

Background:

The clinical application of fluorescent contrast agents (fluorescein, indocyanine green, and aminolevulinic acid) with intraoperative microscopy has led to advances in intraoperative brain tumor imaging. Their properties, mechanism of action, history of use, and safety are analyzed in this report along with a review of current laser scanning confocal endomicroscopy systems. Additional imaging modalities with potential neurosurgical utility are also analyzed.

Methods:

A comprehensive literature search was performed utilizing PubMed and key words: In vivo confocal microscopy, confocal endomicroscopy, fluorescence imaging, in vivo diagnostics/neoplasm, in vivo molecular imaging, and optical imaging. Articles were reviewed that discussed clinically available fluorophores in neurosurgery, confocal endomicroscopy instrumentation, confocal microscopy systems, and intraoperative cancer diagnostics.

Results:

Current clinically available fluorescent contrast agents have specific properties that provide microscopic delineation of tumors when imaged with laser scanning confocal endomicroscopes. Other imaging modalities such as coherent anti-Stokes Raman scattering (CARS) microscopy, confocal reflectance microscopy, fluorescent lifetime imaging (FLIM), two-photon microscopy, and second harmonic generation may also have potential in neurosurgical applications.

Conclusion:

In addition to guiding tumor resection, intraoperative fluorescence and microscopy have the potential to facilitate tumor identification and complement frozen section analysis during surgery by providing real-time histological assessment. Further research, including clinical trials, is necessary to test the efficacy of fluorescent contrast agents and optical imaging instrumentation in order to establish their role in neurosurgery.

Intra-arterial fluorescence angiography with injection of fluorescein sodium from the superficial temporal artery during aneurysm surgery: technical notes

Intra-arterial fluorescence angiography from a catheter inserted into the external carotid artery (ECA) via the superficial temporal artery (STA) allowed us to satisfactorily evaluate cerebral arterial and venous blood flow. We report this novel method that allowed for repeated angiography within minutes with a low risk of complications due to catheter placement from the STA. The STA was secured at the edge of the standard skin incision during cerebral aneurysm surgery. A 3 Fr catheter was inserted approximately 5 cm to 10 cm into the STA. After manual injection of 5 ml of 20 times diluted 10% fluorescein sodium (fluorescein), fluorescein reached the intracranial internal carotid artery (ICA) through the common carotid artery or anastomoses between the ECA and ICA. Fluorescence emission from the cerebral arteries, capillaries, and veins was clearly observed through the microscope and results were recorded. Quick dye clearance makes it possible to reexamine within 1 minute. In addition, we made a graph of the fluorescence emission intensity in the arteries, capillaries, and veins using fluorescence analysis software. With intravenous fluorescence angiography, dye remains in the vessels for a long time. When repeated examinations are necessary, intervals of approximately 10 minutes are required. There were some cases we could not correctly evaluate with intravenous injection due to weak fluorescence emission. Fluorescence angiography with intra-arterial injection from a catheter inserted into the carotid artery or another major vessel, like conventional angiography, has a risk of procedure-related complications. We report our new method since it solved these problems and is useful.

Technical principles and neurosurgical applications of fluorescein fluorescence using a microscope-integrated fluorescence module

BACKGROUND

Fluorescent technology has recently become a valuable tool in the surgical management of neoplastic and vascular lesions. The availability of microscope-integrated fluorescent modules has facilitated incorporation of this technology within the microsurgical workflow. The currently available microscope integrated modules use 5-aminolevulinic acid (5-ALA) and indocyanine green (ICG) as fluorophores.

METHODS

Fluorescein sodium is a fluorescent molecule that has been used specifically in ophthalmology for the treatment of retinal angiography. A new microscope-integrated fluorescent module has been recently developed for fluorescein. We employed this technology to maximize resection of tumors and perform intraoperative angiography to guide microsurgical management of aneurysms and arteriovenous malformations.

RESULTS

Fluorescein fluorescence allows the surgeon to appreciate fluorescent structures through the oculars while visualizing non-fluorescent tissues in near natural colors. Therefore, the operator can proceed with microsurgery under the fluorescent mode. We present three representative cases in which the use of fluorescein fluorescence was found useful in the surgeon's decision making during surgery.

CONCLUSIONS

The applications of this new microscope-integrated fluorescent module are multiple, and include vascular and oncologic neurosurgery. Further clinical investigations with large patient cohorts are needed to fully establish the role of this new technology.

High intracranial pressure effects on cerebral cortical microvascular flow in rats

To manage patients with high intracranial pressure (ICP), clinicians need to know the critical cerebral perfusion pressure (CPP) required to maintain cerebral blood flow (CBF). Historically, the critical CPP obtained by decreasing mean arterial pressure (MAP) to lower CPP was 60 mm Hg, which fell to 30 mm Hg when CPP was reduced by increasing ICP. We examined whether this decrease in critical CPP was due to a pathological shift from capillary (CAP) to high-velocity microvessel flow or thoroughfare channel (TFC) shunt flow. Cortical microvessel red blood cell velocity and NADH fluorescence were measured by in vivo two-photon laser scanning microscopy in rats at CPP of 70, 50, and 30 mm Hg by increasing ICP or decreasing MAP. Water content was measured by wet/dry weight, and cortical perfusion by laser Doppler flux. Reduction of CPP by raising ICP increased TFC shunt flow from 30.4±2.3% to 51.2±5.2% (mean±SEM, p<0.001), NADH increased by 20.3±6.8% and 58.1±8.2% (p<0.01), and brain water content from 72.9±0.47% to 77.8±2.42% (p<0.01). Decreasing CPP by MAP decreased TFC shunt flow with a smaller rise in NADH and no edema. Doppler flux decreased less with increasing ICP than decreasing MAP. The decrease seen in the critical CPP with increased ICP is likely due to a redistribution of microvascular flow from capillary to microvascular shunt flow or TFC shunt flow, resulting in a pathologically elevated CBF associated with tissue hypoxia and brain edema, characteristic of non-nutritive shunt flow.

Diagnosis influences response of cerebral near infrared spectroscopy to intracranial hypertension in children

OBJECTIVE

To describe cerebral regional oxygen saturation measured by near infrared spectroscopy in the setting of normal and increased intracranial pressure in children to evaluate the association between cerebral regional oxygen saturation and intracranial pressure in comparison with other clinical variables.

DESIGN

Prospective observational cohort study.

SETTING

Two academic tertiary care centers' pediatric intensive care units.

PATIENTS

Thirty patients with intracranial pressure and near infrared spectroscopy monitoring (median age, 11.5 yrs; interquartile range, 5.2-13 yrs) for a range of neurologic diagnoses, including brain tumor, trauma, intracerebral hemorrhage, and hydrocephalus.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Temporally correlated cerebral regional oxygen saturation with hematologic (hematocrit), biochemical (pH), and physiological (intracranial pressure, mean arterial pressure, cerebral perfusion pressure, temperature, heart rate, pulse oximetry and end-tidal carbon dioxide) variables. Cerebral regional oxygen saturation during episodes of increased intracranial pressure was lower than with normal intracranial pressure (mean +/- sd intracranial pressure >20 = 71% +/- 13% vs. intracranial pressure <20 = 75% +/- 10%), although the mean difference (-4%) is small compared with variability in the measurement. Neither isolated values nor change in cerebral regional oxygen saturation were significantly associated with intracranial pressure or cerebral perfusion pressure in the overall population. Isolated values and change in end-tidal CO2 were significantly correlated with cerebral regional oxygen saturation and change in cerebral regional oxygen saturation (all p < 0.01). In exploratory analyses, the diagnostic group significantly modified the effect of intracranial hypertension on regional oxygen saturation: regional oxygen saturation decreased during intracranial hypertension in patients with brain tumors (p = .05) and hydrocephalus (p < .001) but increased during intracranial hypertension in those with intracranial hemorrhage (p < .001).

CONCLUSIONS

These data suggest that cerebral regional oxygen saturation is lower with intracranial hypertension. However, the relationship between cerebral regional oxygen saturation and intracranial pressure is strongly influenced by diagnosis.

Confirmation of blood flow in perforating arteries using fluorescein cerebral angiography during aneurysm surgery

Object

The authors performed fluorescein cerebral angiography in patients after aneurysm clip placement to confirm the patency of the parent artery, perforating artery, and other arteries around the aneurysm.

Methods

Twenty-three patients who underwent aneurysm surgery were studied. Aneurysms were located in the internal carotid artery in 12 patients, middle cerebral artery in six, anterior cerebral artery in three, basilar artery bifurcation in one, and junction of the vertebral artery (VA) and posterior inferior cerebellar artery in one. After aneurysm clip placement, the target arteries were illuminated using a beam from a blue light-emitting diode atop a 7-mm diameter pencil-type probe. In all patients, after intravenous administration of 5 ml of 10% fluorescein sodium, fluorescence in the vessels was clearly observed through a microscope and recorded on videotape.

Results

The excellent image quality and spatial resolution of the fluorescein angiography procedure facilitated intra-operative real-time assessment of the patency of the perforating arteries and branches near the aneurysm, including: 12 posterior communicating arteries; 12 anterior choroidal arteries; four lenticulostriate arteries; three recurrent arteries of Heubner; three hypothalamic arteries; one ophthalmic artery; one perforating artery arising from the VA; and one posterior thalamoperforating artery. All 23 patients experienced an uneventful postoperative course without clinical symptoms of perforating artery occlusion.

Conclusions

Because the fluorescein angiography procedure described here allows intraoperative confirmation of the patency of perforating arteries located deep inside the surgical field, it can be practically used for preventing unexpected cerebral infarction during aneurysm surgery.

Hypoxia inducible factor-1α and expression of vascular endothelial growth factor and its receptors in cerebral arteriovenous malformations

BACKGROUND

Vascular endothelial growth factor (VEGF) and its tyrosine kinase family of receptors (VEGFR) (Flt-1, Flk-1, Flt-4) have been implicated in vascular angiogenesis and remodelling in cerebral arteriovenous malformations (CAVM). In this study, we investigate the role of hypoxia inducible factor-1 (HIF-1alpha) in CAVM and its relationship to VEGF and VEGFR.

METHODS

Surgical specimens from 26 patients undergoing CAVM resection were studied for HIF-1alpha , VEGF, Flt-1, Flk-1 and Flt-4. The mean age was 34.08 +/- 14.18 years. Twenty-one patients presented with intracerebral haemorrhage.

RESULTS

VEGF, Flt-1 and Flt-4 were expressed in all specimens. Flk-1 was expressed in 15 of 26 patients. HIF-1alpha was expressed in 15 of 26 patients. HIF-1alpha expression was significantly associated with VEGF, Flt-1 and Flk-1 expression (p < 0.05)

CONCLUSIONS

HIF-1alpha is expressed in human CAVM. The expression of HIF-1alpha is significantly related to VEGF and VEGFR expression, suggesting a possible role for its induction and role in maintaining angiogenesis and vascular remodelling.

Prospective study of neuropsychological and psychosocial outcome following surgical excision of intracerebral arteriovenous malformations

In this prospective study the neuropsychological and psychosocial function of 64 patients undergoing surgical resection of cerebral arteriovenous malformations was examined prior to surgery (T1), one month post-surgery (T2) and one year post-surgery (T3). A mild but widespread cognitive decline was observed pre-operatively. There was a trend toward decreased neuropsychological function at T2. All neuropsychological tests showed a trend toward improvement at T3 compared with both pre-operative (generally not statistically significant) and early post-operative values (generally significant). Patients were assessed for change between testing times. At T2 patients were more likely to have deteriorated than improved, whereas at T3 the group which had altered from baseline were more likely to have improved than deteriorated. Deterioration in some verbal/language tasks was more common for left sided AVMs. Outcome did not differ significantly for patients presenting with haemorrhage. Psychosocial function was unchanged at late follow-up for the majority of patients.

Cerebral hemodynamics and metabolism in patients with cerebral arteriovenous malformations: an evaluation using positron emission tomography scanning

OBJECT

The purpose of this study was to evaluate cerebral hemodynamic and metabolic features in patients with arteriovenous malformations (AVMs) by using positron emission tomography (PET) scanning.

METHODS

Twenty-four patients with supratentorial cerebral AVMs participated in PET studies in which 15O inhalation steady-state methods were used. The authors recorded the values of regional cerebral blood flow (rCBF), regional cerebral blood volume (rCBV), the regional oxygen extraction fraction (rOEF), and the regional cerebral metabolic rate of O2 (rCMRO2) at three designated regions of interest (ROIs) in each patient. These ROIs included perilesional (ROI-p), ipsilateral remote (ROI-i), and contralateral symmetrical (ROI-c) brain regions. To identify the factors that exert a direct effect on the hemodynamics of brains affected by AVM, we also separated the lesions according to their size and flow type shown on angiograms, and grouped the patients according to the presence or absence of progressive neurological deficits. We then compared the PET parameters at different ROIs in individual patients and evaluated the mean values obtained for all 24 patients according to AVM flow type and size, and the presence or absence of progressive neurological deficits.

CONCLUSIONS

Overall, mean rCBV and rOEF values were significantly higher in ROI-p than in ROI-c (p = 0.00046 and p = 0.015, respectively). No significant differences were seen between the ROI-i and ROI-c with respect to rCBF, rCBV, and rOEF. Mean rCMRO2 values were similar in the three ROIs; however, the mean rCBF was significantly lower in the ROI-p than in the ROI-c in patients with high-flow AVMs (p = 0.019), large AVMs (p = 0.017), and progressive neurological deficits (p = 0.021). Furthermore, the mean rOEF values were significantly higher in the ROI-p than in the ROI-c in patients with high-flow AVMs (p = 0.005), large AVMs (p = 0.019), and progressive neurological deficits (p = 0.017). The PET studies revealed hemodynamic impairment characterized by decreased rCBF and increased rOEF and rCBV values in the ROI-p of patients with large, high-flow AVMs regardless of whether they exhibited progressive neurological deficits.

Steal affecting the central nervous system

Steal is a pathophysiological process in which increased blood flow through a low-resistance vascular bed is sufficient to divert flow away from a region of the central nervous system. Three disease states in which steal may cause neurological deficits due to central nervous system ischemia are reviewed. Subclavian steal occurs when stenosis of the subclavian artery proximal to the vertebral origin causes retrograde flow in the left vertebral artery. Patients with anatomic subclavian steal usually do not develop neurological symptoms but may rarely present with posterior circulation ischemia. Arteriovenous malformations alter cerebral blood flow patterns and regional perfusion pressure. It has been hypothesized that cerebral arteriovenous malformations may cause neurological deficits due to steal and that these deficits may be cured with arteriovenous malformation treatment. Intra-arterial pressure measurements and transcranial velocity studies show regional hemodynamic alterations. However, these changes have not been correlated with presenting symptoms. Evidence from single-photon emission computed tomography does suggest a relationship between regional hypoperfusion and neurological deficits. Coarctation of the aorta may divert flow from the spinal cord circulation through intercostal arteries distal to the stenosis. This is a possible but unproven mechanism of myelopathology. Steal syndromes may be amenable to treatment by open surgical or endovascular approaches. Experimental studies of the pathophysiology of steal are strengthened by precise definitions of the measured parameters and innovative applications of technology.

What factors are related to impairment of cerebrovascular reserve before and after arteriovenous malformation resection? A cerebral blood flow study using xenon-enhanced computed tomography

OBJECTIVE

To examine the incidence and possible determinants of impaired vascular reserve in arteriovenous malformation (AVM)-affected brain, before and after surgery.

METHODS

In a prospective study of 30 patients, the regional cerebrovascular reserve capacity (rCRC) and the vasodilated regional cerebral blood flow (rCBF) were assessed during an acetazolamide challenge, using xenon-enhanced computed tomography, before and after complete AVM resection. Single brain slices at the level of the basal ganglia were examined, and scanning through the AVMs was avoided. Five regions of interest in the AVM-bearing hemisphere were compared with their counterparts in the unaffected hemisphere. Vasodilated rCBF reductions of at least 20% in one or more regions of interest and rCRC values of less than 10 ml/100 g/min were considered to be significant.

RESULTS

Ipsilateral vasodilated rCBF was significantly reduced in 17 patients before surgery and 15 patients after surgery. Ipsilateral rCRC was impaired in 14 patients before surgery and 12 patients after surgery. Large AVM size, venous congestion, and AVM-related vascular territories were correlated with impaired vascular reserve in AVM-nonadjacent brain tissue before surgery. Similar correlations were observed after surgery, except that not AVM size but a large number of AVM-supplying vascular territories was correlated. Moreover, the smallest AVMs and those supplied by a single vascular territory, as well as hemorrhage and nonhemorrhagic neurological deficits as presenting symptoms, were correlated with reduced ipsilateral vasodilated rCBF before surgery. Among patients with AVMs and nonhemorrhagic epilepsy, a trend of impaired cerebrovascular reserve was observed. In the only case of postresectional "breakthrough," the preoperative rCRC was not impaired but abnormally high.

CONCLUSION

Among the determinants of impaired cerebrovascular reserve, AVM size is already a constituent of current grading scales and decision-making paradigms, whereas factors such as venous congestion have been less closely considered or less obvious but may deserve increased attention in the future. Nonhemorrhagic epilepsy in patients with AVMs may constitute the clinical equivalent of chronic cerebral ischemia in a murine model. Postresectional breakthrough may be partly attributable to individual predisposition to excessive vasoreactivity in the whole brain.

Distributions of local oxygen saturation and its response to changes of mean arterial blood pressure in the cerebral cortex adjacent to arteriovenous malformations

BACKGROUND AND PURPOSE

To test the hypothesis that neither "steal" as cortical ischemia caused by reduced perfusion pressure nor "breakthrough" on the grounds of loss of pressure autoregulation exist in brain tissue surrounding arteriovenous malformations (AVMs), we established patterns of cortical oxygen saturation (SO(2)) adjacent to AVMs and its behavior after alterations of mean arterial blood pressure.

METHODS

With a microspectrophotometer, SO(2) was scanned in the cortex around AVMs of 44 patients before and after resection and in that of a non-AVM group (n=42) before transsylvian dissection. Autoregulation was evaluated by linear regression analysis after elevation of mean arterial blood pressure (5 microg/min IV noradrenaline). SO(2) values were calculated as medians, percentage of critical values (<25% SO(2)), and coefficients of variance (approximate heterogeneity of SO(2) distributions). All values are given as mean+/-SD.

RESULTS

Forty patients with AVM had an uneventful postoperative course (group A). Four hyperemic complications ("breakthrough") occurred (group B). Autoregulation was tested intact in all groups at all times. Preoperative SO(2) distributions in groups A and C (non-AVMs) were identical. In group B, significantly (P<0.05) lower medians (group A, 52.9+/-16.3%; group B, 44.2+/-17.1%; group C, 51.9+/-11.5% SO(2)), more critical values (group A, 6.5+/-5.1%; group B, 14.7+/-11.1%; group C, 7.1+/-4.9%), and heterogeneous SO(2) distributions (group A, 20.2+/-12.7%; group B, 27.9+/-12.4%; group C, 26.8+/-10.9%) were seen. Increase of median values was significantly higher in group B (76.3+/-10.4% SO(2)) than in group A (65.9+/-13.4% SO(2)) after resection.

CONCLUSIONS

Severely hypoxic areas are uncommon in the cortex adjacent to AVMs and occur predominantly in patients prone to hyperemic complications. Reduced perfusion pressure is compensated in most cases, and moderate hyperemia prevails after excision. Reperfusion into unprotected capillaries of severely hypoxic cortical areas results in "breakthrough," for which vasoparalysis appears not to be the underlying mechanism.

Development of a fluorescein operative microscope for use during malignant glioma surgery: a technical note and preliminary report

BACKGROUND

Fluorescein has been used in the field of neurosurgery; however, fluorescein enhancement or contrast proved to be inadequate because of a lack of appropriate light sources or filters. A new operative microscope system, in which the microscope itself is equipped with excitation and barrier filters, and the application of this system to surgery for malignant glioma are reported.

METHODS

BP 450-490, a glass interference filter used as the excitation filter for the light source optical system, and a Kodak Wratten No. 12 filter used as the barrier filter for the microscope optical system, were incorporated in the operative microscope. A switching apparatus was devised so that filters could be inserted instantly when fluorescence was to be observed. Ten cases in which the location of malignant glioma was enhanced by computed tomography (CT) or magnetic resonance imaging (MRI) were selected for this study. After incision of the dura mater, 8 mg/Kg body weight of fluorescein Na was injected intravenously. Tumor removal was begun some 20 min after the injection with the aid of this newly developed fluorescein operative microscope system.

RESULTS

Fluorescence enhancement and contrast were remarkable when this system was used. It enabled surgical maneuvering while viewing the fluorescent image of objects. The boundaries between the tumor areas enhanced by CT or MRI and the surrounding brain could be clearly distinguished in the fluorescent image; the tumor was totally removed, except for deep lesions, without any neurological deterioration. When the tumor was relatively hard, the area surrounding the tumor was aspirated with a cavitron ultrasonic surgical aspirator; as a result, the tumor could be removed en bloc; otherwise, the fluorescent tumor was removed piece by piece.

CONCLUSIONS

This system provides adequate fluorescent enhancement and contrast and is useful for observing intravenously injected fluorescein during an operation. Though long-term follow-up of such cases is needed, the conditions of our patients immediately after surgery for malignant glioma were satisfactory. These results suggest that our fluorescein operative microscope system is highly effective in surgery for malignant glioma.

Intra-arterial nitrovasodilators do not increase cerebral blood flow in angiographically normal territories of arteriovenous malformation patients

BACKGROUND AND PURPOSE

The mechanism of adaptation to chronic cerebral hypotension in normal brain adjacent to cerebral arteriovenous malformations (AVMs) is unknown. To clarify these mechanisms, we performed cerebral blood flow (CBF) studies in structurally and functionally normal vascular territories during 53 distal cerebral angiographic procedures in 37 patients with AVMs.

METHODS

CBF was measured using the superselective intra-arterial 133Xe method before and after a 3-minute infusion of either verapamil (1 mg.min-1, n = 23), acetylcholine (1.33 micrograms.kg-1.min-1, n = 7), nitroprusside (0.5 microgram.kg-1.min-1, n = 16) or nitroglycerin (0.5 microgram.kg-1.min-1, n = 7).

RESULTS

Mean +/- SD systemic (76 +/- 13 mm Hg) and distal cerebral arterial (55 +/- 16 mm Hg; range, 20 to 97 mm Hg) pressures were not different among groups. Verapamil increased CBF (45 +/- 12 to 65 +/- 21 mL.100 g-1.min-1, P < .001). There was no effect of acetylcholine (no change [46 +/- 9 to 46 +/- 9 mL.100 g-1.min-1], NS) or nitroglycerin (36 +/- 14 to 36 +/- 13 mL.100 g-1.min-1, NS). Nitroprusside decreased CBF (40 +/- 12 to 31 +/- 11 mL.100 g-1.min-1, P < .001). The percent change in CBF after drug administration was proportional to cerebral arterial pressure for verapamil only (r = .57, P = .0051).

CONCLUSIONS

When infused intra-arterially in clinically relevant doses in both hypotensive and normotensive normal vascular territories remote from an AVM nidus, calcium channel blockade caused vasodilation, but there was an absence of response to nitric oxide-mediated vasodilators. These data suggest that (1) the nitric oxide pathway probably is not involved in the adaptation to chronic cerebral hypotension in AVM patients and (2) if our findings in vessels remote from or contralateral to the AVM are applicable to vessels of patients with other forms of cerebrovascular disease, clinically relevant doses of intra-arterial nitrovasodilators may not be useful in the manipulation of cerebrovascular resistance.

Brain tissue gases and pH during arteriovenous malformation resection

OBJECTIVE

The purpose of this study was to determine whether baseline partial pressure of oxygen (PO2), partial carbon dioxide pressure (PCO2), and pH in brain tissue adjacent to arteriovenous malformations (AVMs) are different from those in control patients. In addition, PO2, PCO2, and pH changes were measured during resection of the AVMs.

METHODS

Two groups were studied. Group 1 (n = 8) was composed of nonischemic patients scheduled for cerebral aneurysm clipping. Group 2 (n = 13) was composed of patients undergoing neurosurgery for resection of AVMs. After the craniotomy, the dura was retracted and a combined PO2, PCO2, and pH sensor was inserted into nonischemic brain tissue in Group 1. In Group 2, the sensor was inserted into tissue 2 to 3 cm from the margin of the AVMs, within the same arterial blood supply. After equilibration of the sensor, tissue gases and pH were measured during steady-state anesthetic conditions in Group 1 and during resection of AVMs in Group 2.

RESULTS

Under baseline conditions before the start of surgery, tissue PO2 was decreased in patients with AVMs compared with control patients, but PCO2 and pH were not changed. During resection of the AVMs, PO2 and pH increased and PCO2 decreased compared with baseline measurements. These parameters did not change in control patients during a similar time period.

CONCLUSION

The results suggest that cerebrovascular or metabolic adaptation occurs in patients with AVMs with decreased tissue perfusion pressure as an adjustment for decreased oxygen delivery. During resection of AVMs, this adaptation produces a relative hyperemic environment with tissue hyperoxia, hypocapnia, and alkalosis that is not corrected by the end of surgery.

Superselective intraarterial papaverine administration: effect on regional cerebral blood flow in patients with arteriovenous malformations

In this study the authors determined the effect of papaverine on regional cerebral blood flow (rCBF) in the angiographically normal arteriolar beds of patients with arteriovenous malformations (AVMs) who underwent transfemoral superselective angiography. Middle cerebral artery (MCA) branch vessels were catheterized during 10 procedures performed in nine patients. The mean (+/- standard deviation) largest AVM diameter was 4.4 +/- 1 cm. Regional CBF was measured by recording the washout of a bolus of xenon-133 injected through the microcatheter. In a dose-ranging study. rCBF and MCA pressure in two patients were repeatedly measured after 3-minute infusions of papaverine at 0.07, 0.7, and 7 mg/minute. In a single-dose study, an additional eight patients received only the highest dose of papaverine administered over a 3-minute period. In the dose-ranging study, CBF increased from baseline in a dose-dependent fashion. In the single-dose study, papaverine increased in rCBF 103%, from 48 +/- 11 to 95 +/- 23 ml/100 g/minute at an MCA pressure of 55 +/- 23 mm Hg. Increase in rCBF was linearly related (y = 2.2x - 17, r2 = 0.84; p = 0.001) to baseline MCA pressure (range 22-84 mm Hg). Papaverine increases rCBF in a direct proportion to baseline MCA pressure, even at low baseline pressures. Selective infusion of vasodilators should be investigated in acute cerebral hypotension to facilitate either primary or collateral recruitment of CBF by aiding spontaneous autoregulatory vasodilation. In addition, rCBF monitoring may be useful in determining the most effective intraarterial dose of papaverine while minimizing complications due to hyperemia.

Presence and removal of arteriovenous malformation: Impact of regional cerebral blood flow, as assessed with xenon/CT

In 26 patients with supratentorial AVMs and 1 patient with a dural arteriovenous fistula, the regional cerebral blood flow (rCBF) was assessed by means of xenon-enhanced computed tomography (Xe/CT) before and after complete resective surgery. Each assessment comprised an acetazolamide challenge in order to check the cerebrovascular reserve capacity. While scanning through the AVM was purposely avoided, a single brain slice at the level of the basal ganglia was examined. Five regions of interest (ROIs) in gray matter of the AVM-bearing hemisphere were compared to the contralateral ROIs and categorized into 7 CBF groups. Interhemispherical differences exceeding 20% of the contralateral value in either direction were considered to the significant. AVM-related (AVM-R) and AVM-non-related (AVM-NR) ROIs were looked at separately. Before surgery, all possible rCBF patterns were found, including a normal rCBF as well as a reduced or an increased rCBF, either in AVM-R, AVM-NR, or both. After AVM removal, a rCBF increase in AVM-R is relatively rare, whereas a rCBF decrease is twice as frequent. A rCBF drop to a level of impaired reserve capacity correlates with the occurrence of post-operative neurological deficit.

Intraoperative assessment of aneurysm clip placement by intravenous fluorescein angiography

Clip occlusion of intracranial aneurysms can be difficult, and intraoperative misadventures are still fairly frequent. Many complications are caused by improper clip placement. Intraoperative conventional arteriography is time consuming and requires expensive image-intensifying equipment. As an alternative, we have found that the bolus intravenous administration of fluorescein produces the necessary intra-arterial dye concentration for the direct visualization of the perforating medium and large arteries in proximity to a clipped aneurysm. The technique is straightforward, requiring little additional operating time and an inexpensive set of disposable optical filters. No adverse reactions to fluorescein have been observed.

Continuous monitoring of jugular bulb oxygen saturation as a measure of the shunt flow of cerebral arteriovenous malformations

Jugular bulb oxygen saturation (SjO2) was monitored during preoperative embolization procedures in a consecutive series of 15 patients with large supratentorial arteriovenous malformations (AVM's) in order to test the hypothesis that changes in the shunt flow ratio can be continuously evaluated from the SjO2. A fiberoptic catheter was placed at the dominant jugular bulb. The SjO2 measured using jugular blood withdrawn before embolization was significantly higher than the SjO2 measured at the end of the final embolization procedure (mean +/- standard deviation 84.1% +/- 12.7% vs. 74.2% +/- 10.9%, p < 0.0001), showing a positive correlation with the AVM volume (r = 0.66, p < 0.001). Continuous monitoring of SjO2 via the fiberoptic catheter revealed a progressive decrease in association with the embolization procedures. Microsurgical resection of the AVM was performed at 1 to 2 weeks after the final embolization. Cases in which postoperative hemispheric deformation was revealed on computerized tomography demonstrated a higher SjO2 at the end of embolization compared to that in the remaining cases (81.6% +/- 8.6% vs. 67.8% +/- 8.4%, p < 0.008). Hemispheric deformation was observed in all cases in which the SjO2 did not decline to a level below 90% following embolization. The risk of severe hyperemic complications appeared to be greatly diminished when the SjO2 fell to below 80%. Assuming that the oxygen saturation of the perfusion flow (SjpO2) ranges from 50% to 75%, the ratio of the shunt flow to total flow at an SjO2 of 90% was estimated to be 0.6 to 0.8 based on the following equation: shunt flow/(perfusion flow + shunt flow) = (SjO2 - SjpO2)/(arterial oxygen saturation - SjpO2). These results suggest that monitoring the SjO2 provides real-time information concerning the progress of embolization and helps to determine whether the embolization has progressed sufficiently to avoid postoperative hyperemic complications.

CO2 reactivity in arteriovenous malformations of the brain: a transcranial Doppler ultrasound study

Arteriovenous malformations (AVM's) are congenital tangles of vessels that have a high blood flow through a low-resistance nidus. The vessels in the nidus may lack normal vasoreactivity in response to changes in PaCO2 or perfusion pressure (autoregulation). Arteriovenous malformation hemodynamics have been assessed based on the response of AVM feeding arteries to hypocapnia. Twenty-five AVM patients, aged 34 +/- 11 years (mean +/- standard deviation), were admitted to the Massachusetts General Hospital for proton-beam radiation therapy. Fourteen healthy volunteers aged 30 +/- 7 years served as control subjects. Angiograms with calibrated markers permitting magnification correction were available for all patients. The limits of the middle cerebral artery, as determined by transcranial Doppler ultrasonography, were compared to measurements made on the angiograms. Hyperventilation was induced at a rate set by a metronome. Fixed bilateral Doppler probes allowed almost simultaneous sampling of two vessels. Volunteer control subjects were hyperventilated in two steps. The two PaCO2 step decreases were significant (mean resting PaCO2 40.6 +/- 3.5 mm Hg, Step 1 level 29.4 +/- 3.5 mm Hg and Step 2 level 23.8 +/- 3.5 mm Hg; p < 0.01). These decreases induced a significant decrease in mean flow velocity (Vm) and an increase in the pulsatility index (p < 0.001). Mean carbon dioxide reactivity (% delta Vm/delta PaCO2) was 2.74 +/- 1.0 for Step 1 and 1.44 +/- 1.8 for Step 2 (p < 0.003). The mean PaCO2 decrease in patients was from 39.5 +/- 4.0 mm Hg to 27.0 +/- 3.5 mm Hg. Carbon dioxide reactivity was 0.92 +/- 1.12 for feeding vessels and 2.59 +/- 1.78 for nonfeeding vessels (p < 0.001). Transcranial Doppler ultrasound and angiographic depth measurements correlated well. Hyperventilation induced significantly more hemodynamic changes in control and nonfeeding middle cerebral arteries than in feeding vessels. Impaired CO2 reactivity may help to identify AVM feeding vessels as well as the relative magnitude of the flow provided to the malformation.

Evidence for adaptive autoregulatory displacement in hypotensive cortical territories adjacent to arteriovenous malformations. Columbia University AVM Study Project

We hypothesized that chronic hypotension in normal vascular territories fed by arteriovenous malformation pedicles may reset the lower limit of autoregulation and allow flow to remain constant over a lower pressure range. We studied the effect of increasing systemic mean arterial pressure (SMAP) with phenylephrine on cerebral blood flow using a novel technique. Fourteen patients undergoing 15 procedures were studied before endovascular embolization of arteriovenous malformations under neuroleptic conscious sedation. Mean pressures were transduced via a 1.5-F intracranial microcatheter, which was passed under fluoroscopic guidance into the target feeding artery. The microcatheter was positioned (unwedged) at a point that was relatively hypotensive to systemic pressure but that irrigated normal cortex on angiography; feeding mean arterial pressure (FMAP) and SMAP were recorded. A bolus of 133Xe in saline was injected into the microcatheter, and washout was recorded for 3 minutes by a scintillation detector placed over the vascular territory of the injected pedicle. SMAP was then increased approximately 25 mm Hg by phenylephrine infusion, a second bolus was given, and washout was recorded. After exclusion of the shunt spike, initial slope was calculated. The SMAP (mean +/- standard error) increased from 65 +/- 3 to 89 +/- 2 mm Hg (P < 0.0001), and FMAP increased from 46 +/- 3 to 63 +/- 3 mm Hg (P < 0.0001); cerebral blood flow did not change (40 +/- 2 to 40 +/- 2 ml/100 g per min, P = 0.9199). Dividing the cases on the basis of the baseline FMAP into a "severe" hypotensive group (FMAP = 38 +/- 2; n = 7) and a "moderate" hypotensive group (FMAP = 54 +/- 3; n = 8), cerebral blood flow did not change in either group during phenylephrine challenge. Chronic hypotension does not necessarily result in "vasomotor paralysis" with loss of the ability to vasoconstrict to acute increases in perfusion pressure. Instead, it appears to displace adaptively the lower limit of autoregulation in affected vascular territories by a shift of the autoregulatory curve to the left, conceptually analogous to the adaptive displacement seen with chronic hypertension and its treatment.

Prospective neuropsychological investigation of patients with supratentorial arteriovenous malformations

Thirty-one patients subjected to direct radical excision of a supratentorial arteriovenous malformation (AVM) participated in a comprehensive neuropsychological assessment both pre-operatively and at 4 and 12 months after surgery. The sample constitutes a consecutive series of patients who could complete the neuropsychological assessment before surgery. At the pre-operative assessment the mean results of all the tests fell very close to the average performance of age-equivalent normative samples. Postoperatively, the mean results showed a mild to moderate deterioration of performance on most cognitive and perceptual tasks by 4 months after surgery and a return approximately to the premorbid level by 12 months. Twelve of 15 patients with pre-operative epileptic seizures remained seizure free on medication during this first postoperative year, while two developed seizures de novo. Headache was cured or markedly reduced in all 16 patients incapacitated by headache prior to surgery. No definite emotional or affective changes after surgery were reported by the patients or their relatives. Comparison of pre-operative test results of patients with right-sided and left-sided AVM showed significant differences on five of 24 test parameters. Postoperatively, the number of statistically significant differences increased to nine test parameters at 4 months and ten at the 12-month assessment. The increase in number of statistically significant differences appears to reflect a moderate focal impact of the surgical intervention on cognitive functions in six of the 31 patients (19%). Postoperative improvement of test performance exceeding what may be ascribed to practice effects and chance fluctuations was only encountered in one patient. The present results therefore lend little empirical support to the so-called 'cerebral steal' hypothesis.

Preoperative versus postoperative neuropsychological sequelae of arteriovenous malformations

In a cohort of 14 patients suffering from cerebral arteriovenous malformations (AVM), neuropsychological functioning was examined before and after AVM resection. Improvements after surgery were assumed to be due to enhanced neurocognitive functions associated with the hemisphere ipsilateral to the AVM, and to a lesser extent, with the contralateral hemisphere. Before surgical intervention, the performances of AVM patients were deficient relative to matched normals. Postoperatively, neuropsychological gains were observed particularly in the areas of learning, memory, and higher integrative thought, not only for ipsilateral, but also for contralateral functioning. Contralateral and ipsilateral improvement is consistent with the premise that cerebrovascular steal is lessened; thus, neurosurgical intervention to eliminate arteriovenous shunts was found to result in overall neurobehavioral gains.

Arteriovenous malformation hemodynamics: a transcranial Doppler study

Congenital arteriovenous malformation (AVM) of the brain represents a defect in capillary development resulting in a high flow fistula between arterial and venous systems. In this study, AVM hemodynamics were related with clinical findings. Volume flow was calculated based on transcranial Doppler (TCD) and angiographic data. Forty patients admitted to the Massachusetts General Hospital for proton beam therapy (33 +/- 10 yr old; mean +/- SD) were studied. Four symptoms were considered: intracranial bleeding, progressive neurological deficit, seizures, and headache. Fourteen control subjects aged 30 +/- 7 years (mean +/- SD) were normal volunteers. Angiography with calibrated markers permitting magnification correction was available for all patients. Lateral and medial depth limits of the intracranial basal arteries in relation to the TCD temporal window were determined by TCD and angiogram with excellent correlation. Selected depth for data acquisition was determined independently in the angiogram and by TCD. The difference between the two techniques was less than 4 mm. Mean flow velocity, pulsatility index, and vessel diameter were studied. Flow volume was calculated from these data. Mean flow velocity, pulsatility index, vessel diameter, and flow volume were significantly different among AVM feeders, non-feeders, and control arteries. The non-feeding middle cerebral artery, anterior cerebral artery, and posterior cerebral artery flows were 254 +/- 13, 136 +/- 14, and 79 +/- 8 ml/min, respectively. Accordingly, the estimated cerebral flow volume was 938 ml/min. The feeding middle cerebral artery, anterior cerebral artery, and posterior cerebral artery flows were 552 +/- 47, 369 +/- 70, and 484 +/- 67 ml/min, respectively (P < 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)

Pressure autoregulation is intact after arteriovenous malformation resection

The loss of autoregulatory control of cerebral perfusion to changes in perfusion pressure in tissue remote from an arteriovenous malformation (AVM) has been proposed as the mechanism underlying "normal perfusion pressure breakthrough." This study is the first direct test of this mechanism. Studies were performed during the resection of moderate to large AVMs in 25 patients undergoing 28 procedures under isoflurane anesthesia. Cerebral blood flow (CBF) was measured (xenon-133 method) in the hemisphere adjacent to the nidus before resection after dural exposure (pre), after AVM removal before dural closure at spontaneous systemic blood pressure (post), and, finally, with the mean arterial pressure increased by 20 mm Hg, using phenylephrine (post-BP). AVM resection resulted in a significant enhancement of perfusion in the adjacent hemisphere (30 +/- 2 vs. 25 +/- 1 ml/100g/min, P < 0.01), but no further increase of CBF occurred during increased perfusion pressure (30 +/- 2 ml/100g/min). One patient suffered a postoperative hemorrhage and another developed intraoperative brain swelling during the course of the resection that necessitated staging the procedure. These two patients had the highest increases in CBF, but intact pressure autoregulation. Preserved autoregulation to increased mean arterial pressure after resection does not support a hemodynamic mechanism for the observed increase in CBF from before the resection to after the resection. Pathological events, however, do appear to be related to increases in hemispheric perfusion.

Effects of cerebral angiomas on perifocal and remote tissue: a multivariate positron emission tomography study

BACKGROUND AND PURPOSE

Using multitracer positron emission tomography, I investigated regional hemodynamic and metabolic changes in both perifocal and remote tissues of cerebral angiomas, with special reference to steal phenomena.

METHODS

In 22 patients (14 with arteriovenous malformations and eight with cavernomas) cerebral blood flow, cerebral blood volume, mean vascular transit time, cerebral metabolic rate for oxygen, oxygen extraction fraction, cerebral metabolic rate for glucose, and glucose extraction fraction were measured using standard positron emission tomographic methods. Twelve patients also had their cerebral glucose metabolism assessed during psychophysical activation. Regions of interest representing the angioma, perifocal and remote tissues, contralateral mirror regions, and standard brain regions were analyzed.

RESULTS

There were no significant changes in hemodynamic variables or oxygen metabolism in the ipsilateral cerebral hemisphere, but ipsilateral glucose metabolism was reduced both at rest (p less than 0.01) and during activation (p less than 0.05). Glucose (p less than 0.001) and oxygen (p less than 0.001) metabolism in regions of perifocal tissue with low blood flow were decreased, with substrate extraction fractions showing no increase to compensate for insufficient blood flow. Functional recruitment of the cortex overlying the angioma beyond its periphery and supplied by the same arterial branches was subnormal (p less than 0.05) despite relatively unchanged hemodynamics in this tissue compartment.

CONCLUSIONS

These data suggest that dysfunction of the cortex supplied by arterial branches also feeding the vascular malformation is related to neuronal deafferentation, while the proportionate decrease in blood flow and metabolism of perifocal tissue may be ascribed to neuronal loss in chronically hypoperfused areas, rather than to persistent hemodynamic steal effects.

Lateral ventricular arteriovenous malformations: natural history and surgical indications

The authors report 21 cases of lateral ventricular arteriovenous malformation (AVM) focusing on their natural history and surgical indications. Eighteen of 21 patients (86%) had bleedings prior to admission. We performed definitive surgery in 9 patients and conservative treatment in 12 patients. In 7 of 9 patients (78%) in the operative group and in 5 of 12 patients (42%) in the non-operative group, the nidus of the AVM was less than 4 cm in diameter. The other 2 AVMs in the operative group, more than 4 cm in diameter, were located in the temporal lobe and widely extended to the temporal horn of the lateral ventricle. The nidi of the AVMs were totally removed in all 9 cases in the operative group. All patients have been doing well except one who had a mental breakdown 6 years after surgery. In 10 patients in the non-operative group, AVMs were located in the left cerebral hemispheres. Three patients in the non-operative group had rebleeding of the AVM. Two of the 12 patients in the non-operative group (17%) died of the recurrent haemorrhages. The rate of bleeding of the lateral ventricular AVM seems to be higher than that of the cerebral cortical AVM, but the mortality due to recurrent bleeding might be similar between the two. The operative indications we made depended on the sizes and locations of the AVMs. For AVMs in the temporal horn, even though they were large, we performed total removals of the nidi and had good results. We did not perform any definitive surgery for AVMs more than 4 cm in diameter except for those in the temporal horn. The mortality and morbidity of the 21 cases were 10% and 14%, respectively. Patients with AVMs in the temporal horn and patients with small AVMs in the frontal horn were good candidates for definitive surgery. We undertook conservative treatment for patients with large AVMs, and the results were acceptable.

A model of the pathophysiology of cerebral arteriovenous malformations by a carotid-jugular fistula in the rat

A model of a carotid-jugular fistula in the rat was created such that the arterial feeding vessel is derived from the intracranial arterial circulation and the venous drainage communicates with a major intracranial venous drainage system. This fistula was created in 28 rats on the right side with an additional 11 rats designed as controls with a right carotid ligation and 6 rats without previous surgery. After 12 weeks convalescence, 6 rats with a carotid-jugular fistula and 6 rats without previous surgery underwent cerebral angiography. All fistulae were patent and the model was verified. All of the 33 remaining rats underwent regional cerebral blood flow (rCBF) determination by [14C]iodoantipyrine autoradiography under barbiturate anesthesia. Of the rats with the fistula, 11 had this fistula obliterated 5 min prior to rCBF determination ('closed' group) while 11 had rCBFs determined with the fistula ('open' group). The rCBF was measured from each hemisphere from 7 anatomical regions. The rCBF in the control animals ranged from a median of 82 to 112 ml/100 g/min, in the 'open fistula' group 46 to 68 ml/100 g/min, and in the 'closed' group 118 to 187 ml/100 g/min. This experimental model stimulates the pathophysiologic perturbations in the parenchyma induced by cerebral arteriovenous malformations. It supports the findings that non-infarctional hypoperfusion can result from arteriovenous malformations and that following extirpation of arteriovenous malformations hyperemia may ensue.

Hemodynamic and metabolic effects of cerebral arteriovenous malformations studied by positron emission tomography

Seventeen patients with an intracranial arteriovenous malformation were studied with positron emission tomography. Cerebral blood flow, cerebral blood volume, oxygen extraction fraction, and glucose and oxygen metabolism were evaluated in both hemispheres, excluding the area of the malformation itself. Patients were divided into three groups according to the size of their malformation, and results obtained were compared with studies in healthy volunteers. The glucose metabolism was significantly (p less than 0.01) decreased in the ipsilateral hemisphere in all patients. The cerebral blood volume was significantly increased (p less than 0.001) ipsilaterally in the three groups, and contralaterally in patients with medium- and large-sized arteriovenous malformations. The cerebral blood volume to cerebral blood flow ration, an index of vascular mean transit time, was significantly increased (p less than 0.005) ipsilaterally in patients with medium- and large-sized malformations and contralaterally in patients with large ones. Cerebral blood flow, oxygen extraction fraction, and oxygen metabolism were within the normal range bilaterally in all three groups, but oxygen extraction fraction tended to be higher in patients with larger lesions. The lack of significant change in oxygen metabolism suggests that oxygen metabolism in cortical areas remote from the arteriovenous malformation has been maintained by compensatory hemodynamic mechanisms. These data reveal widespread metabolic and hemodynamic consequences of arteriovenous malformations and suggest that they are associated with impairment of glucose metabolism, both in ipsilateral regions remote from the lesion and in the contralateral hemisphere in patients with large lesions.