Search

A 37-year-old female presented with gradually progressive asymmetric ascending paraesthesia and weakness involving bilateral upper and lower limbs. The MRI spine images revealed expansile intramedullary, solid cystic, peripherally enhancing lesion with a haemosiderin cap along the lower margin. The lesion extended into the left C5 and C6 nerve root exit zones, along with thickening and enhancement of the nerve roots. She underwent excision of the lesion, which revealed intramedullary schwannoma on histopathological examination. Presence of the cap, an extension of the lesion into the nerve root exit zone, with associated thickening and enhancement of the dorsal nerve roots should alert the radiologist to consider the possibility of intramedullary schwannoma rather than ependymoma. Schwannoma showing compact Antoni A area with Schwannian whorls and nuclear palisades (A,B) and loose Antoni B area with haemosiderin pigment (C). The tumour exhibits diffuse positivity for S-100 protein (D) and negativity for GFAP (E). [Stain: A-C: Haematoxylin and Eosin; D,E: Immunoperoxidase. Magnification = Scale Bar, A-E: 100μm].

On T2-weighted images, most solid lesions exhibit nonspecific intermediate signal intensity, whereas most cystic lesions exhibit marked hyperintensity. In contrast, on T2-weighted images, a relatively small number of lesions exhibit hypointensity. This review aimed to differentiate, according to the histopathologic findings, head and neck lesions showing hypointensity on T2-weighted images.In this review article, hypointense head and neck lesions on T2-weighted images are classified into the following nine categories: calcified or osseous lesions, granulomatous lesions, fibrous lesions, mucous- or proteincontaining lesions, hemosiderin-containing lesions, melanin-containing lesions, thyroglobulin-containing lesions, rapid blood flow, and air-filled spaces.Knowledge regarding hypointense head and neck lesions on T2-weighted images allows radiologists to make accurate differential diagnoses. Key points • Hypointense head and neck lesions on T2-weighted images include calcified or osseous lesions, granulomatous lesions, fibrous lesions, mucous- or protein-containing lesions, hemosiderin-containing lesions, melanin-containing lesions, thyroglobulin-containing lesions, rapid blood flow, and air-filled spaces. • Radiologists should recognize the hypointense head and neck lesions on T2-weighted images for the final correct diagnosis, resulting in appropriate patient management.

Introduction: It is still unclear how often subarachnoid hemorrhage (SAH) leads to chronic hemosiderin depositions. In this study, we aimed to determine the frequency of chronic hemosiderin depositions after aneurysmal SAH in patients who did not undergo surgery. Furthermore, we analyzed typical MRI patterns of chronic SAH and sought to obtain information on the temporal course of MRI signal changes. Methods: We retrospectively analyzed 90 patients who had undergone endovascular treatment for acute aneurysmal SAH. In all patients, initial CT studies and at least one T2*-weighted MRI obtained 6 months or later after SAH were analyzed for the presence and anatomical distribution of SAH or chronic hemosiderin depositions. In total, 185 T2*-weighted MRI studies obtained between 2 days and 148 months after SAH were evaluated (mean follow-up 30.2 months). Results: On MRI studies obtained later than 6 months after SAH, subpial hemosiderin depositions were found in 50 patients (55.5 %). Most frequent localizations were the parenchyma adjacent to the frontal and parietal sulci and the insular cisterns. While the appearance of hemosiderin depositions was dynamic within the first 3 months, no changes were found during subsequent follow-up. MR signal changes were not only conclusive with subarachnoid hemosiderin depositions but in many cases also resembled those that have been associated with cortical hemosiderosis. Conclusions: T2*-weighted MRI is an effective means of diagnosing prior SAH. Our study suggests that chronic hemosiderin depositions can be found in a considerable number of patients after a single event of subarachnoid hemorrhage.

Gradient-echo (GE) MRI has been demonstrated to be the most sensitive current technique for detection of intracerebral haemosiderin, especially in the chronic stage of haemorrhage. Our purpose was to see whether GE MRI shows old haemorrhage indefinitely. We reviewed serial GE images of 105 adults with imaging features consistent with post-traumatic intracerebral haemorrhage, who had serial MRI at 1, 4-6, 12, and 24 months after trauma. Of 1235 scattered low-signal foci consistent with isolated intracerebral haemosiderin deposits on images at 4-6 months, 248 (20.1%) were not seen at 24-month assessment. Reviewing individual patients, we saw that in 71.8% of those with scattered haemosiderin deposits and 46.4% of those with haemosiderin surrounded by gliosis, the low-signal foci appeared less conspicuous with time. Even given certain limitations to the interpretation of these findings, it would appear that, even with the use of GE MRI, time affects the visibility of haemorrhagic intracerebral lesions. We therefore conclude that a time of 4-6 months to 1 year or slightly more should be recommended for most precise detection of haemosiderin deposits on MRI of head-injured patients, should this be thought desirable. Normal GE images may not exclude old haemorrhage.

MRI of intraventricular haemorrhage (IVH) has not been studied formally. We aimed to describe the degradation rate and patterns shown on 1.5 T MRI in IVH, comparing them to other coexisting brain hemorrhage. We studied 50 consecutive cases using T1-, proton-density, and T2-weighted images. IVH was seen in two forms: layered (free-flowing in ventricles) (37 cases) and/or clotted (31). Both were best shown by proton-density image. Layered IVH was seen in the dependent portions of the lateral ventricles with fluid (“blood-CSF”) levels, degrading more slowly than both clotted IVH and intraparenchymal hemorrhages (IPH) (acute blood products persisting for several more days; P < 0.05). Clotted IVH degraded at a rate comparable to IPH. IVH cleared rapidly and did not form hemosiderin. Subarachnoid hemorrhage (SAH) cleared faster and was less conspicuous than IVH. Hypertensive (22), aneurysmal (11), traumatic (2), idiopathic (9), or vascular malformation-related (6) IVH were seen. IVH coexisted with IPH (30) or SAH (12), or both (12). The high rate of layering with blood-CSF levels in IVH is most likely due to different densities of blood components and CSF and the fibrinolytic capability of the latter. Delayed degradation of layered IVH probably reflects high intra-ventricular oxygen and glucose content. Further study is necessary to determine if MRI characteristics of IVH are helpful in excluding other intraventricular diseases such as neoplasia and pyocephalus.

Our aim was to evaluate the diagnostic reliability of turbo spin-echo (TSE) sequences compared to a conventional dual-echo spin-echo (SE) sequence in routine brain MRI at 1.0 T. The following demands were made on TSE sequences: acquisition time-reduction of at least 50% and true proton density (PD) contrast (low-signal cerebrospinal fluid). A conventional spin-echo and two single-echo TSE sequences were used in 150 patients, a dual-echo TSE sequence in addition in 50 patients. Demonstration of most anatomical structures and disorders was equivalent with TSE and SE sequences. Advantages of TSE were reduced flow artefacts on T2-weighted images, better lesion contrast on PD-weighted TSE images (especially in the dual-echo sequence) and acquisition time reduction to about 5 min (single-echo TSE) and 3:35 min (dual-echo TSE). Disadvantages of TSE were: reduced contrast of iron-containing substances such as haemosiderin and of areas of calcification. By virtue of the shorter acquisition time and diagnostic reliability dual-echo TSE proved the best sequence. If it is used with only one acquisition--whereby image quality but not diagnostic reliability is slightly decreased--acquisition time can be further reduced to 1:48 min. Application of a susceptibility-sensitive gradient-echo sequence, such as FLASH, compensates for the disadvantages mentioned above.

We present a case of cerebral aspergillosis in an immunocompetent patient. The MRI signal characteristics were compared with the histologic findings. Irregular low-signal zones were demonstrated between the wall of the abscess and the central necrosis on T2-weighted images; the pathology specimen revealed concentrated iron in these transitional zones but no hemosiderin. Iron is an essential element for the growth of fungal hyphae. The low-signal zones may represent the areas where there was active proliferation of aspergillus, and the unique location of the low signal may be a helpful imaging characteristic for the diagnosis of an aspergillus abscess.

Giant cavernous malformations (GCM) are very large, low-flow vascular malformations, which usually have atypical imaging features and are commonly misdiagnosed preoperatively as neoplasms or vascular malformations. These lesions have mostly been reported in children. As cavernomas show different features in children compared to adults, we evaluated the imaging features of pediatric GCMs in order to help in the preoperative diagnosis of these malformations. Brain MR studies of nine children (mean age of 4 years; 8 months–9 years) with biopsy-proven GCM were retrospectively evaluated. We defined GCMs as cavernomas of ≥4 cm. Lesions were evaluated regarding their size, location, signal characteristics, general appearance (uni/multilocular) as well as regarding the presence of mass effect, edema, and fluid–fluid levels and were classified according to the Mottolese classification of pediatric cavernomas. Lesion locations were parietal (n = 5), frontal (n = 2), temporal, and intraventricular. Seven lesions were in the periventricular region (with five in the periatrial region). Six patients had T1 hyperintense multilobulated lesions with “bubbles of blood” appearance and three patients had heterogeneous lesions with reticular core. All lesions had mass effect, edema (marked in four cases), and peripheral hemosiderin rim. Fluid–fluid levels were also common (n = 7). Most of our lesions (six of nine) were classified as type IIIA, two as type IIIC, and one as type IA. In children, a GCM should be considered in case of very large hemorrhagic intra-axial mass with “bubbles of blood” multicystic appearance, surrounding hemosiderin ring, fluid–fluid levels, and accompanying edema–mass effect, especially in the periatrial location.

Intracerebral haemorrhage may be visible indefinitely on MRI, due to persistence of haemosiderin in macrophages around the lesion, but it is not clear whether all haemorrhages produce haemosiderin or, if not, what proportion cannot be identified as former haemorrhages on routine MRI. We performed routine MRI (spin-echo T2- and proton-density weighted images) in 116 survivors of moderate to severe head injury, 1–5 years after injury. We reviewed the images blindly and correlated them with CT in the acute stage, to determine how many haemorrhages from the acute stage were identifiable by virtue of haemosiderin deposition on late MRI. Of 106 haemorrhages in 78 patients on CT at the time of injury, 96 (90 %) were visible as haemosiderin on late MRI. Of the old haemorrhages without haemosiderin, seven of ten were in patients where another haemorrhage with haemosiderin was still visible elsewhere in the brain. No patient or haemorrhage features explained the formation or absence of haemosiderin. Thus about 10 % of definite haematomas show no trace of haemosiderin on routine spin-echo MRI. Radiologists should be alerted to supplement routine spin-echo with gradient-echo sequences if there is a reason to suspect, or specifically exclude, prior haemorrhage.

To examine the chronological changes characteristic of cortical laminar necrosis caused by brain infarction, 16 patients were repeatedly examined using T1-, T2-weighted spin-echo, T2*-weighted gradient echo, fluid attenuated inversion recovery (FLAIR) images, and contrast enhanced T1-weighted images at 1.0 or 1.5 T. High intensity cortical lesions were visible on the T1-weighted images from 2 weeks after ictus and became prominent at 1 to 3 months, then became less apparent, but occasionally remained at high intensity for 2 years. High intensity cortical lesions on FLAIR images became prominent from 1 month, and then became less prominent from 1 year, but occasionally remained at high intensity for 2 years. Subcortical lesions did not display high intensity on T1-weighted images at any stage. On FLAIR images, subcortical lesions initially showed slightly high intensity and then low intensity from 6 months due to encephalomalacia. Cortical lesions showed prominent contrast enhancement from 2 weeks to 3 months, but subcortical lesions were prominent from 2 weeks only up to 1 month. T2*-weighted images disclosed haemosiderin in 3 of 7 patients, but there was no correlation with cortical short T1 lesions. Cortical laminar necrosis showed characteristic chronological signal changes on T1-weighted images and FLAIR images. Cortical short T1 lesions were found not to be caused by haemorrhagic infarction.

The case of a 70-year-old woman with cerebral amyloid angiopathy (CAA) is presented. MRI of the head showed widespread miliary foci of haemorrhage within the cerebrum and cerebellum, with some additional linear lesions within the cerebral cortex and patchy lesions in the white matter. This is in contrast to the more usual pattern of intracranial haemorrhage in CAA, i. e., a lobar haematoma.

An unusual case of hemorrhage within a cyst in the choroid plexus of the lateral ventricle is presented. Choroid plexus cysts are commonly found incidentally on imaging studies, particularly MRI, where their characteristic appearance produces little diagnostic difficulty. In this case, high signal intensity of the cyst contents on T1-weighted images and apparent hemosiderin deposition in the cyst wall, not previously described, represented a dramatic change from a study 10 weeks previously and allowed a presumptive diagnosis of hemorrhage into the cyst, a presumption further supported by the changes in signal 20 weeks later.

Fifty-one patients with 59 angiographically proven cerebral arteriovenous malformations (AVMs) were examinded by high-field MRI to detect blood breakdown products. Results were correlated with the history of intracranial bleeding. Evidence of previous episodes of haemorrhage was seen in 10 of 12 patients (83.3%) with verified bleeding, in 4 of 9 patients (44.4%) with symptoms which could suggest bleeding and in 6 of 30 patients (20%) with negative histories. Because of the known rebleeding rate and the increased risk of associated complications, identification of the subgroup who had had haemorrhage and should therefore be considered for surgery may be beneficial. MRI can make a contribution to management by demonstrating prior haemorrhage in patients with an inadequate clinical history.

A gemistocytic astrocytoma demonstrating high intensity on both T1W and T2W MR images is reported. Astrocytoma usually shows low density in T1WI. This peculiar astrocytoma showed no hemorrhage, hemosiderin deposits, melanin or iron. Shortening of T1 relaxation time may be caused by protein hydration layer due to protein rich tumor cells.

We report a patient with multiple angiographically occult vascular malformations in the brain and spine. Magnetic resonance imaging showed multiple lesions in brain and spine with hypointense areas on both T1 and T2-weighted images. These hypointense areas are usually secondary to hemosiderin deposits consistent with remote bleeding in the lesions. We conclude that when magnetic resonance reveals an intraspinal lesion with signal intensity characteristics consistent with a vascular malformation, an examination of the brain should be performed to rule out associated intracranial lesions. The finding of multiple lesions in the brain with identical signal intensity characteristics reinforces the diagnosis of vascular malformation.