Cerebral Hemispheric Low-Flow Infarcts in Arterial Occlusive Disease
Lesion Patterns and Angiomorphological Conditions
Background and Purpose Among the pathogenetic causes of subcortical hemispheric infarcts are small-vessel disease, thromboembolic occlusions of small arteries, and hemodynamic compromise in low-flow conditions. A topographic classification of these infarcts based on CT and MRI can be misleading.
Methods We evaluated 30 consecutive patients with presumed supratentorial low-flow infarcts. CT was available in all cases, with additional MRI in 14 patients. In all cases the occlusion pattern of the extracranial and intracranial arterial system was studied in detail with angiography.
Results The dominant lesion patterns seen on CT and MRI were multilocal chainlike lesions in 19 and confluent striated lesions in 8 cases located in the supraventricular and paraventricular deep white matter. In 8 patients subcortical lesions extended into the adjacent cortex. Angiography revealed that extracranial occlusive disease (n=24) or stenosis of the middle cerebral artery (n=6) was always accompanied by impairment of the circle of Willis, in either the anterior part (n=25) and/or the posterior part (n=16). Moreover, leptomeningeal pathways indicative of vascular hemispheric compromise were identified in 26 cases. In total, 29 of 30 patients displayed a noncompetent circle of Willis.
Conclusions Low-flow infarcts show typical but not pathognomonic lesion patterns on CT and MRI. Definite diagnosis requires knowledge of the complex vascular compromise of the extracranial and/or intracranial arterial system. A noncompetent circle of Willis should be regarded as the additional predisposing condition in hemispheric low-flow infarcts.
Classification of brain infarcts is commonly based on location, size, and shape of parenchymal damage as revealed on CT or MRI.1 2 3 4 5 6 Convincing evidence has been presented that low-flow infarcts (also called terminal supply area or subcortical watershed infarcts) are rare compared with infarcts of thromboembolic origin.5 7 8 Infarcts in so-called watershed areas between neighboring territories of the main hemispheric arteries are difficult to distinguish from embolic territorial infarctions, and it is doubtful whether they should be designated as a separate entity.7 Identification of the underlying vascular pathology is mandatory since therapeutic strategies differ considerably.3 9
Noninvasive methods such as extracranial and transcranial Doppler sonography or, as recently reported, MR angiography can help to reveal morphological and functional aspects of the underlying vascular pathology.10 11 12 It is therefore surprising that to our knowledge detailed angiomorphological analysis of the causative pathogenetic mechanisms in presumed low-flow infarcts, even in larger series, is not available.2 5 12 13 14 15 16
In this study we present clinical, CT, MR, and angiographic findings in 30 symptomatic patients with subcortical low-flow infarcts. We have attempted to answer the following questions: Which angiomorphological conditions are the basis for low-flow infarcts? What is the spectrum of CT and MRI findings in these infarcts?
Subjects and Methods
Patients with low-flow infarcts were enrolled in this study if the following criteria were fulfilled: (1) clinical symptoms correlated with a hemispheric lesion on CT, (2) a subcortical lesion on CT was suspicious of a low-flow infarct, (3) Doppler ultrasonography and angiography revealed occlusive cerebrovascular disease, and (4) angiography allowed definite assessment of the extracranial and intracranial arterial systems. Patients with evidence of severe microangiopathy or of territorial infarctions on CT or MR were excluded. Moreover, monosymptomatic patients with amaurosis fugax alone were not considered. On the basis of these criteria, 30 consecutive patients were selected for this study between January 1990 and December 1995.
All patients underwent CT and angiographic investigations. Initial CT was performed within 3 weeks of the onset of symptoms; in 14 cases CT was available within 1 week. Angiographic studies included bilateral carotid and vertebral digital subtraction angiograms (mostly indirect opacification through the subclavian arteries) and were performed within 4 weeks of the onset of symptoms. No procedure-related complication was observed during or after angiography.
Fourteen patients underwent routine spin-echo MRI of the brain with double-echo T2-weighted sequences (repetition time, 3000 ms; echo time, 15 and 90 ms; 1 excitation). Six patients had additional T1-weighted pulse sequences (before and after intravenous gadolinium-DTPA application; repetition time, 700 ms; echo time, 15 ms; 1 excitation). MRI examinations were performed on a 1.5-T unit (Magnetom V63, Siemens) within 4 weeks of the onset of symptoms.
Further examinations included extracranial continuous-wave Doppler ultrasonography in 30 patients and transcranial Doppler ultrasonography in an additional 23 patients. Cerebral vasomotor reactivity was measured during hypocapnia and hypercapnia by transcranial Doppler in 8 patients, while single-photon emission CT was performed in 14 patients. Results of cerebral vasomotor and single-photon emission CT studies have been reported elsewhere in part.10 13
Neuroradiological evaluation was performed by two experienced neuroradiologists. CT and MRI lesions in the supratentorial terminal supply area were classified based on location and extent, comparable with the recently proposed classification of Nakano et al.17 Lesions in the supraventricular and paraventricular white matter were distinguished; concomitant cortical involvement was documented. The angiographic occlusion pattern of the extracranial and intracranial arterial systems was analyzed, and special attention was directed to the arteries of the circle of Willis and the primary and secondary collateral pathways, such as ophthalmic and leptomeningeal arteries. Angiographic data were entered in a scheme case by case. For the evaluation of the internal carotid artery, occlusive disease was defined as high-grade stenosis or occlusion (stenosis >80% according to the measurement method used in the North American Symptomatic Carotid Endarterectomy Trial).18 Moderate or low-grade stenoses (<80%) were not considered because of lack of hemodynamic evidence. Patients were classified into three types based on the predominant angiographic occlusion pattern: type A, unilateral occlusive disease of the internal carotid artery; type B, bilateral occlusive disease of the internal carotid artery; and type C, unilateral occlusive disease of the middle cerebral artery.
Thirty patients (21 males, 9 females) with symptomatic low-flow infarcts were studied. The mean age was 54.5 years (range, 15 to 78 years). Transient ischemic attacks were noted in 15 (50%) of the 30 patients and minor strokes in 7 (23%). Five patients (17%) had a major stroke, and 3 (10%) had a progressive stroke. Fluctuating and/or repetitive symptoms were observed in 16 patients (12 patients with transient ischemic attacks and 4 with strokes). Predisposing hemodynamic factors such as heart failure or cardiac arrest were not observed. Stroke risk factors were hypertension in 13, smoking in 12, hypercholesterolemia in 8, and diabetes mellitus in 6 patients. Clinical impairment was greater than expected from the obvious lesion size in most cases. Seven of the 15 patients with stroke showed a good clinical recovery.
Clinical findings, lesion side on CT/MRI, and the predominant occlusion pattern on angiography correlated well in all cases.
CT and MRI
Both CT and MRI revealed lesions of different size, shape, extent, and location. Typical multilocular chainlike lesions were seen in 19 patients, while a confluent striated lesion pattern was observed in 8 patients (Fig 1⇓). Solitary lesions were detected in 3 patients. Lesion size (maximum diameter) was less than 2.0 cm in 22 patients, and very small lesions less than 1.0 cm could be detected in 11 of these patients. The maximum diameter of all 8 confluent lesions exceeded 2.0 cm. The location of low-flow infarcts is listed in Table 1⇓. Lesions were most frequently detected in the supraventricular white matter in 12 patients and in the paraventricular white matter in 4. Combined lesions, with involvement of the supraventricular and paraventricular white matter, were observed in 14 patients. In 8 patients the predominant subcortical lesion extended into the cortex and could be differentiated from focal atrophy. Five of these small concomitant cortical infarcts were located frontoparietally between the territories of the middle and anterior cerebral arteries. One patient additionally had a severe hemiatrophy. Four patients displayed bilateral small low-flow infarcts. In 5 of 6 patients with isolated stenosis or occlusion of the middle cerebral artery, lesions were detected in the supraventricular white matter. Two of these had additional involvement of the paraventricular white matter. In 9 of 14 patients, information additional to that provided by CT was obtained from MRI. In 4 patients additional white matter lesions were observed; in 1 case a small concomitant cortical infarct was detected. Contrast enhancement in a cortical (frontal or parietal) area and a subcortical area was detected in 2 patients (Fig 2⇓); purely cortical or subcortical enhancement in another 2 patients was detected within 3 weeks of the onset of symptoms.
High-grade stenosis or occlusion of the internal carotid artery was demonstrated angiographically in 24 patients. Thirteen of these had unilateral occlusive disease (extracranial in 4, intracranial in 3, combined extracranial and intracranial in 6 patients) classified as type A. Eleven patients had bilateral occlusive disease (extracranial in 1, intracranial in 2, combined extracranial and intracranial in 8) classified as type B. In 6 patients unilateral middle cerebral artery occlusive disease was the predominant angiographic occlusion pattern, corresponding to type C. A schematic synopsis of the angiographic extracranial and intracranial occlusion patterns for all patients is given in Fig 3⇓. In addition, analysis of the intracranial occlusion pattern and of the collateral pathways is shown in Table 2⇓. Type A and B patients had evidence of severe extracranial and associated intracranial occlusive vascular disease, while in type C the circle of Willis was not intact angiographically in 5 of 6 patients with middle cerebral artery occlusive disease. In predominant internal carotid artery disease, 20 of 24 patients had a noncompetent anterior part of the circle of Willis (A1 portion of the anterior cerebral artery in 18, anterior communicating artery in 2 patients), and absence or severe hypoplasia of the posterior communicating artery and/or the P1 portion of the posterior cerebral artery was evident in 14. Twenty of these patients revealed collateral leptomeningeal pathways. In patients with unilateral middle cerebral artery disease (n=6), absence or severe hypoplasia of the anterior communicating artery was noted in 4, and the anterior part of the circle of Willis was affected with a narrow A1 portion of the anterior cerebral artery in 1 patient contralateral to the middle cerebral artery stenosis. All patients in this group C had evidence of collateral leptomeningeal pathways.
In summary, evaluation of the angiographic data showed that 29 patients displayed a noncompetent circle of Willis.
Ischemic lesions exclusively located in the subcortical white matter represent a diagnostic challenge. Size, extent, and location of the ischemic lesion are not specific with regard to the pathogenetic mechanism.4 7 13 17 Low-flow infarcts are defined as a result of reduced perfusion pressure in terminal supply areas possibly reinforced by episodes of hypotension. Neuropathological studies have described supratentorial watershed areas between the deep and superficial territories of the middle cerebral artery in the deep white matter and between the anterior and middle cerebral arteries in the supraventricular white matter.19 20 Infarct classification based mainly on the morphological pattern has proved to be helpful in typical larger infarctions but to have limitations with regard to smaller subcortical lesions. Moreover, the postulated terminal supply and watershed areas are still under controversial discussion.7
All of the subcortical lesions in our selected patient group were located in the supraventricular and paraventricular white matter. They were characterized by circumscribed chainlike or confluent hypodense areas in the white matter representing the core of an hemodynamically induced infarction. Nakano et al17 recently reported that lesions of the centrum semiovale in large subcortical infarcts are not associated with middle cerebral artery disease. This is not the case for smaller subcortical low-flow infarcts investigated in the present study. Here middle cerebral artery occlusive disease can be an important factor and induce lesions in the supraventricular white matter, as demonstrated in 5 of our 6 patients with this occlusion pattern. In most cases the terminal supply area was located in the supraventricular white matter and could extend into the paraventricular white matter. The concept of an ischemic penumbra can help to explain the discrepancy between clinical impairment and obvious lesion detectable by CT and MRI in this type of cerebral ischemia.21
In some patients the predominantly subcortically located lesions reached the adjacent cortex. Cortical involvement was evident in 2 of 4 patients without leptomeningeal anastomoses. Lesions were grossly identical in CT and MRI. In 9 of 14 patients, additional information was available on MRI that in some cases was performed later in the course of the disease. The most interesting additional finding on MRI was a cortical and/or subcortical contrast enhancement in 4 patients within 3 weeks after onset of symptoms. In 2 patients this was exclusively cortical and in the remaining patients cortical and subcortical. This finding has not been reported for low-flow infarcts. In 5 patients small concomitant cortical lesions were located frontoparietally, indicating that small infarctions existed in the postulated watershed area. This finding supports the hypothesis that low-flow infarctions can extend to the postulated watershed area between anterior and middle cerebral arteries. The pathogenetic mechanism is not yet clear, but a possible explanation may be adjacent thrombosis induced by the low-flow condition or insufficient leptomeningeal anastomoses.
Almost all authors mention angiography in the diagnostic workup of low-flow infarcts. Extracranial occlusive disease below the circle of Willis can be asymptomatic and compensated by sufficient collateral pathways. Our findings help to explain why hemodynamically induced infarcts of the brain are comparatively rare and are seldom encountered in patients without additional lesion or abnormality of the circle of Willis. Recently, the role of the posterior communicating artery has been described in a cohort of 29 patients studied by means of MRI angiography, but the morphology of the entire circle of Willis could not be completely analyzed because of methodological reasons.12 Our findings support the assumption that the anterior part of the circle of Willis is of equal importance and that various patterns of restricted collateral flow may contribute to low-flow lesions. Although noninvasive methods such as extracranial and transcranial Doppler ultrasonography are useful in the screening of cerebrovascular disease, only selective cerebral angiography can thus far reliably demonstrate the complexity of the compromised primary and secondary collateral pathways.
Intracranial angiographic predictors of low-flow infarcts were different in patients with predominant internal carotid and middle cerebral artery occlusive diseases. In internal carotid artery occlusive disease, occlusion, stenosis, and/or hypoplasia of the anterior part (in 20 of 24 patients) and the posterior part of the circle of Willis (in 14 of 24 patients) were the dominant intracranial findings. In middle cerebral artery occlusive disease, collateral flow via the anterior communicating artery could not be demonstrated in 4 of 6 patients.
In conclusion, low-flow infarcts show typical but not pathognomonic lesion patterns on CT and MRI with involvement of the supraventricular and paraventricular white matter, which can be defined as terminal supply areas. These findings support the concept of the most vulnerable distal field. There is evidence in our larger series that deep subcortical lesions can extend into the adjacent frontoparietal cortex between the territories of the middle and anterior cerebral arteries. On the other hand, we did not observe purely superficial cortical lesions. Definite diagnosis of low-flow infarcts requires information about the underlying vascular pathology. In most cases unilateral or bilateral extracranial and intracranial occlusive disease only results in a severe hemispheric hemodynamic compromise if the circle of Willis is not intact.
This study was supported in part by IZKD “Prävention und Kompensation von Störungen des ZNS” (Neuro 03) of the Medical Faculty, Aachen Technical University.
Reprint requests to Michael Mull, MD, University Hospital of the Technical University, Department of Neuroradiology, Pauwelsstr 30, D-52057 Aachen, Germany.
- Received July 19, 1996.
- Revision received October 10, 1996.
- Accepted October 14, 1996.
- Copyright © 1997 by American Heart Association
Zülch KJ. The Cerebral Infarct. New York, NY: Springer-Verlag; 1985.
Ringelstein EB, Zeumer H, Angelou D. The pathogenesis of strokes from internal carotid artery occlusion: diagnostic and therapeutical implications. Stroke. 1983;14:867-875.
Bladin PF, Berkovics SF. Striatocapsular infarction: large infarcts in the striatiolenticular arterial territory. Neurology. 1984;34:1423-1430.
Torvik A. The pathogenesis of watershed infarctions in the brain. Stroke. 1984;15:221-223.
Howard R, Trend P, Ross Russell RW. Clinical features of ischemia in cerebral arterial border zones after periods of reduced cerebral blood flow. Arch Neurol. 1988;44:943-940.
Bogousslavsky J. The plurality of subcortical infarction. Stroke. 1992;23:629-631.
Ross MR, Pelc NJ, Enzmann DR. Qualitative phase contrast MRA in the normal and abnormal circle of Willis. AJNR Am J Neuroradiol. 1993;14:19-25.
Weiller C, Ringelstein EB, Reiche W, Buell U. Clinical and hemodynamic aspects of low-flow infarcts. Stroke. 1991;22:1117-1123.
Harrison MJC, Marshall J. The variable clinical and CT findings after carotid occlusion: the role of collateral blood supply. J Neurol Neurosurg Psychiatry. 1988;51:269-272.
Bogousslavsky J, Regli F. Unilateral watershed cerebral infarcts. Neurology. 1986;36:373-377.
Nakano S, Yokogami K, Ohta H, Goya T, Wakisaka S. CT-defined large subcortical infarcts: correlation of location with site of cerebrovascular occlusive disease. AJNR Am J Neuroradiol. 1995;16:1581-158
Adams JH, Brierley JB, Connor RCR, Treip CS. The effects of systemic hypotension upon the human brain: clinical and neuropathological observations in 11 cases. Brain. 1966;89:235-268.
Romanul FCA, Abramowitcs A. Changes in brain and pial vessels in arterial border zones. Arch Neurol. 1964;11:40-65.