Recovery of Vision after Ischemic Lesions: Positron Emission Tomography Thomas M. Bosley, MD,"f Robert Dann, BS,t Frank L. Silver, MD," Abass Alavi, MD,? Michael Kushner, MD," John B. Chawluk, MD," Peter J. Savino, MD,S :Robert C. Sergott, MD,I Norman J. Schatz, MD,$ and Martin Reivich, MD" We used ['8F)fluoro-2-deoxyglucose positron emission tomography (PET) to study serial changes in the local cerebral metabolic rate of glucose in 5 patients with ischemic lesions of the posterior afferent visual system causing homonymous visual field defects. All 5 patients had striking impairment of glucose metabolism in the striate cortex shortly after ictus. In 2 patients, visual field defects abated, and repeat PET scans showed reduced size of the metabolic lesion and improvement of striate metabolism. Three patients did not recover vision, and repeat PET scans did not show improvement. Patients who recovered vision had ischemic lesions outside the occipital lobe, while those who did not experience improvement had primary damage to the occipital lobe itself. Bosley TM, Dann R, Silver FL, Alavi A, Kushner M, Chawluk JB, Savino PJ, Sergott RC, Schatz NJ, Reivich M: Recovery of vision after ischemic lesions: positron emission tomography. Ann Neurol 21:444-450, 1987 Positron emission tomography (PET) has been used to study clihical-metabolic correlations in patients who have suffered ischemic lesions of the occipital lobe or optic radiations [4, 12, 153. These studies have shown that: (1)large, dense metabolic defects may be present prior to the appearance of abnormalities on computed tomograms (CT); (2) deafferentation due to ischemia in the optic radiation causes mod_estly reduced glucose uptake in undamaged occipital lobes; ( 3 ) metabolic abnormalities in the occipital cortex are generally larger in size than the corresponding anatomic defect shown on CT; (4) areas distant from the ischemic lesion (visual association cortex, thalamus) may have reduced metabolism; and (5) the correlation between visual field defects and metabolic abnormalities seen on PET is excellent. The visual system is ideal for the study of recovery of function after stroke because the degree of clinical deficit can be defined more objectively than in other cortical lesions. We obtained serial PET scans on 5 patients with ischemic lesions of the posterior afferent visual system that caused dense homonymous visual field defects. We report here the clinical, anatomic, and cerebral metabolic characteristics of the 2 patients who recovered vision, compared to those patients with unchanging visual deficits. From the "Cerebrovascular Research Center of the Department of Neurology, and the tDepartment of Radiology, University of Pennsylvania;and the SNeuro-ophthalmology Eye Hospital, Philadelphia, PA. 444 Methods The study included 5 patients with dense homonymous visual field defects caused by ischemic lesions of the posterior afferent visual system. All patients had best-corrected visual acuity of 20/30 or berter in at least one eye. Neuroophthalmological evaluation, including Goldmann kinetic perimetry, was performed within one week of each PET scan. High-resolution CT was performed within one month of PET scanning. Informed consent was obtained from each patient, and all procedures were approved by the Committee on Studies Involving Hiunan Beings, of the University of Pennsylvania. Each patient was studied with ['8F~fluoro-2-deoxyglucose (FDG) PET to determine the local cerebral metabolic rate of glucose (ICMRglc) shortly after the ictus and again one to ten months later. Patients 1, 2 , 4, and 5 were studied with eyes open, while Patient 3 was studied with eyes closed in order to accentuate metabolic activity related to spontaneous visual hallucinations in the dkcted hemifield. In all cases the scanning protocol was identical on initial and repeat PET studies, A modified version of the PE'IT V system [261 was employed to obtain brain images of lCMRglc using Phelps and colleagues' 1141 modification of the method of Reivich and associates 1191. Characteristics of the PET system and the protocol for data collection have been described previously 1131. The size of metabolically depressed areas was assessed visually on each PET. In addition, metabolic regions of interest Received May 19, 1986, and in revised form Aug 21. Accepted for publication Aug 29, 1986. Address correspondence to I)r Bosley, Neuro-ophth&-,ology Service, Wills Eye Hospital, Ninth and Walnut Sts, Philadelphia, PA 19107. Table 1 . Clinical Presentations Patient No. 1 2 3 4 5 Age (yr)/ Sex Visual Field Defect 531M 60IM 841F RHH LHH LHH RSQ LHH 67lM 641F Ictus to PET 1 (days) Ictus to PET 2 (mo) 7 5 19 3 10 1 1 3 4 9 equations of Sokoloff and colleagues [22}, upon which the deoxyglucose method is based. Ischemia may alter the rate constants for forward and reverse transport of FDG between plasma and tissue and for phosphorylation and dephosphorylation of FDG [ l o ] .Likewise, the value for the lumped constant may be altered in ischemia 13 1J but may return toward normal during reperfusion "97. Our subjects were not studied acutely, and conditions were most likely close to those found in the nonischemic state. Residual perturbations in the involved cerebral cortices are, nevertheless, a potential source of uncertainty. ~ PET = positron emission tomography; R = right; L = left; HH = homonymous hemianopia; SQ = superior quadrantanopia; M = male: F = female. (ROI) were placed, using a computerized overlay system based on normal human anatomy @I, over the anterior striate cortex, posterior striate cortex, visual association cortex, and thalamus, in the fashion described previously [4]. Hemispheric midline and X-Y scaling were adjusted visually according to the individual patient's anatomy on CT. Quantitative ICMRglc data were extracted from ROIs and compared with homologous regions of the contralateral hemisphere using the laterality index (LI): LI = ICMRglc(R) - ICMRglc(L) ICMRglc(R) + ICMRglc(L) 200 Positive values for the laterality index indicate relatively less lCMRglc on the left ( L in equation), and negative numbers indicate relatively less lCMRglc on the right (R in equation). Laterality index data were compared qualitatively among patients because of the small sample size. Some potential sources of error must be recognized. Metabolic values for ROIs were averaged over multiple slice levels to compensate for the possibility that tilting of the head within the scanner may induce artifactual asymmetries. Conditions of ischemia may affect factors in the operational Results The patient group was composed of 3 men and 2 women, aged 53 to 84 years (Table 1).No subject was somnolent or stuporous at the time of PET scanning. Patients were divided into two groups according to clinical criteria alone. Patients 1 and 2 experienced improvement in their visual field between PET scans. Patients 3, 4, and 5 had no change in their visual field defect (Table 2). There was no significant difference between these two groups in general medical history, size of visual field defect, laterality of visual loss, or time from ictus to first or second PET scan. The patients who experienced improvement were younger than those who did not, and the clinical course of their neurological deficits differed. Patient I This 53-year-old right-handed man was seen after the sudden onset of a dense right homonymous hemianopia, alexia without agraphia, color anomia, short-term memory loss, and subtle weakness of the right upper extremity. More clinical details are described elsewhere [21}. CT scan showed discrete lucencies of the left hemisphere that involved the lateral geniculate body, splenium of the corpus callosum, and anterior limb of the internal capsule (Fig 1). PET scanning Table 2. Results of Positron Emission Tomography Patient No. 1 2 3 4 5 Scan 1 2 Difference Size of PET Defect Smaller 1 2 Difference 1 2 Difference 1 2 Difference 1 2 Difference Smaller Slightly larger Same Same Laterality Index CALCA CALCP VA 33.5 24.0 9.5 - 13.1 - 8.4 4.7 - 8.7 - 19.2 - 10.5 42.0 43.6 - 1.6 - 32.6 22.0 18.3 3.7 - 14.8 - 1.8 13.0 - 17.5 - 29.1 - 11.6 19.2 34.8 - 15.6 - 43.5 - 46.7 - 3.4 35.3 30.1 5.2 - 16.2 - 13.9 2.3 - 13.9 9.0 22.9 16.6 24.6 - 8.6 - 25.4 - 34.0 - 8.6 - 36.4 - 3.8 Thalamus 20.4 13.6 6.8 2.1 - 1.1 - 3.0 - 4.9 - 1.1 3.8 1.5 - 3.7 - 2.2 - 17.2 - 5.3 11.9 - CALCA = anterior calcarine cortex; CALCP = posterior calcarine cortex; VA = visual association cortex. Bosley et al: PET of Visual Recovery 445 was unchanged, but repeat PET scan documented reduced size and amplitude of the metabolic asymmetry. Patient 2 This 60-year-old man was seen after the sudden onset of left homonymous hemianopia, confusion, and left hemineglect. CT scanning documented a right posterior parietal hypodensity with loss of sulci. Initial PET scanning showed a moderately large region of metabolic asymmetry with relatively less glucose uptake in an area involving the right posterior temporal lobe, posterior parietal lobe and occipital lobe (Fig 2). Three months later, the results of Goldmann kinetic perimetry were normal while confrontation visual fields revealed a residual mild left homonymous red desaturation. Repeat PET scanning showed a smaller area of metabolic asymmetry that was restricted predominantly to the posterior parietal lobes. Patient 3 This 84-year-old woman was seen after the sudden onset of headache and left homonymous hemianopia, with visual hallucinations in the blind hemifield consisting of flashing golden lights. CT scan showed a relatively small ischemic lesion involving the optic radiations and the occipital lobe on the left. There was an area of modest metabolic asymmetry with lower glucose uptake in the right posterior occipital lobe seen on the initial PET scan (Fig 3). Over the next ten months, her visual halllucinations gradually disappeared, but the left homonymous hemianopia persisted while her right visual field remained intact. Repeat PET scanning demonstrated a larger area of more severe metabolic asymmetry involving the posterior parietal and occipital lobes. This scan documented a new area of relative hypometabolism in the left posterior parietal lobe, resulting in a positive laterality index for the visual association areas (see Table 2). A repeat CT scan was unchanged. Fig I . Computed tomographic (CT) scan (top), initial positron emission tomographic (PET) scan (middle),and repeat PET scan (bottom) of Patient I who had a dense right homonymom hemianopia and alexia without agraphia. Symptoms largely resolved between the two PET scans. C T scanning was performed at -20 to the orbitomeatal line to conform with the axis of the PET scan. Scans shown are at approximately the same tomographic leveh, with the lower image on the kfi. Large arrows in each image point to areas of glucose uptake corresponding to kjit and right striate cortex; small arrows point to areas appropriate for left visual association cortex. Glucose uptake in ldt visual association and striate cortices is rekztiuely improved in the second PETscan. (L = 1ejt;R = rigbt.) performed shortly after presentation demonstrated an extensive area of metabolic asymmetry with relatively less glucose uptake into the left occipital lobe, posterior parietal lobe, mesial temporal lobe, and thalamus. Symptoms resolved gradually, and after four months he had a mild right homonymous dyschromatopsia that was denser superiorly than inferiorly; his reading was slow but correct; he had no difficulty with color naming or short-term memory; and there was a mild increase in muscle tone on the right. Repeat CT scan 446 Annals of Neurology Vol 21 N o 5 May 1987 Patient 4 This 67-year-old man was seen after the sudden onset of a right superior quadrantanopia CT scanning documented a left inferior occipital lobe infarct. Initial PET scanning showed an area of metabolic asymmetry posteriorly with relatively less glucose uptake in the left inferior occipital lobe without involvement of visual association areas or thalamus. His visual field remained the same over the next month, and a repeat PET scan wm unchanged. Patient 5 This 64-year-old woman was seen after the sudden onset of a dense left homonymom hemianopia. CT scanning showed an infarct in the right occipital lobe. Initial PET scanning revealed striking metabolic asymmetry of an area involving the striate cortex, visual association cortex, and thalamus, with relatively less glucose uptake on the right. One month later, the visual field defect was unchanged and repeat PET scanning documented continued metabolic asymmetry. Discussion All 5 patients were iin the vasculopathic age range, and the diagnosis of stroke involving predominantly the Fig 2. Initial (top)and repeat (bottom)positron emission tomographic (PET)scan of Patient 2 with resolving l4t homonymous hemianopia due to a smll ischemic lesion of the right posterior parietal lobe. The lwer imge is on the lejk Large arrows in each image point to areas of glucose uptake corresponding to ldt and right striate cortex; smll open arwnus point to area c o w sponding to infarct in the right visual association area. Note partial recwm of metabolism in the right striate cortex between PET scans. ( L = kfi; R = right.) posterior afferent visual system was supported in each patient by the sudden onset of the deficit, the character of the neurological symptoms, and the location and appearance of the CT abnormalities. Two patients recovered vision in areas of homonymous visual field loss, while homonymous visual field defects in the other 3 patients did not change. The recovery of vision after stroke causing homonymous visual field loss is well recognized 12, 3 , 18, 25, 291, although the neurobiological mechanisms involved in recovery are controversial and poorly understood [ 5 , 17, 23, 27, 321. PET scans of all 5 patients revealed hypometabolic areas that were appropriate for the visual deficit and corresponded to lesions seen on CT scans 141. Serial PET scans documented partial correction of the metabolic asymmetries in the striate and posterior parietal cortex of the 2 patients who recovered vision. In both cases, the area of hypometabolism decreased strikingly (see Figs 1, 2) and the magnitude of metabolic asymmetry also decreased. The size of lesion on repeat study was probably the more reliable index of improvement, because metabolic rates on repeat PET scans of even normal individuals vary considerably C71 and because decreased lesion size may have artificially reduced laterality indices by volume averaging [ll]. Repeat PET scans on the 3 patients who did not recover vision revealed no improvement in either the size of the metabolic lesion or the laterality index in affected areas of striate and visual association cortex. Neither the size of the metabolic lesion nor the laterality indices correlated with visual recovery. The area of decreased glucose uptake in Patient 1 was considerably larger than that of any other patient, while the affected area in Patient 2 was larger than that of Patient 4 but smaller than that of Patient 3. Laterality indices for striate and visual association cortex in Patient l were greater than those of Patient 3, but were less than those of Patients 4 and 5. Conversely, comBosley et ak PET of Visual Recovery 447 Fig 3. lnitial (top) and repeat (bottom)positron emission tomographic scan of Patient 3 with dense left homonymous hemianopia due to stroke in the right occipital lobe. The lower image is on the left. Large arrows in each image point to areas of&cose uptake appropriatefor right and ldt striate cortex. Unformed visual hallucinations in the blind left hemifield resolved between scans. Relative hypometabolism of the right striate cortex is increased in the repeat scan, with enlargement of the hypometabolic area corresponding to right visual association cortex (small open arrow). (L = left; R = right.) parison of repeat PET scans proved quite sensitive in individual patients, both to metabolic changes accompanying visual recovery and to persistent metabolic abnormalities when that recovery was absent or incomplete. Visual improvement in this small series of patients was best predicted by the location of the anatomic lesion documented by CT early in the clinical course. Patients 3 , 4, and 5 all had lesions causing direct damage to the occipital lobe, and these patients retained 448 Annals of Neurology Vol 2 1 No 5 May 1987 dense homonymous visual field defects. Patient 1 suffered ischemic damqge to the lateral geniculate body and deep white matter that caused deafferentation of the striate cortex, while the ischemic lesion of Patient 2 destroyed visual association cortex in the posterior parietal lobe. In both cases, primary visual cortex was spared direct ischemic damage, and in both cases vision recovered. This observation does not address the mechanism of visual recovery, but it does imply that striate cortex is necessary, if not wholly sufficient, for a full visual field. Visual recovery was not complete in Patients 1 and 2. In both cases, visilal fields were full to Goldmann kinetic perimetry several months after ictus, but the patients were aware of mild homonymous visual distortion and dyschromatopsia. This type of residual defect is not uncommon after incomplete damage to the posterior afferent visual system El, 201, although it is frequently overlooked on clinical examination. Abnor- mal regional glucose uptake of striate and visual association cortex also persisted in both patients. In fact, it is striking in Patient 1 that laterality indices improved only marginally. On repeat PET scanning the laterality indices in striate cortex were within the range noted previously in dense visual field loss due to striate deafferentation 14) and continued to be greater than those of Patient 3 with a total homonymous hemianopia. Therefore, neither residual visual function nor future visual recovery were accurately predicted on a single PET scan by lateratity indices in the range represented in this small series of patients. Patient 3 experienced unformed visual hallucinations (flashing golden lights and balls) in her blind left visual hemifield after ischemic damage that affected the contralateral optic radiations and occipital lobe. Visual hallucinations have been well described after a variety of cerebral lesions 16, 301, but they remain poorly understood because this uniquely human experience cannot be studied in experimental animals. This patient's homonymous hemianopia persisted, but her visual hallucinations gradually resolved prior to the second PET scan, which on several images revealed an enlarged hypometabolic area in the right posterior parietal cortex and a new hypometabolic area involving the left posterior parietal lobe (see Fig 3). Iaterality indices in the striate cortex were increased. Patient 3 had no symptoms, clinical signs, or CT evidence of a new lesion in the left hemisphere, and the cause of this unusual change in relative glucose uptake and its possible role in the disappearance of visual hallucinations remain uncertain. The primate striate cortex is the major cortical center for the evaluation of visual information and the dominant (but not exclusive) relay station for that information between the lateral geniculate and visual association areas in the lateral occipital and posterior parietal lobes 1321. Complete destruction of the striate cortex caused by ischemia may well result in total loss of vision in corresponding areas of homonymous visual field that is unlikely to recover. Incomplete damage to the striate cortex or damage to optic radiations or visual association cortex might allow the preservation or recovery of the visual field, possibly with some element of visual distortion or hemi-inattention. Clinical studies from the era prior to CT cannot address this hypothesis because the exact location of the ischemic lesion was usually unknown. Even pathological data are difficult to interpret because striate cortex is quite variable in extent and symmetry 116, 24, 28). It is hoped that careful clinical-anatomic correlations made by newer-generation CT and magnetic resonance imaging will be able to resolve these issues. Supported by U.S.P.H.S. Program Project Grant NS-14867-07. References 1. Bender MB, Furlow L T Phenomenon of visual extinction in homonymous fields and psychologic principles involved. Arch Neurol Psychiatry 53:29-33, 1945 2. Bergman PS: Cerebral blindness. Arch Neurol Psychiatry 78:568-584, 1957 3. Bogousslavsky J, Regli F, Van Melle G Unilateral occipital infarction: evaluation of the risks of developing bilateral loss of vision. J Neurol Neurosurg Psychiatry 46:78-80, 1983 4. Bosley TM, Rosenquist AC, Kushner M, et al: Ischemic lesions of the occipital cortex and optic radiations: positron emission tomography. Neurology 35:470-484, 1985 5. Campion J, Latto R, Smith YN: Is blindsight an effect of scattered hght, spared cortex, and near-threshold vision? Behav Brain 6:423-486, 1983 6. Cushing H: Distortions of the visual fields in cases of brain tumor. Trans Am Neurol Assoc, 1921, p 374 7. Cutler NR, Haxby JV, Duma R, et ak Brain metabolism as mekured with positron emission tomography: serial assessment in a patient with familial Alzheimer's disease. Neurology 35: 1556-1561, 1985 8. Dann R, Muehllehner G, Rosenquist A: Computer aided data analysis of ECT data. J Nucl Med 24232, 1983 use of the 2-deoxyglucose method of 9. zation in the abnormal brain: evaluation of the lumped constant during ischemia. Stroke 10483, 1979 10. Hawkins RA, Phelps ME, Huang SC, Kuhl DE: Effect of ischemia on quantification of local cerebral glucose metabolic rate in man. J Cereb Blood Flow Metab 1:37-51, 1981 11. Hoffman ET, Huang SC, Phelps ME: Quarititation in positron emission tomography. I. Effect of object size. J Comput Assist Tomogr 3:299-308, 1979 12 Kuhl DE, Phelps ME, Kowell AP, et al: Effects of stroke on Local cerebral metabolism and perfusion: mapping by emission computed tomography of "FDG and 13NH3. Ann Neurol 8~47-60, 1980 13 Kushner M, Alavi A, Reivich M, et al: Contralateral cerebellar hypometabolism fotlowing cerebral insult: a positron emission tomographic study. Ann Neurol 15:425-434, 1984 14 Phelps ME, Huang SC, Hoffman EJ, et ak Tomographic measurement of local cerebral glucose metabolic rate in humans with (F-18)2-fluoro-2-deoxy-~glucose: validation of method. Ann Neurol6:371-388, 1979 15 Phelps ME, Mazziotta JC, Kuhl DE, et al: Tomographic mapping of human cerebral metabolism: vIsual stimulation and deprivation. Neurology 31:517-529, 1981 16 Polyak S L The Vertebrate Visual System. Chicago, University remarks on neuronal mechanisms in Neurobiol 1:3-4, 1982 18. Reese F M Bilateral homonymous hemianopsia Am J Ophthalmol 38:44-57, 1954 19. Reivich M, Kuhl DE, Wolf A, et al: 18-F-Fluorodeoxyglucose method for the measurement of local cerebral glucose utilization in man. Circ Res 44:127-137, 1979 20. Riddoch G: Visual disorientation in homonymous half-fields. Brain 58:376-382, 1935 21. Silver FL, Bosley TM, Chawluk JB, et al: Resolving metabolic abnormalities in a case of pure alexia. Neurology 35(suppl 1):180, 1985 22. Sokoloff L, Reivich M, Kennedy C, et al: The 14-C-deoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J Neurochem 28:897916, 1977 23. Sprague JM: Interaction of cortex and superior colliculus in 17 Bosley et al: PET of Visual Recovery 449 mediation of visually guided behavior in the cat. Science 153:1544- 1547, 1966 24. Stensaas SS, Eddington DK, Dobelle WH: The topography and variability of the primary visual cortex in man. J Neurosurg 401747-755, 1974 25. Symonds C, Mackenzie I: Bilateral loss of vision from cerebral infarction. Brain 80:415-455, 1957 26. Ter-Pogossian MM, Mullan NA, Hood JT: Design considerations for a positron emission tomograph ( P I T I V) for imaging of the brain. J Comput Assist Tomogr 2:539-544, 1978 27. Teuber H-L: Recovery of function after brain injury in man. In Outcome of Severe Damage to the Central Nervous System (Ciba Foundation Symposium 34), Amsterdam, Elsevier, 1974, pp 159-186 Correction Please note that in the paper by Grotta and associates (21:171-175, 1987), the name of the fourth author has been misspelled. It should be Catherine Reid, BS. Our apologies for this oversight. 450 Annals of Neurology Vol 21 No 5 May 1987 28. Van Essen D: Functional organization of primate visual cortex. In Peters A, Jones 13G (eds): Cerebral Cortex, Vol 3. New York, Plenum, 1985, pp 259-330 29. Walsh FB, Hoyt WF: Clinical Neuro-Ophthalmology. Baltimore, Williams & Wilkins, 1969 30. Weinberger LM, Grant FC: Visual hallucinations and their neuro-optical correlates. Arch Ophthalmol 23: 116-199, 1940 31. Wise RJS, Rhodes ClG, Gibbs JM, et ak Disturbance of oxidarive metabolism of glucose in recent human cerebral infarcts. Ann Neurol 14:627--637, 1983 32. Zihl J: Recovery of visual functions in patients with cerebral blindness: effect of specific practice with saccadic localization. Exp Brain Res 44:159-169, 1981