Journal of Neuro-Ophthalmology, June 1987, Volume 7, Issue 2

Glucose utilization of visual cortex following extra-occipital interruptions of the visual pathways by tumor. A positron emission tomography study.

To assess the effect of extra-occipital lesions on the local cerebral glucose utilization of the primary and associative visual cortex, 29 patients were studied in the unstimulated state by positron emission tomography and [18F]2-deoxyglucose. Quantitative Goldmann perimetry was done in each patient at the time of the positron emission tomographic study. Nine patients showed homonymous defects, either hemianopsia or quadrantanopsia, whereas nine patients had heteronymous defects. Eleven control subjects, free of any neurological symptoms and with normal visual fields, were also studied with [18F]2-deoxyglucose positron emission tomography. In the normal control subjects and in patients with a heteronymous defect, left-to-right differences in the local cerebral metabolic rate for glucose of the visual cortex varied less than 10%. In patients with hemianopic defects, differences ranged from 8 to 38%, with the hypometabolic cortex always contralateral to the field defect. In patients with quadrantanopic defects, the visual cortex contralateral to the field defect demonstrated differences from 14 to 24% above and below the calcarine fissure, the cortex that received greater input from the affected field being hypometabolic.

!oUTI1Ill of Clinical Neuro-ophtlullmology 7(2): 63-68, 1987. ~ 1987 Raven Press, New York Glucose Utilization of Visual Cortex Following Extra-occipital Interruptions of the Visual Pathways By Tumor A Positron Emission Tomography Study Donn S. Fishbein, M.D., Georgia Antonakou Chrousos, M.D., Giovanni Di Chiro, M.D., Robert E. Wayner, M.D., Nicholas J. Patronas, M.D., and Steven M. Larson, M.D. To assess the effect of extra-occipital lesions on the local cerebral glucose utilization of the primary and associa­tive visual cortex, 29 patients were studied in the un­stimulated state by positron emission tomography and [lsF]2-deoxyglucose. Quantitative Goldmann perimetry was done in each patient at the time of the positron emission tomographic study. Nine patients showed homonymous defects, either hemianopsia or quadran­tanopsia, whereas nine patients had heteronymous de­fects. Eleven control subjects, free of any neurological symptoms and with normal visual fields, were also studied with [l8FJ2-deoxyglucose positron emission to­mography. In the normal control subjects and in pa­tients with a heteronymous defect, left-to-right differ­ences in the local cerebral metabolic rate for glucose of the visual cortex varied <10%. In patients with hemi­anopic defects, differences ranged from 8 to 38%, with the hypometabolic cortex always contralateral to the field defect. In patients with quadrantanopic defects, the visual cortex contralateral to the field defect demon­strated differences from 14 to 24% above and below the calcarine fissure, the cortex that received greater input from the affected field being hypometabolic. Key Words: Brain tumor-Cerebral glucose utilization -Positron emission tomography-Visual cortex. From the Neuroimaging Section, National Institute of Neuro­logical and Communicative Disorders and Stroke (D.S.B., G.D.C., R.E.W., N.].P.), the Neuro-ophthalmology Section, National Eye Institute, and the Center for Sight, Georgetown University (G.A.c.), and the Nuclear Medicine Department, Clinical Center (S.M.L.) of the National Institutes of Health, Bethesda, Maryland. Address correspondence and reprint requests to Giovanni Di Chiro, M.D., National Institutes of Health, Building 10, Room . 1C451, Bethesda, MD 20205, U.S.A. 63 Several techniques exist to measure cerebral me­tabolism in local regions such as the visual cortex (1). These techniques may utilize biochemical anal­ysis of tissues, radioisotope incorporation, an ox­ygen electrode, arteriovenous differences, or mea­surement of cerebral blood flow. The [14C]2-deoxy­glucose technique developed by Sokoloff et at (2) allows the determination of the local cerebral met­abolic rate for glucose (LCMRGlu) after killing of the experimental animal, sectioning of the tissue of interest, and autoradiography. The technique relies upon the intracellular trapping of a labeled glucose analog, which reaches a metabolic dead end after undergoing the first step (6-phosphory­lation) of glycolysis. An adaptation of the 2-deoxy­glucose technique alows measurement of the LCMRGlu in intact subjects, including humans. Using the Neuro-PET positron emission tomo­graph and 2-deoxyglucose labeled with a positron emitter, 18F, the LCMRGlu can be determined in the brain to a resolution of 6-7 mm in plane and 10 mm longitudinally. We used the technology just described to eval­uate LCMRGlu for the occipital cortex in patients with homonymous and heteronymous visual field defects caused by brain tumors. Recent studies describing the metabolic dysfunction of the occip­ital cortex in relation to specific visual field defects (3,4) contain cases of occipital lobe ischemia or neoplasia. The present study was designed to in­clude patients with visual field defects caused by tumors remote to the occipital cortex. This allows one to assess the effect of visual pathway inter­ruption upon occipital metabolism without having 64 D. S. FISHBEIN ET AL. to consider lesions that suppress metabolism by virtue of their destruction of the occipital cortex. MATERIALS AND METHODS Eighteen patients, 10 men and 8 women be­tween the ages of 22 and 71, were studied be­tween May 1981 and September 1984. Patients were eligible for the study if they had a stable vi­sual field defect due to an extra-occipital lesion. Patients with heteronymous defects usually had sellar or suprasellar lesions. Patients with homon­ymous defects had neoplastic lesions of the tem­poral and parietal lobes, or suprasellar lesions with either demonstrated or presumed retrochias­matic involvement of the optic pathways. Gold­mann perimetry was performed in every patient. All patients with visual field defects were studied with high-resolution x-ray computed tomography. A group of 11 individuals, 8 men and 3 women with no known neurological deficit and normal re­sults of visual field examination, were also studied to serve as a control group. All patients were studied in accordance with the appropriate guide­lines of the National Institutes of Health after giving informed consent. Following preliminary neurological, ophthalmo­logical, and radiological examinations, all patients were studied with (18F]2-deoxyglucose positron emission tomography. The procedure described by Phelps et al. (5) was used for the injection of (18F]2-deoxyglucose and subsequent blood sam­pling. Each patient received between 3.4 and 5.0 mG of isotope (5.0 mG is the maximum dose per­mitted at our institution). Scans were started -30 min postinjection. Patients were studied in the un­stimulated state, with the eyes patched and the ears plugged. The patients were instructed to re­frain from movement and speech, and to relax, but to avoid sleeping. Glucose metabolic rates were calculated as previously described (6,7). Typically, three scans oriented parallel to the canthomeatal line were obtained, each lasting 10 min. The first scan was performed between 10 and 20 mm above the canthomeatal line. In order to assure visualization of the entire occipital cortex, scans were repeated at progressively lower levels until the cerebellum was included in the field of view. In order to insure good axial sampling, each scan after the first was offset either 4 or 6mm from the preceding scan. Two methods were used to identify the striate cortex and its associated areas in region-of-interest analysis. The first method uses standard x-ray computed tomography and neuroanatomy atlases to establish correlations with positron emission to­mograms (8). The second method uses positron emission tomogram metabolic data to identify the striate cortex. For example, Bosley et al. (3) noted increased metabolic activity after visual stimula­tion in the medial occipital lobe, from the level of the tentorium cerebelli inferiorly to the atria of the lateral ventricles superiorly. The anterior margin of the striate cortex was at the same level as the splenium of the corpus callosum; the posterior margin formed the posterior pole of the hemi­sphere. Phelps et al. (9) chose regions for study in the striate cortex by outlining medial occipital cortex in which the metabolic rate increased after visual stimulation. In this study, these two methods were com­bined to increase the accuracy in the placement of regions of interest. Circular regions of interest be­tween 5 and 10 mm in diameter were placed bilat­erally in the macular and peripheral sections of Brodmann areas 17 (area striata), 18 (area para­striata), and 19 (area peristriata). Regions to be measured were chosen above and below the cal­carine fissure, respecting anatomic correspon­dence and encompassing uniform regions. X-ray computed tomograms were consulted to assist in anatomic localization as described. Measurements were also taken in extra-occipital locations for inter-comparison with control values. Region-of-interest measurements yield an abso­lute value for the LCMRGlu and a value for the relative uniformity of the pixels enclosed within the region. A metabolic rate for a region was com­pared with that of the homologous contralateral region. A percentage difference was calculated as LCMRGlul - LCMRGlur - 1 RESULTS In the group of 11 normal control subjects, the . average right-to-Ieft differences in glucose metabo­lism measured in Brodmann areas 17, 18, and 19 were 8 ± 3%, 10 ± 4%, and 9 ± 5%, respectively. Nine patients with nonhomonymous defects demonstrated left-to-right differences in glucose metabolism similar to the control group (area 17: 6 ± 4%, area 18: 10 ± 9%, area 19: 6 ± 4%). Nine patients evaluated had homonymous de­fects. Six patients were hemianopic: three right­sided and three left-sided. Three patients were quadrantanopic: one with a right inferior defect, one with a right superior defect, and one with a left superior defect. The greatest right-to-Ieft dif­ference in this group was seen within the macular CORTICAL GLUCOSE METABOLISM WITH VISUAL FIELD DEFECTS 65 representation in Brodmann area 17. The average difference was 20 ± 7%, with values ranging from 11 to 38%. The right-to-Ieft difference in Brod­mann area 18 was 10 ± 4%, and in Brodmann area 19,8 ± 5%. In studying quandrantanopsias, it was difficult to separate the cortical representations of the su­perior and inferior visual fields, because the reso­lution of the scanner used is lower in the longitu­dinal plane than the transverse plane. An attempt was made to accomplish this by sampling the vi­sual cortex on the lowest and highest slice in which it was visualized. Using this method in the three patients with quadrantanopsias, left-to-right differences of 14, 16, and 24% were seen between the macular representations with Brodmann area 17. Pertinent clinical data and metabolic rates are summarized in Table 1. CASE REPORTS The following reports are representative of the patients with significant left-to-right differences in the glucose metabolic rate of the visual cortex. Corresponding positron emission scans, relevant x-ray computed tomography scans, and visual fields are shown in Figs. 1 and 2. Case 1 At age 39 a left-handed man began complaining of vague visual disturbances. At that time an oph­thalmological examination, including visual fields, showed normal results. Six months later he devel­oped a right hemianopsia. X-ray computed tomog­raphy revealed a mass in the area of the left chiasma, hypothalamus, and optic tract area. The patient underwent craniotomy and excision of a nodule in the region of the lamina terminalis and anterior hypothalamus, which on histological ex­amination was found to be a low-grade astrocy­toma. A year later he underwent a second craniotomy because of evidence of enlargement of the tumor on x-ray computed tomography and worsening of his visual field examination. During the procedure the optic chiasm was dissected free from scar tissue, and a piece of tantalum foil was placed over the left optic nerve as a shield against planned radiation therapy. The patient received a course of proton beam irradiation (5,600 rads), which was effective in reversing the changes. Five years later, after deterioration of his visual acuity in the left eye (20/40) and a new deteriora­tion of his visual fields, the patient was given ste­roids and received additional radiation therapy. Since that time his condition has remained stable. Sixteen years later his ophthalmological examina­tion shows the best corrected visual acuity in the right eye to be 20/20, and in the left eye, 20/50. His visual fields show a right incongruous hemi­anopsia. There is an efferent pupillary defect in the left eye, and bilateral optic atrophy (1 + pallor right eye, 3+ pallor left eye). Recent x-ray com­puted tomograms show an enhancing lesion at the base of the left frontal lobe, an area of left frontal TABLE 1. Summary of clinical and metabolic data for patients with homonymous and heteronymous field defects LCMRGlu left:right differences ("!o) Patient Sex Age Visual fields eNS lesion (macular) (peripheral) (area 18) (area 19) 1 M 45 R homonymous hemianopsia Hypothalamic glioma 0.21 0.12 0.17 0.08 2 F 40 L homonymous hemianopsia Pituitary adenoma 0.38 0.08 0.04 0.15 3 F 25 L homonymous hemianopsia R parietal glioma 0.17 0.01 0.07 0.02 4 F 55 R homonymous hemianopsia Hypothalamic glioma 0.19 0.05 0.12 0.03 5 M 68 L homonymous hemianopsia R frontoparietal glioma 0.16 0.07 0.06 0.04 6 F 41 R homonymous hemianopsia Suprasellar glioma 0.11 0.01 0.12 0.12 7 F 55 R homonymous inferior Pituitary adenoma 0.14" 0.20 0.16 0.03 quadrantanopsia 8 M 66 R homonymous superior L temporal glioma 0.24b 0.17 0.12 0.11 quadrantanopsia 9 M 44 L homonymous superior R parietotemporal glioma 0.16b 0.12 0.08 0.11 "quadrantanopsla 10 M 64 Bitemporal hemianopsia Pituitary adenoma 0.07 0.08 0.17 0.07 11 M 26 Bitemporal hemianopsia Pituitary adenoma 0.10 0.08 0.12 0.05 12 F 43 Bitemporal hemianopsia Suprasellar ependymoma 0.04 0.04 0.26 0.03 13 M 53 Bitemporal hemianopsia Pituitary adenoma 0.12 0.03 0.00 0.13 14 F 27 Bllemporal hemianopsia Pituitary adenoma 0.08 0.10 0.19 0.02 15 M 68 Bitemporal hemianopsia Suprasellar glioblastoma 0.10 0.11 0.13 0.11 16 M 47 Bitemporal hemianopsia Pituitary adenoma 0.02 0.24 0.03 0.06 17 M 48 Bitemporal hemianopsia Pituitary adenoma 0.00 0.01 0.00 0.Q1 18 M 31 Bitemporal hemianopsia Pituitary adenoma 0.03 0.04 0.02 0.03 " Supracalcarine cortex only. b Infracalcarine cortex only. JClin Neuro-ophthalmol, Vol. 7, No.2, 1987 66 D. S. FISHBEIN ET AI. FIG. 1. Noncontrast x-ray computed tomogram (A), quantitative Goldmann perimetry (B), and positron emission tomogram (C) of patient 1, with right hom­onymous hemianopsia (B) secondary to previously operated astrocytoma involving left chiasma, left an· terior hypothalamus, and left optic tract. Note tumoral calcification (arrow) and surgical defect (A), and hy­pometabolism (arrow) of left visual cortex (C). surgical ablation, and the absence of edema. Nu­clear magnetic resonance scanning shows ventric­ular dilation and periventricular edema. (lsF]2-De­oxyglucose positron emission tomography dem­onstrates hypometabolism of the primary and associative visual areas in the occipital cortex on the left side (Fig. 1). Case 2 At age 22 a right-handed woman noted a gradual deterioration of visual acuity in her right eye over a period of 6 months. She saw an oph­thalmologist, who treated her with steroids for 3 months without improvement. A fewmoi1:ths~~ . h"',' be~t corrected visual acuity was 20/200 wi.th 2+ optic pallor in the right eye, and visual acuity was 20/20 with trace optic pallor in the left eye. Her visual fields showed a left incongruous hemi­anopsia. X-ray computed tomography done at that time revealed a pituitary mass, for which she un­derwent bifrontal craniotomy with decompression and subtotal removal of a cystic chromophobe ad­enoma. She did well for 17 years until she was found by x-ray computed tomography to have some enlargement of the tumor, and was given bromocryptine. She responded to the treatment, and her condition has been stable ever since. Her recent ophthalmological examination re­vealed visual acuity of 20/30 in the left eye and count fingers at 1 m in the right eye, Her visual fields showed the same incongruous left hemi- CORTICAL GLUCOSE METABOLISM WITH VISUAL FIELD DEFECTS 67 FIG. 2. Quantitative Goldmann perimetry (A) and pos­itron emission tomogram (8) of patient 2, with left homonymous hemianopsia (A) secondary to pre­viously operated chromophobe adenoma with poste­rior extension involving retrochiasmatic structures. Note hypometabolism of the right visual cortex (8, arrow). anopsia. There was an afferent pupillary defect in the right side, and bilateral optic atrophy (3 + pallor right eye, 1+ pallor left eye). X-ray com­puted tomography scanning at the same time as positron emission tomography revealed a cystic lesion in the sella with a rim of enhancement, but it did not show retrochiasmatic extension of tumor. [l8F]2-Deoxyglucose positron emission to­mography showed right occipital lobe hypometab­olism in the primary and associative visual areas (Fig. 2). DISCUSSION The primary visual cortex has one of the highest baseline glucose utilization of the regions that can be resolved by positron emission tomography. A large number of neurons performing a similar task are concentrated in a small anatomic area. Because it is a primary sensory area, its metabolic rate is reasonably easy to manipulate through sensory stimulation or deprivation. These facts have com­bined to make the visual system, especially the vi­sual cortex, a favorite subject of investigation using positron emission tomography. The functional organization of the primary and associative visual cortex had been studied by Hubel and Wiesel (10-12) and Hubel (13) before determination of glucose metabolic rates was common practice. Kennedy et al. (14) applied the [14C]2-deoxyglucose technique to the study of the primary visual system of the Macaque monkey, finding the greatest metabolic activity in layer IV, the termination of the geniculocortical pathway. Phelps et al. (9) and Phelps and Kuhl (15) studied the response of the human visual cortex to stimu­lation and deprivation, including patients with homonymous hemianopsias. Their hemianopic patients had a normal anatomic appearance of the occipital cortex by x-ray computed tomography. Bosley et al. (3) reported eight patients with ischemic lesions of the occipital cortex and optic radiation studied with positron emission tomog­raphy. In the majority of patients in this study, an occipital lesion was clearly identifiable on com­puted tomographic scans, increasing the expecta­tion of identifying a hypometabolic region by posi­tron emission tomography. Kiyosawa et al. (4) re­ported studies of eight patients with hemianopsia secondary to ischemic lesions of the occipital cortex. The present study attempted to assess the baseline (nonstimulated) metabolic function of the denervated occipital cortex, eliminating as much as possible factors other than remote interruption of the visual pathways. Previous investigators have noted the difficulty in precisely identifying regions within the occipital cortex (3,9). In order to study patients with quad­rantanopsias, it was necessary to further divide the calcarine cortex into areas representing the su­perior and inferior visual fields. By using interca­lated slices centered 4 mm apart, it was possible to identify the lower and upper limits of the calcarine cortex and to deduce that they must lie on oppo­site sides of the calcarine fissure. Patients with nonhomonymous defects did not show significant differences in occipital lobe glu­cose metabolic rates from side to side, or overall in comparison with normal patients. Although no difference would be expected between or within hemispheres at the measurement resolution of positron emission tomography in a patient with a bitemporal hemianopsia, the overall metabolic rate for visual cortex might be expected to be lower. A region-of-interest measurement simply reflects the sum of tracer activity occurring within its bounds. , C/ill Nt'llrO-o!,!Ithnllllol, Vol. 7, No.2. 1987 68 D. S. FISHBEIN ET AI. If fewer neurons within a region concentrate the tracer (Le., metabolize glucose), the activity mea­sured within the region should be decreased. In bitemporal hemianopsia, approximately one half of the synaptic input to the occipital cortex is lost. However, measured glucose metabolism does not decrease by half. Neurons of the visual cortex receive varying contributions from both eyes. Loss of one half of synaptic input does not translate into denervation of one half of visual cor­tical neurons. In addition, positron emission to­mography measures tracer in vascular, interstitial, and intracellular compartments. Tracer concentra­tion in the first two compartments is relatively in­dependent of neuronal activity, and studies are performed in a manner to minimize the activity measured in the vascular and interstitial compart­ments. Also, neurons do not constitute the only cell type within the visual cortex. The glucose me­tabolism of glial cells is measured along with that of neurons. Finally, visual cortex, including pri­mary visual cortex, receive synaptic input from other than the optic radiations (16). The major finding of this study is that the result of loss of trophism upon the visual cortex by le­sions outside the visual cortex causing homony­mous hemianopic or quadrantanopic defects is demonstrated in the primary visual cortex by high-resolution PSF]2-deoxyglucose positron emission tomography, even when such lesions cannot be demonstrated by other imaging modali­ties, such as x-ray computed tomography. The in­ability to demonstrate changes in the associative visual cortex with homonymous lesions may in part be due to the fact that all patients were studied in the unstimulated state. The ability of a functional imaging modality to demonstrate this is due in part to the unique structure of the visual pathways. Positron emission tomography allows the dem­onstration of physiological change in visual cortex caused by an extra-occipital lesion where no change can be demonstrated by other imaging mo­dalities. The major change demonstrated in this study is within the macular representation of Brodmann area 17, an area that receives a large concentration of fibers from a relatively small area of the retina. REFERENCES 1. Sokoloff L. Circulation and energy metabolism of the brain. In: Siegal GJ, Albers RW, Agranoff BW, Katzmann R, eds. Basic neurochemistry. 3rd ed. Boston: Little, Brown, 1981:471-6. 2. Sokoloff L, Reivich M, Kennedy C, et al. The [l4C]deoxy­glucose method for the measurement of local cerebral glu­cose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J Neurochem 1977;28:897-916. 3. Bosley TM, Rosenquist AC, Kushner M, et al. Ischemic le­sions of the occipital cortex and optic radiations: positron emission tomography. Neurology 1985;35:470-84. 4. I<iyosawa M, Mizuno K, Hatazawa J. Metabolic imaging in hemianopsia using positron emission tomography with [lSF]2-deoxyfluoroglucose. Am J OphthalmolI986;101:310-9. 5. Phelps ME, Huang SC, Hoffman EJ, Selin C, Sokoloff .L, Kuhl DE. Tomographic measurement of local cerebral glu­cose metabolic rate in humans with [18F]2-fiuoro-2-deoxy­o- glucose: validation of method. Ann NeuroI1979;S:371-88. 6. Di Chiro G, DeLaPaz RL, Brooks RA, et al. Glucose utiliza­tion of cerebral gliomas measured by [l8F}fluorodeoxygiu­cose and positron emission tomography. Neurology 1982;32:1323-9. . 7. Huang SC, Phelps ME, Hoffman EJ, Sideris K, Selin q, Kuhl DE. Noninvasive determination of local cerebral met­abolic rate of glucose in man. Am JPhysioI1980;238:E69-82. 8. Mazziotta JC, Phelps ME, Miller J, Kuhn DE. Tomographic mapping of human cerebral metabolism: normal unstimu­lated state. Neurology 1981;31:503-16. 9. Phelps ME, Mazziotta JC, Kuhl DE, et al. Tomographic mapping of human cerebral metabolism: visual stimulation and deprivation. Neurology 1981;31:517-29. 10. Hubel DH, Wiesel TN. Receptive fields and functional ar­chitecture in two nonstraite visual areas (18 and 19) of the cat. J NeurophysioI1965;28:229-89. 11. Hubel DH, Wiesel TN. Sequence regularity and geometry of orientation columns in the monkey striate cortex. JComp NeuToI1974;158:267-93. 12. Hubel DH and Wiesel TN: Uniformity of monkey striate cortex: a parallel relationship between field size, scatter, and magnification factor. JComp Neuroll974;158:295-305. 13. Hubel DH. The visual cortex of normal and deprived monkeys. Am Sci 1979;67:532-43. 14. Kennedy C, Des Roseiers MH, Sakurada 0, et al. Metabolic mapping of the primary visual system of the monkey by means of the autoradiographic 14-C-deoxyglucose tech. nique. Proc Natl Acad Sci 1976;73:4230-4. 15. Phelps ME, Kuhl DE: Metabolic mapping of the brain's re­sponse to visual stimulation: studies in humans. Science 1981;211:1445-8. 16. Brodal A. Neurological anatomy in relation to clinical medicine. 3rd ed. New York: Oxford University Press, 1981:9-35, 578-98.