Diffusion imaging of the human brain in vivo using high-speed STEAM MRI.

This paper describes a new method for diffusion imaging of the human brain in vivo that is based on a combination of diffusion-encoding gradients with high-speed STEAM MR imaging. The single-shot sequence 90 degrees-TE/2-90 degrees-TM-(alpha-TE/2-STE)n generates n = 32-64 differently phase-encoded stimulated echoes STE yielding image acquisition times of 576 ms for a 48 x 128 data matrix. Diffusion encoding is performed during the first TE/2-interval as well as during each readout period. Phantom studies reveal a quantitative agreement of calculated diffusion coefficients with literature values. EKG triggering completely eliminates motion artifacts from diffusion-weighted single-shot STEAM images of human brain in vivo. While signal attenuation of the cerebrospinal fluid (CSF) is predominantly due to flow, that observed for gray and white matter results from diffusion. Evaluated diffusion coefficients yield (1.0 +/- 0.1) x 10(-5) cm2 s-1 for gray matter, (0.5 +/- 0.1) x 10(-5) cm2 s-1 for white matter with the diffusion encoding parallel to the main orientation of the myelin sheath of the neurofibrils, and (0.3 +/- 0.1) x 10(-5) cm2 s-1 for white matter and a perpendicular orientation. All studies were performed at 2.0 T using a conventional 10 mT m-1 gradient system.     

Multipurpose NMR imaging using stimulated echoes

STEAM (stimulated-echo acquisition mode) imaging techniques recently introduced by the authors are demonstrated to provide a versatile tool for improving the parametric specificity in NMR imaging. Stimulated echoes can be excited by a sequence of at least three rf pulses with flip angles of 90 degrees or less. The main characteristics of the STEAM method are based on the great functional flexibility of an imaging sequence comprising three rf pulses unequal to 180 degrees and three intervals prior to acquisition of the data. Major advantages are the easy access to contiguous multiplanar images, to CHESS (chemical-shift-selective) images, and to T1 information. Moreover, the rf power deposition is considerably reduced as compared to spin-echo NMR imaging sequences. Here first in vivo results on human extremities are presented including contiguous multislice images, multiple CHESS images, and spin-lattice relaxation time images calculated from a series of simultaneously recorded T1-weighted STEAM images.     

Rapid isotropic diffusion mapping without susceptibility artifacts: whole brain studies using diffusion-weighted single-shot STEAM MR imaging.

A subsecond magnetic resonance imaging (MRI) technique for isotropic diffusion mapping is described which, in contrast to echo-planar imaging (EPI), is insensitive to resonance offsets, i.e., tissue susceptibility differences, magnetic field inhomogeneities, and chemical shifts. It combines a diffusion-weighted (DW) spin-echo preparation period and a high-speed stimulated echo acquisition mode (STEAM) MRI sequence and yields single-shot images within measuring times of 559 msec (80 echoes). Here, diffusion encoding involved one scan without DW, three DW scans with b = 490 sec mm(-2), and three DW scans with b = 1000 sec mm(-2) (orthogonal gradient orientations). An automated on-line evaluation resulted in isotropic DW images as well as ADC maps (trace of the diffusion tensor). Experiments at 2.0 T covered the brain of healthy subjects in 20 contiguous sections of 6 mm thickness and 2.0 x 2.0 mm(2) in-plane resolution within a total measuring time of 78 sec. High-resolution studies at 1.0 x 1.0 mm(2) (interpolated from 2.0 x 1.0 mm(2) acquisitions) were obtained within 5 min 13 sec using four averages. In comparison with EPI, DW single-shot STEAM MRI exhibits only about half the SNR, but completely avoids regional signal losses, high intensity artifacts, and geometric distortions.     

Burst excitation for quantitative diffusion imaging with multiple b-values.

A quantitative imaging sequence has been developed to exploit the intrinsic sensitivity of Burst NMR data to molecular diffusion. In the scan time of a single spin echo experiment, it is possible to acquire many images of the same slice, with a different T(2) and diffusion weighting. Under favorable conditions, it is possible to obtain both the diffusion coefficient and T(2) from the same experiment; or, by correcting for T(2) relaxation using a control image, more precise diffusion coefficients may be measured. The quantitative values in rat brain are in agreement with those from conventional experiments. The major gains of this method are the potentially reduced scan time, the higher number of acquired images corresponding to different diffusion weightings, the reduced sensitivity to inter-scan motion artifact and to local variations in magnetic susceptibility, and an automatic co-registration between T(2) and diffusion images. Problems with the sequence include a lower signal-to-noise ratio than is achievable with diffusion-weighted spin-echo imaging, the limitation of measuring only in-plane components of diffusion and, at present, single-slice acquisition.     

Diffusion-prepared fast imaging with steady-state free precession (DP-FISP): a rapid diffusion MRI technique at 7 T

Diffusion MRI is a useful imaging technique with many clinical applications. Many diffusion MRI studies have utilized echo-planar imaging (EPI) acquisition techniques. In this study, we have developed a rapid diffusion-prepared fast imaging with steady-state free precession MRI acquisition for a preclinical 7T scanner providing diffusion-weighted images in less than 500 ms and diffusion tensor imaging assessments in ～1 min with minimal image artifacts in comparison with EPI. Phantom apparent diffusion coefficient (ADC) and fractional anisotropy (FA) assessments obtained from the diffusion-prepared fast imaging with steady-state free precession (DP-FISP) acquisition resulted in good agreement with EPI and spin echo diffusion methods. The mean apparent diffusion coefficient was 2.0 × 10(-3) mm(2) /s, 1.90 × 10(-3) mm(2) /s, and 1.97 × 10(-3) mm(2) /s for DP-FISP, diffusion-weighted spin echo, and diffusion-weighted EPI, respectively. The mean fractional anisotropy was 0.073, 0.072, and 0.070 for diffusion-prepared fast imaging with steady-state free precession, diffusion-weighted spin echo, and diffusion-weighted EPI, respectively. Initial in vivo studies show reasonable ADC values in a normal mouse brain and polycystic rat kidneys.     

Diffusion imaging with the MP-rage sequence

Diffusion-weighted magnetic resonance (MR) images obtained with conventional spin-echo techniques are known to be sensitive to subject motion because of long image acquisition times. To reduce the acquisition time, use of a magnetization-prepared rapid gradient-echo (MP-RAGE) sequence modified for diffusion sensitivity was studied. In this sequence, a preparation phase with a 90°–180°–90° pulse train is used to sensitize the magnetization to diffusion. Centric-ordered phase encoding, short TRs (5.2–6.5 msec), and small flip angles (5°–8°) are necessary to minimize saturation effects from tissues with short relaxation times. Phantom studies with various concentrations of copper sulfate (T1 ranging from 2,459 to 90 msec) were performed to validate that the diffusion-weighted signal obtained with the MP-RAGE sequence was independent of relaxation time. Diffusion-weighted images of water, isopropyl alcohol, and acetone were acquired to confirm the accuracy of measured diffusion coefficients. Brain images of healthy normal volunteers were obtained to demonstrate motion insensitivity and general image quality of the technique. The results indicate that accurate diffusion-weighted images can be obtained with a diffusion-weighted MP-RAGE sequence, with imaging times of about 1 second.     

Stimulated-echo acquisition mode (STEAM) MRI for black-blood delayed hyperenhanced myocardial imaging

PURPOSE: To develop a breathhold method for black-blood viability imaging of the heart that may facilitate identifying the endocardial border. MATERIALS AND METHODS: Three stimulated-echo acquisition mode (STEAM) images were obtained almost simultaneously during the same acquisition using three different demodulation values. Two of the three images were used to construct a black-blood image of the heart. The third image was a T(1)-weighted viability image that enabled detection of hyperintense infarcted myocardium after contrast agent administration. The three STEAM images were combined into one composite black-blood viability image of the heart. The composite STEAM images were compared to conventional inversion-recovery (IR) delayed hyperenhanced (DHE) images in nine human subjects studied on a 3T MRI scanner. RESULTS: STEAM images showed black-blood characteristics and a significant improvement in the blood-infarct signal-difference to noise ratio (SDNR) when compared to the IR-DHE images (34 +/- 4.1 vs. 10 +/- 2.9, mean +/- standard deviation (SD), P < 0.002). There was sufficient myocardium-infarct SDNR in the STEAM images to accurately delineate infarcted regions. The extracted infarcts demonstrated good agreement with the IR-DHE images. CONCLUSION: The STEAM black-blood property allows for better delineation of the blood-infarct border, which would enhance the fast and accurate measurement of infarct size.     

SPLICE: Sub-second diffusion-sensitive MR imaging using a modified fast spin-echo acquisition mode

Recent human studies for measuring of the apparent diffusion coefficient in tissue by magnetic resonance imaging have been conducted by time-consuming standard spin-echo acquisition sequences and phase correction with navigator echoes. Diffusion-weighted echo-planar sequences have been shown to be rapid alternatives for brain imaging. Both methods show inherent disadvantages in applications on thoracic or abdominal sites. A new approach combining single-shot diffusion-weighted imaging with a modified fast spin-echo acquisition mode is reported here. The modification is necessary, because normal fast spin-echo acquisition requires a particular phase relation between the magnetization and the refocusing pulses. Unfortunately, this phase relation is not provided after diffusion sensitive preparation. Therefore, the split echo acquisition mode was developed and is shown to be insensitive to the phase of the magnetization. The advantages of both fast spin-echo acquisition and diffusion weighting can be combined in the SPLICE sequence (split aqcuisition of fast spin-echo signals for diffusion imaging). The applicability of the new technique is shown by series of sub-second diffusion-weighted images from different parts of the body.     

MR imaging using stimulated echoes (STEAM)

The introduction of STEAM (stimulated echo acquisition mode) magnetic resonance (MR) sequences provides access to a variety of MR parameters. T1-weighted and calculated T1 proton MR images of the head of healthy volunteers and a patient with an astrocytoma are presented. MR examinations were performed with a 2.0-T whole-body system. The STEAM T1 method can be used to characterize multiexponential relaxation behavior, to evaluate T1 relaxation times, and to improve the T1 contrast within MR images. Both the measuring time and the spatial resolution are the same as for a conventional image.     

Molecular self-diffusion of intracellular metabolites in rat brain in vivo Investigated by Localized Proton NMR Diffusion Spectroscopy

Molecular self-diffusion coefficients of water (0.75 ± 0.05), Nacetylaspartate (0.27 ± 0.04), creatines (0.27 ± 0.04), and cholines (0.28 ± 0.08) × 10^?5 cm² s^?1 were obtained from localized proton NMR spectra of rat brain in vivo using diffusion-weighted stimulated-echo (STEAM) sequences with a diffusion time of (Δ ? δ/3) = 17 ms.     

Diffusion-weighted MR imaging of intracerebral masses: comparison with conventional MR imaging and histologic findings

BACKGROUND AND PURPOSE: The purposes of this study were to find the role of diffusion-weighted MR imaging in characterizing intracerebral masses and to find a correlation, if any, between the different parameters of diffusion-weighted imaging and histologic analysis of tumors. The usefulness of diffusion-weighted imaging and apparent diffusion coefficient (ADC) maps in tumor delineation was evaluated. Contrast with white matter and ADC values for tumor components with available histology were also evaluated. METHODS: Twenty patients with clinical and routine MR imaging/CT evidence of intracerebral neoplasm were examined with routine MR imaging and echo-planar diffusion-weighted imaging. The routine MR imaging included at least the axial T2-weighted fast spin-echo and axial T1-weighted spin-echo sequences before and after contrast enhancement. The diffusion-weighted imaging included an echo-planar spin-echo sequence with three b values (0, 300, and 1200 s/mm(2)), sensitizing gradient in the z direction, and calculated ADC maps. The visual comparison of routine MR images with diffusion-weighted images for tumor delineation was performed as was the statistical analysis of quantitative diffusion-weighted imaging parameters with histologic evaluation. RESULTS: For tumors, the diffusion-weighted images and ADC maps of gliomas were less useful than the T2-weighted spin-echo and contrast-enhanced T1-weighted spin-echo images in definition of tumor boundaries. Additionally, in six cases of gliomas, neither T2-weighted spin-echo nor diffusion-weighted images were able to show a boundary between tumor and edema, which was present on contrast-enhanced T1-weighted and/or perfusion echo-planar images. The ADC values of solid gliomas, metastases, and meningioma were in the same range. In two cases of lymphomas, there was a good contrast with white matter, with strongly reduced ADC values. For infection, the highest contrast on diffusion-weighted images and lowest ADC values were observed in association with inflammatory granuloma and abscess. CONCLUSION: Contrary to the findings of previous studies, we found no clear advantage of diffusion-weighted echo-planar imaging in the evaluation of tumor extension. The contrast between gliomas, metastases, meningioma, and white matter was generally lower on diffusion-weighted images and ADC maps compared with conventional MR imaging. Unlike gliomas, the two cases of lymphomas showed hyperintense signal on diffusion-weighted images whereas the case of cerebral abscess showed the highest contrast on diffusion-weighted images with very low ADC values. Further study is required to find out whether this may be useful in the differentiation of gliomas and metastasis from lymphoma and abscess.     

Diffusion imaging of the in vivo heart using spin echoes--considerations on bulk motion sensitivity.

Cardiac diffusion MRI based on stimulated-echo acquisition mode (STEAM) techniques is hampered by its inherent low signal-to-noise ratio (SNR) efficiency. Diffusion imaging using standard spin-echo (SE) techniques, on the other hand, offers higher SNRs but has been considered impractical for the beating heart due to excessive signal attenuation from cardiac bulk motion. In this work the effect of systolic cardiac motion on different diffusion-encoding schemes was studied in detail. Numerical simulations based on in vivo motion data (acquired by MRI tagging techniques) demonstrate an up to 10-fold decrease in bulk motion sensitivity of the diffusion encoding if the first-order moment of the diffusion-encoding gradients is nullified. It is shown that the remaining systolic phase pattern on the myocardium does not influence the magnitude images if the spatial resolution is chosen to be higher than 4 mm. Given these relatively low resolution requirements, we obtained in vivo diffusion-weighted (DW) short-axis images from four healthy volunteers using an SE-based diffusion-encoding sequence with excitation and refocusing in orthogonal planes for field of view (FOV) reduction. The results showed no significant signal loss due to cardiac motion, and the direction of the principal eigenvalues was found to be in good agreement with known myocardial fiber orientation.     

Magnetic resonance manifestations of decidualized endometriomas during pregnancy

OBJECTIVE: Decidual changes of endometrial tissue in endometriomas during pregnancy may manifest as mural nodules and mimic malignant transformation. We evaluated magnetic resonance findings of decidualized ovarian endometriomas for the differentiation from malignant transformation. METHODS: Magnetic resonance manifestations of 5 decidualized ovarian endometriomas were evaluated. High b-value diffusion-weighted images were performed in 3 cases. RESULTS: The mural nodules were demonstrated as linear, small nodular, broad-based nodular, or polypoid structures, which showed prominent hyperintensity on T2-weighted images and on diffusion-weighted images similar to normal endometrium or placenta. The apparent diffusion coefficient of decidualized mural nodules was significantly higher than that of malignant mural nodules of 7 ovarian cancers. CONCLUSIONS: Endometrioma with prominent hyperintense mural nodules on T2-weighted images in a pregnant woman is highly suggestive for decidualized endometriomas. Diffusion-weighted images with the apparent diffusion coefficient measurement may help the diagnosis of this rare entity.     

Diffusion-weighted MR of acute cerebral infarction: comparison of data processing methods.

BACKGROUND AND PURPOSESome investigators have proposed that either calculated diffusion trace images or apparent diffusion coefficient (ADC) maps, which require imaging with multiple diffusion sensitivities and/or postacquisition image processing, are essential for the accurate interpretation of diffusion-weighted images in acute stroke because of the possible pitfalls of regional diffusion anisotropy, magnetic susceptibility artifacts, and confounding T2 effects, all of which alter signal on diffusion-weighted MR images. The purpose of our study was to compare the sensitivity, specificity, and accuracy of simple, orthogonal-axis diffusion-weighted imaging for the diagnosis of early cerebral infarction with three other sets of postacquisition-processed images: isotropic diffusion-weighted, diffusion trace-weighted, and diffusion trace images.METHODSTwenty-six consecutive adult patients with signs and symptoms consistent with a clinical diagnosis of early cortical and/or subcortical cerebral infarction and 17 control subjects were studied with multisection, single-shot, spin-echo echo-planar diffusion-weighted imaging at 1.5 T to generate a set of three orthogonal-axis diffusion-weighted images. Isotropic diffusion-weighted, diffusion trace-weighted, and diffusion trace (mean ADC) images were then generated off-line and all four sets of images were interpreted blindly by two neuroradiologists.RESULTSThe average sensitivity, specificity, and accuracy for the orthogonal-axis diffusion-weighted images were 98.1%, 97.1%, and 97.7%, respectively. The average sensitivity, specificity, and accuracy for isotropic diffusion-weighted images were 88.5%, 100%, and 93% respectively. The average sensitivity, specificity, and accuracy for diffusion trace-weighted images were 82.7%, 73.6%, and 79.1%, respectively. The average sensitivity, specificity, and accuracy for diffusion trace images were 50.0%, 85.3%, and 64.0%, respectively.CONCLUSIONOrthogonal-axis diffusion-weighted images have the highest sensitivity and accuracy and very high specificity for early cerebral infarction. Our data contradict the contention that quantitative diffusion maps, requiring imaging with multiple diffusion sensitivities and/or subsequent image processing, are necessary for clinical stroke imaging.     

Restricted diffusion within ring enhancement is not pathognomonic for brain abscess.

BACKGROUND AND PURPOSE: Recent experience suggests that diffusion-weighted MR imaging may be decisive in the differential diagnosis of ring-enhancing cerebral lesions. Whether restricted diffusion within a ring-enhancing cerebral mass lesion is pathognomonic for abscess was studied. METHODS: Seventeen patients with ring-enhancing cerebral lesions (three abscesses, six glioblastomas, eight metastases) on conventional contrast-enhanced T1-weighted images were examined with echo-planar diffusion-weighted MR imaging. Apparent diffusion coefficient (ADC) maps and the ADCs were calculated for all lesions. Lesions with signs of intralesional hemorrhage on unenhanced T1-weighted images were excluded. RESULTS: The central portion of all six glioblastomas and seven of eight metastases showed unrestricted diffusion, whereas two of three abscesses showed restricted diffusion (low ADC values) in their cavity. However, restricted diffusion also was found in one metastasis, and one abscess within a postoperative cavity showed unrestricted diffusion within a larger nondependent portion. CONCLUSION: In patients with ring-enhancing cerebral mass lesions, restricted diffusion might be characteristic but is not pathognomonic for abscess, as low ADC values also may be found in brain metastases.     

Eddy current correction in diffusion-weighted imaging using pairs of images acquired with opposite diffusion gradient polarity.

In echo-planar-based diffusion-weighted imaging (DWI) and diffusion tensor imaging (DTI), the evaluation of diffusion parameters such as apparent diffusion coefficients and anisotropy indices is affected by image distortions that arise from residual eddy currents produced by the diffusion-sensitizing gradients. Correction methods that coregister diffusion-weighted and non-diffusion-weighted images suffer from the different contrast properties inherent in these image types. Here, a postprocessing correction scheme is introduced that makes use of the inverse characteristics of distortions generated by gradients with reversed polarity. In this approach, only diffusion-weighted images with identical contrast are included for correction. That is, non-diffusion-weighted images are not needed as a reference for registration. Furthermore, the acquisition of an additional dataset with moderate diffusion-weighting as suggested by Haselgrove and Moore (Magn Reson Med 1996;36:960-964) is not required. With phantom data it is shown that the theoretically expected symmetry of distortions is preserved in the images to a very high degree, demonstrating the practicality of the new method. Results from human brain images are also presented.     

Black-blood imaging of the human heart using rapid stimulated echo acquisition mode (STEAM) MRI

PURPOSE: To develop a rapid stimulated echo acquisition mode (STEAM) MRI technique for "black-blood" imaging of the human heart that overcomes the single-slice limitation and partially compromised blood suppression associated with double inversion-recovery techniques. MATERIALS AND METHODS: Black-blood multislice images of the heart along anatomic orientations and triggered to end diastole were obtained from healthy human subjects at 3T using rapid STEAM MRI sequences with five-eighths partial Fourier encoding and variable flip angles. Single-shot STEAM images at 2.5 x 2.5 mm2 in-plane resolution and 6-mm section thickness were recorded in 230 msec from individual heartbeats. Improved signal-to-noise ratio (SNR) and higher spatial resolution of 2.0 x 2.0 mm2 and 1.5 x 1.5 mm2 were achieved by segmented multishot STEAM MRI with interleaved k-space acquisitions (160 msec each) from several heartbeats. In a single breathhold covering 18 heartbeats selected applications employed either three segments with six sections or six segments with three sections. RESULTS: Because stimulated echoes (STEs) dephase signals from moving spins, rapid STEAM images are free from blood signal contamination. The method offers a flexible tradeoff between spatial resolution, imaging speed (i.e., number of segments), and volume coverage (i.e., number of sections). CONCLUSION: Rapid STEAM MRI of the heart emerges as a simple technique for multislice imaging of the myocardial wall with efficient flow suppression.     

Imaging articular cartilage under compression--cartilage elastography.

We constructed a device to compress small samples of articular cartilage while the samples were imaged in a 1.5 T imager. With the use of a piezoelectric piston, the device compressed 1-cm-diameter cylindrical samples of articular cartilage (200 microm) at a rate of 2 Hz. Simultaneously, we imaged the samples with a displacement-sensitive stimulated-echo acquisition mode (STEAM) sequence. We validated the technique using tissue that mimicked silicone samples. We compared the results from the same cartilage samples before and after they were degraded by digestion in trypsin. The extent of degradation was visualized from T(1)-weighted images of the samples after they were soaked in 0.5 mmolar of GdDTPA. The resulting elastographic images show compression and differential strain in directions both parallel and perpendicular to the surface of the cartilage. The static elastographic images that depict compression made before digestion and after 5 and 15 hr of trypsin digestion show that the elastic modulus of the samples decreased with a spatial variation consistent with the enzymatic digestion as revealed by the T(1) images. We believe this technique will be useful in studies of the mechanical properties of articular cartilage and other tissues, and may in the future be extended to the clinical setting.     

Sensitivity encoding for fast MR imaging of the brain in patients with stroke

PURPOSE: To evaluate sensitivity encoding (SENSE) technique in a clinical setting for magnetic resonance (MR) imaging in patients who are suspected of having infarction. MATERIALS AND METHODS: This intraindividual comparative study included 62 patients suspected of having cerebral ischemia. Patients underwent T2-weighted fluid-attenuated inversion-recovery (FLAIR) (n = 62), T2-weighted turbo spin-echo (TSE) (n = 48), and single-shot echo-planar diffusion-weighted imaging (n = 27) with standard sequential and SENSE MR acquisitions with a 1.5-T magnet and phased-array coil. With SENSE, acquisition time was reduced from 1 minute 12 seconds to 35 seconds for FLAIR and from 1 minute 18 seconds to 39 seconds for T2-weighted TSE imaging. For diffusion-weighted imaging, echo train length was shortened (78 vs 71 msec) to reduce susceptibility effects while acquisition time was maintained. Two radiologists scored quality of standard and SENSE images with a five-point scale and assessed presence of artifacts (motion, susceptibility) and lesion conspicuity. To assess statistical significance, Wilcoxon signed rank and chi2 tests were used. RESULTS: Statistical analysis revealed no significant difference in terms of image quality and presence of artifacts between standard and SENSE T2-weighted TSE (image quality, P =.724; presence of artifacts, P =.378) and FLAIR (image quality, P =.127; presence of artifacts, P =.275) images. Image quality at SENSE diffusion-weighted imaging was scored significantly higher compared with that at standard diffusion-weighted imaging (P =.002). Susceptibility artifacts were significantly reduced at SENSE diffusion-weighted imaging when compared with those at standard diffusion-weighted imaging (P <.001). Conspicuity of 84 lesions was rated equivalent with both standard and SENSE protocols. CONCLUSION: SENSE allowed acquisition of T2-weighted TSE and FLAIR images with image quality and lesion conspicuity that did not differ from those of standard acquisition techniques but in only half the acquisition time. Use of SENSE with diffusion-weighted imaging significantly reduces susceptibility artifacts while lesion conspicuity is maintained.     

Diffusion-weighted MR imaging of an acute venous stroke: case report.

A patient with a superior sagittal sinus thrombosis had progressively worsening symptoms and signs that resolved after IV heparin therapy. MR imaging revealed abnormalities in diffusion, similar to those seen with acute arterial stroke. Abnormalities shown on a T2-weighted fast spin-echo and fluid-attenuated inversion recovery images resolved completely. The findings in this report contradict those from previous reports that suggest diffusion-weighted imaging with quantitative apparent diffusion coefficients may be used in selecting patients for dural venous sinus thrombolysis.