Hypointense thrombus on T2-weighted MR imaging: a potential pitfall in the diagnosis of dural sinus thrombosis.

Purpose: To determine the frequency of hypointense appearance of dural sinus thrombosis on T2-weighted images, which may mimic a normal flow void, and when possible correlate with appearance on T1-weighted images. Methods and Materials: Retrospective review of radiology files showed 51 patients with a discharge diagnosis of dural sinus thrombosis who underwent MR imaging during the period 1986-1998. These images were reviewed by an experienced neuroradiologist for appearance on T2-weighted images. This process yielded five cases in which a hypointense appearance on this pulse sequence simulated a normal flow void. An additional two cases were added from the teaching files of two other institutions giving a total of seven cases (13% of studies). The resulting study population consisted of five women and two men (mean age 27.1 years). T1 weighted images were available in five patients. In two patients MR venography was available, but not T1-weighted images. The diagnosis of dural sinus thrombosis was based solely on absence of flow void on T1-weighted images in one case, solely on absence of flow void on MR venography in two cases and absence of flow void on T1-weighted images in conjunction with MR venography or gradient echo findings in 4 patients. All images were obtained on a 1.5 T magnet (GE Medical Systems; Milwaukee, Wisconsin). Results: In all patients hypointense signal of thrombus was isointense with normal flow voids in other dural sinuses on T2-weighted images. In all cases in which T1-weighted images were available, the signal intensity of thrombus was isointense to gray matter. Conclusion: Hypointense appearance of thrombus on T2-weighted images is a potential pitfall in the MR diagnosis of dural sinus thrombosis. Because thrombus in this stage of evolution appears isointense to gray matter on T1-weighted images, careful attention must be paid to other sequences to avoid this pitfall.     

Magnetic resonance imaging of cartilaginous tumors: a retrospective study of 79 patients

Objective: Hyaline cartilaginous tumors are characterized by extremely high signal intensity on T2-weighted images. Recently, some distinctive MR features of cartilaginous bone tumors were reported in small series. Low signal intensity septa surrounding high signal intensity cartilage lobules were seen on T2-weighted images in low-grade chondrosarcomas. On spin-echo T1-weighted images after Gd contrast injection, marked ‘septal’ or ‘ring-and-arc’ enhancement was observed in low-grade chondrosarcomas and enchondromas. The purpose of this study was to determine sensitivity and specificity of these MR findings in diagnosis of cartilaginous tumors, and to assess the value of MR in diagnostic workup of these lesions. Materials and methods: Retrospective evaluation of MR findings in 79 cartilaginous tumors and in 79 non-cartilaginous tumors. All lesions were biopsy proven. Each MR examination was independently reviewed by two experienced radiologists without knowledge of clinical data, radiographic and/or CT findings, or histological diagnosis. All lesions were evaluated for morphology (lobular or non-lobular), presence of a high signal intensity mass on T2-weighted images, presence of low signal intensity septa separating high signal intensity lobules on T2-weighted images, and evidence of septal (‘ring-and-arc’) enhancement.Results: None of the reviewed parameters is useful in diagnosing osteochondromas. Since osteochondromas have a characteristic appearance on plain radiography, the value of MR imaging in the workup of these lesions remains limited. MR findings in enchondromas have a low specificity and a low sensitivity. Low-grade chondrosarcomas, often hard to diagnose on plain radiography and difficult to differentiate from enchondromas, are characterized by the MR tandem of ‘low signal intensity septa on T2-weighted images’ together with ‘septal or ring-and-arc enhancement’ (sensitivity 92.3%, specificity 76.5%). High-grade chondrosarcomas are easily recognized on plain radiography. Conclusions: In differentiating cartilaginous from non-cartilaginous tumors, MR features are highly specific but lack sensitivity. Grading potentials of MR parameters are promising due to the high accuracy in diagnosing low-grade chondrosarcomas.     

MR imaging of the normal and abnormal clivus

We retrospectively reviewed 330 T1-weighted sagittal images, 80 T2-weighted sagittal images, and 83 gadopentetate-dimeglumine-enhanced scans of normal adults to determine the MR appearance of the normal adult clivus. MR images of 21 patients with an abnormal clivus (19 with tumor invasion and two with marrow reconversion) were also evaluated restrospectively and compared with those of the control group to assess MR features distinguishing the two groups. Our study revealed that a normal adult clivus consisted of low- and high-intensity portions mixed in various proportions on T1-weighted images. The low-intensity portion was isointense or hyperintense relative to the pons and always contained foci of bright signal intensity. The low-intensity tumor of a pathologic clivus tended to be hypointense relative to the pons (17/19), and was completely devoid of foci of bright signal intensity. The normal adult clivus was approximately isointense relative to the pons on T2-weighted images. Clival tumors were grossly hyperintense relative to the pons on T2-weighted images in 11 of 17 patients. In the remaining six patients, either a portion of or the entire lesion was isointense relative to the pons and, therefore, was not detectable on T2-weighted images. A normal adult clivus can enhance to some degree (19/83). Clival tumors were found to enhance intensely. A clivus of very low signal intensity (signal void) on T1- or T2-weighted images was always abnormal. The clivus with marrow reconversion was uniformly hypointense relative to the pons on T1-weighted images and isointense relative to normal marrow on T2-weighted images.(ABSTRACT TRUNCATED AT 250 WORDS)     

Cross-sectional imaging of primary osseous hemangiopericytoma

The aim of this study was to assess cross-sectional imaging features and the value of CT and MRI in primary hemangiopericytoma of bone. In five patients with histologically proven primary osseous hemangiopericytoma CT and MR scans were evaluated retrospectively. Both CT and MRI were available in four patients each. In three patients both imaging techniques were available. On CT primary hemangiopericytoma of bone presents as an expansive lytic lesion with bone destruction and inhomogeneous contrast enhancement. Magnetic resonance imaging depicts osseous hemangiopericytoma as hyperintense lesion on T2-weighted images with intermediate signal intensity on T1-weighted images. Curvilinear tubular structures of signal void in the tumor matrix on T1-weighted images and corresponding hyperintense structures on T2-weighted and on fat-suppressed short tau inversion recovery images were present in three patients. Although cross-sectional imaging findings are non-specific, they add to the diagnosis and provide valuable information about the extent of bone destruction and local tumor spread in patients with primary osseous hemangiopericytoma. While CT demonstrates the extent of bone destruction best, MRI better visualizes medullary and soft tissue extension of the tumor. Curvilinear signal abnormalities support the diagnosis of hemangiopericytoma of bone. This imaging pattern is best visualized on fat-suppressed or contrast enhanced T1-weighted MR images.     

MR imaging of intracranial epidermoid tumors: specific diagnosis with Turbo-FLAIR pulse sequence.

The purpose of this study is to present the imaging findings with particular reference to the FLAIR magnetic resonance (MR) pulse sequence, of four patients with epidermoid cysts. In all cases CT showed hypodense lesions with stippled peripheral calcifications and no enhancement. The MR showed low signal intensity lesions in T1-weighted images, high signal intensity in T2-weighted lesions, no or minor enhancement after gadolinium administration and characteristic heterogeneous appearance in FLAIR images. Including the FLAIR sequence in MR imaging, a specific diagnosis can be suggested.     

CT and MR imaging of benign primary cardiac neoplasms with echocardiographic correlation.

Benign primary cardiac neoplasms are rare but may cause significant morbidity and mortality. However, they are usually treatable and can often be diagnosed with echocardiography, computed tomography (CT), or magnetic resonance (MR) imaging. Myxomas typically arise from the interatrial septum from a narrow base of attachment. Fibroelastomas are easily detected at echocardiography as small, mobile masses attached to valves by a short pedicle. Cardiac fibromas manifest as a large, noncontractile, solid mass in a ventricular wall at echocardiography and as a homogeneous mass with soft-tissue attenuation at CT. They are usually homogeneous and hypointense on T2-weighted MR images and isointense relative to muscle on T1-weighted images. Paragangliomas usually appear as large, echogenic left atrial masses at echocardiography and as circumscribed, heterogeneous masses with low attenuation at CT. These tumors are usually markedly hyperintense on T2-weighted MR images and iso- or hypointense relative to myocardium on T1-weighted images. Cardiac lipomas manifest at CT as homogeneous, low-attenuation masses in a cardiac chamber or in the pericardial space and demonstrate homogeneous increased signal intensity that decreases with fat-saturated sequences at T1-weighted MR imaging. Cardiac lymphangiomas manifest as cystic masses at echocardiography and typically demonstrate increased signal intensity at T1- and T2-weighted MR imaging. Familiarity with these imaging features and with the relative effectiveness of these modalities is essential for prompt diagnosis and effective treatment.     

Radiologic-Pathologic Correlation of Unusual Lingual Masses: Part II: Benign and Malignant Tumors

Because the tongue is superficially located and the initial manifestation of most diseases occurring there is mucosal change, lingual lesionscan be easily accessed and diagnosed without imaging analysis. Some lingual neoplasms, however, may manifest as a submucosal bulge and be located in a deep portion of the tongue, such as its base; their true characteristics and extent may be recognized only on cross-sectional images such as those obtained by CT or MRI.Some uncommon tongue neoplasms may have characteristic radiologic features, thus permitting quite specific radiologic diagnosis. Lipomas typically manifest at both CT and MR imaging as homogeneous nonenhancing lesions. Relative to subcutaneous fat they are isoattenuating on CT images, and all MR sequences show them as isointense. Due to the paramagnetic properties of melanin, metastases from melanotic melanoma usually demonstrate high signal intensity on T1-weighted MR images and low signal intensity on T2-weighted images.Although the radiologic findings for other submucosal neoplasms are nonspecific, CT and MR imaging can play an important role in the diagnostic work-up of these unusual tumors. Delineation of the extent of the tumor, and recognition and understanding of the spectrum of imaging and the pathologic features of these lesions, often help narrow the differential diagnosis.     

MR imaging of the normal and abnormal clivus

We retrospectively reviewed 330 T1-weighted sagittal images, 80 T2-weighted sagittal images, and 83 gadopentetate-dimeglumine-enhanced scans of normal adults to determine the MR appearance of the normal adult clivus. MR images of 21 patients with an abnormal clivus (19 with tumor invasion and two with marrow reconversion) were also evaluated retrospectively and compared with those of the control group to assess MR features distinguishing the two groups. Our study revealed that a normal adult clivus consisted of low- and high-intensity portions mixed in various proportions on T1-weighted images. The low-intensity portion was isointense or hyperintense relative to the pons and always contained foci of bright signal intensity. The low-intensity tumor of a pathologic clivus tended to be hypointense relative to the pons (17/19), and was completely devoid of foci of bright signal intensity. The normal adult clivus was approximately isointense relative to the pons on T2-weighted images. Clival tumors were grossly hyperintense relative to the pons on T2-weighted images in 11 of 17 patients. In the remaining six patients, either a portion of or the entire lesion was isointense relative to the pons and, therefore, was not detectable on T2-weighted images. A normal adult clivus can enhance to some degree (19/83). Clival tumors were found to enhance intensely. A clivus of very low signal intensity (signal void) on T1- or T2-weighted images was always abnormal. The clivus with marrow reconversion was uniformly hypointense relative to the pons on T1-weighted images and isointense relative to normal marrow on T2-weighted images. The intensity patterns of the normal clivus on T1- and T2-weighted MR images change predictably with advancing age. Intensity patterns of abnormal clivi differ from those of normal clivi. When contrast material is used, normal and abnormal clivi generally show different patterns of enhancement.     

MR imaging findings of Rathke's cleft cysts: significance of intracystic nodules

BACKGROUND AND PURPOSE: Rathke's cleft cysts often may be difficult to differentiate from other intrasellar or suprasellar masses on radiologic studies. The purpose of this study was to describe the significance of intracystic nodules, a diagnostic characteristic found in Rathke's cleft cysts, on MR images. METHODS: A retrospective review of MR studies was conducted for 13 patients who, after pathologic analysis, were diagnosed as having Rathke's cleft cyst. These patients underwent unenhanced and contrast-enhanced T1- and T2-weighted axial and coronal spin-echo sequential imaging. The signal intensity and incidence of the intracystic nodules on T1- and T2-weighted images were analyzed. The signal intensity of the nodule was compared with that of white matter and surrounding cyst fluid. The signal intensity of cyst fluid was compared with the intraoperative appearance of the cyst fluid. Biochemical and pathologic analyses of the intracystic nodules were conducted in two cases. RESULTS: An intracystic nodule having high signal intensity on T1-weighted images and low signal intensity on T2-weighted images was observed in 10 (77%) of the cases. At surgery, intracystic nodules were yellow, waxy, solid masses. Pathologic analysis showed this nodule to be a mucin clump. Biochemical analysis of the intracystic nodules showed cholesterol and proteins as the main constituents. In the Rathke's cleft cyst with intracystic nodules, cyst fluid revealed low signal intensity to isointensity relative to the intensity of the nodules on T1-weighted images, and isointensity to high signal intensity on T2-weighted images. Intracystic nodules were clearly visible on T2-weighted images. CONCLUSION: Because cyst fluid of Rathke's cleft cysts shows variable intensities on MR images, the specific diagnosis is often difficult when based on MR signal intensity values alone. The presence of an intracystic nodule with characteristic signal intensities on MR images may be indicative of the diagnosis of Rathke's cleft cyst.     

Uterine contractions: Possible diagnostic pitfall at MR imaging

A total of 206 nongravid patients with various gynecologic problems underwent pelvic magnetic resonance (MR) examinations that included both sagittal T2-weighted and contrast agent–enhanced T1-weighted images. MR images were retrospectively reviewed to identify changes in endometrial configuration on serial images obtained during the same MR examination. In 20 MR examinations (all in women of reproductive age), endometrial distortion due to myometrial bulging was noted on T2-weighted or contrast-enhanced T1-weighted images. It was absent on other MR images obtained at different times. Myometrial bulging exhibited low signal intensity in 18 examinations. The finding resembled adenomyosis or leiomyoma on T2-weighted or contrast-enhanced T1-weighted images. These results evidence the presence of transient myometrial bulging and transient low-intensity myometrium in the nongravid uterus. This phenomenon is thought to represent uterine contraction. Clinicians should be aware of the potential presence of transient low-signal-intensity myometrial bulging that could present diagnostic problems in the normal uterus.     

Relationship between MR imaging and histopathologic findings of the brain in extremely sick preterm infants.

BACKGROUND AND PURPOSE: MR imaging can now be used safely in extremely preterm infants. The aim of this study was to compare the MR imaging appearance of the immature brain with neuropathologic findings at postmortem examination. METHODS: Seven extremely sick preterm infants, born at a median of 24 weeks' gestation, were studied using T1- and T2-weighted MR sequences. Infants died at a median of 3 days after initial MR imaging, and postmortem examinations were carried out. RESULTS: The cortex and germinal matrix were seen as areas of low signal intensity on T2-weighted images, which corresponded to their highly cellular histologic appearance. The periventricular and subcortical layers of white matter had a high signal intensity, corresponding to high fiber and relatively low cellular density; the intermediate layer of low signal intensity corresponded to a dense band of migrating cells. Regions of acute hemorrhage were seen as low signal intensity and regions of infarction as high signal intensity on T2-weighted images. One infant with mild periventricular leukomalacia had some low signal intensity on T1-weighted images, but no focal changes on T2-weighted images. Regions of neuronal mineralization, seen in association with infarction and capillary proliferation, within the basal ganglia and thalami were characterized by very low signal intensity on T2-weighted images and by very high signal intensity on T1-weighted images. There were no imaging abnormalities detected in regions with more subtle histologic abnormalities, such as increased glial or apoptotic cells. CONCLUSION: MR imaging can be used to observe normal developing brain anatomy in extremely premature infants; it can detect areas of hemorrhage and infarction within the developing brain, but conventional MR imaging may not detect more subtle histologic abnormalities.     

Brodie Abscess: MR imaging appearance in 10 patients

The magnetic resonance (MR) imaging features of Brodie abscess have not yet been fully evaluated. Ten patients with Brodie abscess, eight of long bone and two of vertebra, were studied with MR imaging. Long bone abscess had a characteristic “target” appearance with four layers: (a) a center with low signal intensity on T1-weighted images and high signal intensity on T2-weighted and STIR (short-inversion-time inversion recovery) images, (b) an inner ring isointense to muscle on T1-weighted images and with high signal intensity on T2-weighted and STIR images, (c) an outer ring hypoin-tense on all images, and (d) a peripheral halo hypointense on T1-weighted images. In six of eight cases, a soft-tissue mass was found. The two vertebral abscesses had a less specific appearance, with low signal intensity on T1-weighted images and high signal intensity on T2-weighted and STIR images. Only the peripheral halo was clearly identified in both cases.     

MR imaging of intracranial tuberculomas

Eight patients with intracranial tuberculomas were studied with CT and magnetic resonance (MR) imaging. Large, ring enhancing, solid lesions on CT showed low intensity on T2-weighted images and intermediate intensity on T1-weighted images. Small lesions, with ring enhancement on CT, showed central bright signal on T2-weighted images with a peripheral low intensity rim surrounded by high intensity edema. The MR imaging features of the tuberculomas were found to be distinct from those of abscesses, metastases, and gliomas.     

Normal adrenal gland: in vivo observations, and high-resolution in vitro chemical shift MR imaging-histologic correlation

RATIONALE AND OBJECTIVES: The purpose of this study was to determine whether adrenal cortical lipid affects signal intensity on magnetic resonance (MR) images and to evaluate contrast between cortex and medulla. MATERIALS AND METHODS: From their clinical database, the authors selected 37 MR imaging studies of patients with adrenal adenomas. Two independent readers compared in-phase and fat-suppressed T1-weighted images, looking for visible lipid-induced signal intensity loss in the adrenal gland. Six adrenal gland specimens obtained after radical nephrectomy were also studied with high-resolution MR imaging, including in-phase, opposed-phase, and fat-suppressed T1-weighted images, and T2-weighted images. Adjacent histologic sections were stained with oil red O for neutral fats and with hematoxylin-eosin, and they were also viewed with polarization light microscopy. The relative amount of lipid was graded as mild, moderate, or intense, and the appearance of the cortex and medulla was compared with that on the MR images. RESULTS: On the 37 clinical MR studies, there was no visible signal intensity loss within the limbs of the ipsilateral adrenal glands. T2-weighted images of the adrenal specimens showed a thin high-intensity band, corresponding to the appearance of medulla on histologic slices. This could not be seen on any of the T1-weighted images. Region-of-interest measurements were nearly identical for in-phase and opposed-phase images. Histologic analysis showed abundant cortical lipid. CONCLUSION: Adrenal corticomedullary contrast can be depicted on high-resolution T2-weighted images but not on any T1-weighted images. There is abundant cortical lipid in adrenal specimens, but comparison of in-phase with opposed-phase MR images does not depict it.     

MR imaging of tuberculous arthritis: Clinical and experimental studies

Signal-intensity characteristics of magnetic resonance (MR) images were assessed in five patients and in 10 rabbits with tuberculous arthritis. MR imaging findings were compared with histologic findings in the animal study. In both of clinical and experimental cases, tuberculous lesions showed both intermediate and high signal intensity on T2-weighted images, while they showed heterogeneously intermediate signal intensity on T1-weighted images. As T2-weighted images were compared with pathologic specimens, intermediate signal intensity corresponded to caseous necrosis, whereas high signal intensity related to granulomas or effusion. Postcontrast T1-weighted images showed enhancement at the peripheries rather than the centers of tuberculous lesions. These results indicate that tuberculous arthritis should be included in the differential diagnosis when intra-articular lesions with low or intermediate signal intensity are found on T2-weighted images.     

Giant cell tumor of the proximal tibia: MR and CT appearance

The magnetic resonance (MR) appearance of four cases of giant cell tumor (GCT) of the proximal tibia are described and the MR grading of these tumors is compared with CT and conventional radiography. Magnetic resonance showed the lesions to be well defined with respect to adjacent marrow and cortical bone. Homogeneous intermediate signal intensity or low signal within the tumors was seen on T1-weighted images. T2-weighted images showed mixed signal intensity with small "bright patches" of increased signal intensity in all four cases. No fluid levels were identified. Magnetic resonance was superior to CT and plain radiography in radiologic grading of the tumors. Computed tomography was superior in determining if cortical invasion was present. Intraarticular tumor extension was more accurately detected by MR and arthrotomography than CT. An MR manifestation of GCT of the proximal tibia is described which may be a common appearance of this tumor by this modality. Magnetic resonance is the procedure of choice in the radiologic grading of GCT.     

Giant intracranial aneurysms: MR imaging

Fifteen patients with giant intracranial aneurysms were examined with magnetic resonance (MR) imaging, computed tomography (CT), and angiography. MR imaging revealed a rounded extraaxial mass with evidence of intraluminal blood flow in 12 of the 15 cases. Signal void within the lumen was seen in all 12 of these cases. Other flow effects, such as even-echo rephasing, were seen in some cases. Complete thrombosis was seen in three giant aneurysms in which high signal intensity on T1- and T2-weighted images filled the lumen, and no flow effects were seen. Partial thrombosis was detected in four aneurysms as persistent areas of medium to high signal intensity within the lumen on T1- and T2-weighted images. In five cases, the cause of the intraluminal signal that was present was difficult to determine. The ability of MR to indicate flow within the lumen in 12 of 15 cases provided for greater diagnostic confidence than with CT, which was strongly suggestive of an extraaxial tumor in nine patients. Calcification of the wall of the aneurysm was better detected with CT.     

MR and CT of lacunar infarcts

Twenty-two patients with clinical signs and symptoms compatible with lacunar transient ischemic attack or stroke of varying chronicity were evaluated with MR imaging. CT was also performed in 21 of these patients. MR revealed small, deep cerebral lesions in locations appropriate to the clinical symptoms in 19 patients. Lacunar infarcts were imaged by CT in 11 patients; however, no lesions were identified on CT that were not detected with MR. Presumed lacunar infarcts were identified on MR images in 17 additional patients. Lacunae generally appeared as focal areas of decreased signal intensity on T1-weighted images and as focal areas of increased signal intensity on T2-weighted images. T2-weighted MR images detected a greater number of lacunar infarcts than did mixed T1-/T2-weighted images, which in turn detected more lacunae than did T1-weighted images. In general, acute lacunar infarcts (within 1 week of onset or recurrence of clinical symptoms) were seen only on T2-weighted images, while chronic lesions (more than 1 week) were seen on both T1- and T2-weighted images. Our results indicate that MR is superior to CT for evaluating lacunar infarcts, and second, that T2-weighted images are more sensitive than T1- and mixed T1-/T2-weighted images for detecting lacunar infarcts.     

CT and MR imaging of advanced Zollinger-Ellison syndrome

The CT and magnetic resonance (MR) findings in 13 patients with advanced Zollinger-Ellison syndrome are described. In eight patients (62%) one or more primary tumors were found with both methods. All patients with proven liver metastases (n = 7) were identified by MR. Computed tomography was positive in six of these patients. Three patients with lymph node metastases were identified on CT and MR and one patient had bone metastases. Computed tomography and MR were inferior to selective arteriography in the detection of multiple lesions of the pancreas in a patient with multiple endocrine neoplasia-I syndrome. On the T1-weighted MR images, the primary tumors demonstrated no consistency with regard to their signal intensity relative to the adjacent pancreatic parenchyma. All gastrinomas had an increased relative signal intensity on the T2-weighted images with the exception of a calcified tumor. Liver and lymph node metastases had a low signal intensity on the T1-weighted images and an increased signal intensity on the T2-weighted images. The signal intensity of primary tumors and metastases was independent of size. In conclusion, MR was able to detect abnormalities based on its outstanding lesion/normal tissue contrast, whereas CT diagnosis was based mostly on contour distortion. For the current technique, MR is considered at least equal to CT.     

MR imaging findings in patients with bone-chip allografts

Increasing use of MR to evaluate primary bone neoplasms has stimulated numerous articles on the initial assessment of these lesions. However, scant MR literature has been written about the postoperative MR appearance. We studied the MR appearance of lesions treated by curettage followed by packing with cancellous bone-chip allograft. We retrospectively reviewed the MR scans of 18 consecutive patients treated in this way. Pathologic diagnoses of these lesions were confirmed according to commonly accepted criteria. The allograft sites showed a distinctive pattern of speckled bright signal on T1-weighted images in eight cases. Thirteen of 18 grafts showed a whorled or speckled pattern of increased signal on the T2-weighted images. Only four allograft regions had predominantly low signal on both T1- and T2-weighted images. Two of 18 patients had recurrent tumor proved by open biopsy. MR images in these cases showed areas of homogeneous signal that replaced areas of speckled hyperintensity on both T1- and T2-weighted images. In both these recurrences the tumor had signal intensity similar to that seen on the preoperative MR study. We conclude that knowledge of the MR appearance of cancellous bone chip allografts is important to avoid misinterpreting areas of high signal on T1- or T2-weighted images as areas of recurrence of tumor and/or hemorrhage.