Clinical and Pathological Correlations
Background and Purpose Single, oval lesions greater than 5 mm in diameter lying inferior to the lateral putamen (infraputaminal lacunes [IPLs]) seen on CT or MR images are commonly reported as lacunar infarcts. To determine the clinical relevance and underlying pathology of IPLs, we evaluated the imaging appearances, clinical features, vascular risk factors, and histopathology in patients with IPLs.
Methods Consecutive MR scans were reviewed for the presence of IPLs. Serial patients seen in routine clinical practice with IPLs were also included. Vascular risk factors were obtained from a prescan questionnaire. Histology and microangiography were performed on postmortem material. A MEDLINE search for putaminal infarcts was performed to look for imaging lesions typical of IPLs.
Results Three of 100 serial MR scans had IPLs (3%). Nine other patients with in vivo (7) or postmortem (2) MR scans had IPLs. No neurological symptoms could be related to the IPLs. There were no differences in age, hypertension, diabetes, or presence of cortical enlarged perivascular spaces (EPVSs) between patients with and without IPLs. Unlike infarcts, IPLs were isointense with the cerebrospinal fluid on proton density MR sequences. Histological correlation of three MR scans showed IPLs to be a single large EPVS, situated lateral to the anterior commissure. IPLs were located at a point where multiple lenticulostriates turn sharply dorsally. An IPL was the probable cause of the apparent infarct in six publications from peer-reviewed literature that linked different clinical signs to putaminal infarct.
Conclusions IPLs are EPVSs that can be differentiated from infarcts on proton density MR images.
In his detailed classification of lacunes and other different cavitary states, Marie1 illustrated the gross appearance of a particular type of “lacune,” which he described as a solitary cavity, varying from the size of a lentil to that of a small bean, situated in the inferior part of the lenticular nucleus. Marie thought that these lesions were due to a dilatation of the perivascular space surrounding a lenticulostriate artery at its entry into the lentiform nucleus but provided no details of the histology. More recently, Poirier2 reported the microscopic findings in an asymptomatic, 12-mm-diameter “lacuna” situated in the lateral subputaminal region. The lesion was lined by epithelium and traversed by normal arteries, in keeping with an enlarged perivascular space (EPVS). Poirier’s case tended to confirm the existence of these infraputaminal lacunes (IPLs) and also Marie’s hypothesis that they are EPVSs. With MRI, Jungreis et al3 have noted hyperintensities greater than 5 mm in diameter in the infraputaminal region on T2-weighted images compatible with IPLs, but they provided no pathological correlations to support their conclusion that these hyperintensities are caused by EPVSs.
More recently, however, Braffman et al4 reported that a single, oval, clearly defined lesion lying in the region of the lateral putamen, in a location identical to Poirier’s IPL, was histologically an infarct. In addition, we found a case report of a CT lesion with the characteristic size, shape, and location of an IPL, which was identified as an infarct.5 We have also noted that IPLs are frequently reported as lacunar infarcts on routine MR and CT studies.
The present study was undertaken (1) to determine the pathomorphology of IPLs and to decide whether IPLs are infarcts or EPVSs, (2) to determine the radiological characteristics of IPLs and to establish the frequency of IPLs on serial MR scans, and (3) to determine how IPLs are being reported in routine imaging reports and in the literature.
Subjects and Methods
Two independent observers evaluated serial MR scans of consecutive patients undergoing brain MRI in an outpatient setting. We noted the presence of (1) IPLs and their MR characteristics, (2) multiple subcortical EPVSs located near the vertex (small focal hyperintensities that were seen only on heavily weighted T2 sequences), and (3) confluent white matter hyperintensities compatible with subcortical arteriosclerotic encephalopathy. The ventricular index6 was measured.
The presence of prior stroke, hypertension, and diabetes was determined by a questionnaire given to the patients at the time of the MR scan or by chart review. Vascular risk factors and MR findings were compared between patients with and without IPLs. In patients with IPLs, neurological symptoms and signs were further determined by chart review or telephone contact with the primary physician.
In brains with IPLs examined at autopsy, we noted the anatomic location of IPLs. We examined the infraputaminal region microscopically with sections stained with hematoxylin and eosin, Luxol fast blue/periodic acid–Schiff for myelin, Gomori’s trichrome stain for connective tissue, and Verhoeff’s elastic stain. Immunohistochemistry was performed on paraffin sections as necessary, with antibody to smooth muscle actin (Biogenex) to identify the smooth muscle component of arterial walls as well as pericyte proliferation in the adventitia. The relation of the lenticulostriate arteries to the typical location of IPLs was studied with postmortem angiography.
We reviewed the reports of all scans with IPLs on in vivo scans. To determine whether any clinical features were being attributed to IPLs in the literature, we performed a MEDLINE search for putaminal infarcts and reviewed published CT and MR illustrations for lesions resembling IPLs. We also reviewed the inclusion criteria in recent asymptomatic stroke studies.
A total of 100 serial MR scans were reviewed, and three patients with IPLs were found (3%). Nine other scans with IPLs were available: 7 with in vivo MR scans and 2 with postmortem MR scans. IPLs were clearly demarcated oval lesions that were isointense with the cerebrospinal fluid on all imaging sequences, rendering them hypointense on T1-weighted sequences (Fig 1⇓), isointense on proton density, and hyperintense on heavily weighted T2 sequences with respect to brain (Fig 2⇓, middle panel). The longest diameter of the IPLs was 10 mm. IPLs occasionally had a lobulated appearance or were made up of two or three smaller focal hyperintensities (Fig 2⇓, middle panel).
There were no differences in cortical EPVS or in confluent white matter disease between scans with and without IPLs, but the ventricular index was smaller in patients with IPLs (39.1) than in patients without IPLs (46.4). Superimposition of coronal and axial postmortem micoangiographic images and MR images showed that IPLs were located at a point where multiple lateral lenticulostriate arteries course together and turn sharply dorsally. IPLs were either reported as lacunar infarcts (6/10) or else not mentioned in the scan report (4/10).
In the 12 patients with IPLs, no neurological symptoms could be related to the IPLs. In one patient a pure motor hemiparesis with a transient language disturbance was attributed to what was thought to be a small deep infarct of the putamen seen on CT (Fig 2⇑, top panel). However, MR showed the lesion seen on CT to be an IPL and revealed an infarct in the ipsilateral corona radiata that had not been seen on CT that was the likely cause of the symptoms (Fig 2⇑, middle and bottom panels). There were no differences in hypertension or diabetes between patients with and without IPLs. The age tended to be lower in patients with IPL (64.8 years) than in patients without IPL (71.6 years). The youngest patient with an IPL was 36 years old.
Histology was available in five brains with IPLs, and correlation with MR was available in three of these. In one of these the IPL was seen on an in vivo scan (Fig 1⇑), and in two brains IPLs were found on postmortem MRI. No MRI or CT was performed in an additional two brains with IPLs found at autopsy. IPLs were either a single large EPVS or a group of multiple smaller EPVSs. On coronal section, the EPVSs were always situated lateral to the anterior commissure. Gross features that were diagnostic of EPVS were the following: (1) the lesions contained patent, histologically normal blood vessels; (2) the walls of the cavity were regular and clearly demarcated; (3) no occluded arteries were seen in the vicinity of the IPL; and (4) the surrounding tissue was not infarcted.
On histological sections, arteries in longitudinal and transverse profiles were present within IPLs. Some vascular profiles occupying the IPLs were tortuous, suggesting elongation. Often, serial histological sections were required to reveal the vascular elements coursing through the IPL. Elastic stains and immunostains for smooth muscle actin demonstrated the vessel population within an IPL to be composed of either arteries alone or both arterial and venous structures. In serial sections of the dilated spaces that did not contain vascular profiles, thickened collagen bundles tangential to the adventitia or detached leptomeninges were present. The adventitia was often thicker than seen in vessels at the same location in brains without enlarged perivascular spaces. Multiple corpora amylacea were present in the brain parenchyma surrounding the IPL in two of the brains. These tended to outline the immediate subpial parenchyma. The white matter surrounding the IPL in those two brains also showed vacuolation and decreased myelin staining with minimal gliosis. We did not observe the typical histological changes of infarction in any of our IPLs.
Through the literature review we found six publications in peer-reviewed journals5 7 8 9 10 11 that included images of MR or CT lesions identical in size, shape, and location to IPLs. The lesions were all described as infarcts, but the clinical syndromes that were reported in association with these were all different: painful tonic spasms,7 parkinsonism,5 hemiballismus,8 cervical dystonia9 (case 1), arm weakness10 and asymptomatic in association with subarachnoid hemorrhage11 (case 1). Moreover, we could not find any mention that IPLs had been looked for or excluded in studies of asymptomatic infarction that we reviewed. For example, a recent report from the Asymptomatic Carotid Atherosclerosis Study fails to provide a clear description of what should be considered an asymptomatic brain infarction on CT or MRI.12
We have found IPLs to be asymptomatic EPVSs and a relatively frequent finding on serial MR scans. We found no evidence of an association of IPLs with brain ischemia, stroke symptoms, or movement disorders. Like other EPVSs, IPLs appear not to be directly due to focal cerebral ischemia,13 and EPVSs are only rarely symptomatic.14 15 Our report is of clinical relevance because misinterpretation of radiological findings is common5 7 8 9 10 11 and could lead to inappropriate management.
The histology underlying IPLs was diagnostic of EPVS in all five cases we examined. The lesion that Braffman et al4 reported to be an infarct on histology was isodense with cerebrospinal fluid on proton density MR images, like an EPVS. We interpret the photographs in that report also to be compatible with an EPVS. The histology described in that report also seems compatible with an EPVS if, as in many of our serial sections, the section of dilated space examined did not happen to contain vascular profiles. Occlusion of single or multiple lenticulostriate arteries near their origins produces an area of necrosis extending distally along the entire course of the occluded artery16 since penetrating arteries are end-arteries and there are no collaterals to restore flow distal to the occlusion. The lateral lenticulostriate arteries are long and mostly extend up to the lateral ventricle, and occlusion of one or more of these at the inferior putamen would be very unlikely to produce a lacunar infarct the size and shape of an IPL.
The lack of relatable neurological symptoms and signs in IPLs is against these lesions causing local structural damage. Infarction of the putamen may be symptomatic,17 and it is likely that at least some of the IPLs we reported would have produced symptoms if the putamen had been damaged. We have shown, however, that IPLs are infraputaminal and only appear to be in the putamen on imaging because of partial volume effect. EPVSs are usually asymptomatic,14 but Poirier et al15 hypothesized that occasionally they may spontaneously enlarge, producing neurological deficits by local pressure effects. There was no evidence of local mass effect in any of our cases, either on imaging or at autopsy.
Both our series and the cases we reviewed show that IPLs are being diagnosed as infarcts, and several different neurological deficits have been ascribed to them,5 7 8 9 10 11 probably erroneously. The six case reports we found in the literature included reports of very different movement disorders. This probably reflects a tendency among neurologists to link clinical movement disorders to imaging abnormalities in the basal ganglia. If IPLs were symptomatic they would be unlikely to produce five distinct clinical syndromes.
The fact that IPLs are not generally recognized to be distinct from infarcts has implications for studies of asymptomatic stroke. IPLs are probably more likely to be confused with infarcts on CT than on MRI since they can be differentiated from lacunar infarcts on CT only by their location and shape. The reported frequency of asymptomatic infarction of 15%12 would likely be reduced if IPLs were excluded. Potential error may be greatest with interpretation of subcortical lesions in or near the putamen.
Although EPVS less than 5 mm in diameter may be found in the typical infraputaminal location of IPLs (Fig 1⇑, top right panel, right side of brain), these were not counted as IPLs in our study to distinguish IPLs from small EPVSs that are frequently seen in the region of the anterior commissure. Perivascular spaces smaller than 2 mm in diameter in a location similar to that of IPLs are a frequent finding on MRI and like IPLs are not associated with risk factors. IPLs differ from other EPVSs greater than 2 mm in diameter, which are associated with age and hypertension.14 Although the number of patients compared was too small to draw definitive conclusions, the lack of association of IPLs with major vascular risk factors and stroke symptoms in the present study and the tendency to a lower age suggest that IPLs may be part of a continuum of peri- vascular space size related to the development and location of the lenticulostriate arteries.
In our study postmortem angiography demonstrated that the tortuous course taken by the lenticulostriates may be most pronounced at the inferior putamen, where they change direction from a lateral to a dorsomedial path and are grouped closely together. At that location, a proximal perivascular space might “normally” be large because it contains several vessels. This would be analogous to a small cisternal enlargement of the subarachnoid space contiguous with the perivascular space along the course of these arteries. In some cases, perivascular trauma due to pulsations of multiple small arteries may enlarge the space and be a factor in the development of IPLs.18 The presence of histological evidence of perivascular shrinkage in the form of reactive astrocytosis, myelin pallor, and corpora amylacea in two of our cases suggests that there may also be an ex vacuo component in the pathogenesis of IPLs.
In conclusion, we found that IPLs are EPVSs that are relatively frequent and recognizable by their imaging characteristics. IPLs should not be mistaken for lacunar infarcts, and IPLs should be specifically excluded in studies of silent cerebral infarction. Neurological deficits, in particular movement disorders, should not be ascribed to IPLs.
- Received March 8, 1995.
- Revision received May 24, 1995.
- Accepted May 24, 1995.
- Copyright © 1995 by American Heart Association
Mayo J, Arias M, Leno C, Berciano J. Vascular parkinsonism and periarteritis nodosa. Neurology. 1986;36:874-875.
Merchut MP, Brumlik J. Painful tonic spasms caused by putaminal infarction. Stroke. 1986;17:1319-1321.
Mas J-L, Launay M, Derouesné C. Hemiballism and CT-documented lacunar infarct in the lenticular nucleus. J Neurol Neurosurg Psychiatry. 1987;50:104-118.
Bragoni M, Di Piero V, Priori R, Valesini G, Lenzi GL. Sjogren’s syndrome presenting as an acute ischemic stroke. Stroke. 1994;25:2276-2279.
Tatter SB, Buonanno FS, Ogilvy CS. Acute lacunar stroke in asso-ciation with angiogram-negative subarachnoid hemorrhage. Stroke. 1995;6:891-895.
Brott T, Tomsick T, Feinberg W, Johnson C, Biller J, Broderick J, Kelly M, Frey J, Schwartz S, Blum C, Nelson JJ, Chambless L, Toole J. Baseline silent cerebral infarction in the Asymptomatic Carotid Atherosclerosis Study. Stroke. 1994;25:1122-1129.
Van Swieten JC, Van den Hout JHW, Van Ketel BA, Hijdra A, Wokke JHJ, Van Gijn J. Periventricular lesions in the white matter on magnetic resonance imaging in the elderly: a morphometric correlation with arteriolosclerosis and dilated perivascular spaces. Brain. 1991;114:761-774.
Heier LA, Bauer CJ, Schwartz L, Zimmerman RD, Morgello S, Deck MDF. Large Virchow-Robin spaces: MR-clinical correlation. AJNR Am J Neuroradiol. 1989;10:929-936.
Hughes W. Origin of lacunes. Lancet. 1965;1:19-21.