Kluwer - Handbook of Biomedical Image Analysis Vol

174	Rakesh Sharma and Avdhesh Sharma

disease [3]. CE-MRA with automated table movement (MoBI-trak) using a 1.5 T superconducting magnet (Philips Gyroscan ACS NT) was equipped with a Power Trak 6000 gradient. Contrast medium (Gd-DTPA) was administered in two sequential boluses-20 cm3 at 0.6 cm3/sec (starting phase) and 20 cm3 at 0.3 cm3/sec (maintenance phase), using a MedRad Spectris automatic injector. DSA was the gold standard and was performed using a Philips Integris 3000, with a brilliance intensifier of 38 sec. DSA and MRA were evaluated on printed films. DSA provided more than 200 diagnostic assessments including stenosis

<50–99% occlusions. CE-MRA MoBI-trak exhibited good sensitivity, specificity, positive and negative predictive values and high diagnostic accuracy. Using this technique MoBI-trak has been shown to be a reliable technique for the detection of peripheral vascular disease up to the trifurcation, although it underlines the necessity for more diagnostic investigation and improvements in the technique.

3.4.4Magnetic Resonance Angiography with Diffusion-Weighted Imaging

This approach was used for intraoperative magnetic resonance imaging, including magnetic resonance angiography and diffusion-weighted imaging [4]. This integrated approach was used to monitor the surgical treatment of a patient with an intracranial aneurysm. Intraoperative imaging was performed with a ceiling-mounted, mobile, 1.5-T magnet (developed in collaboration with Innovative Magnetic Resonance Imaging Systems, Inc., Winnipeg, MB, Canada) that included high-performance 20-mT/m gradients. Preand postclipping, intraoperative, T1-weighted, angiographic, and diffusion-weighted magnetic resonance images were obtained from a patient with an incidental, 8- mm, anterior communicating artery aneurysm. T1-weighted images demonstrated brain anatomic features, with visible shifts induced by surgery. Magnetic resonance angiography demonstrated the aneurysm and indicated that, after clipping, the A1 and A2 anterior cerebral artery branches were patent. Diffusion-weighted studies demonstrated no evidence of brain ischemia. For the first time, intraoperative magnetic resonance imaging was reported to monitor the surgical treatment of a patient with an intracranial aneurysm (see Fig. 3.29).

Advances in Magnetic Resonance Angiography

175

Figure 3.29: Magnetic resonance angiography is shown to demonstrate the aneurysm with both A1 and A2 anterior cerebral artery branches as patent (shown with arrows in left panel on top). Preand post-clipping, intraoperative, T1-weighted, angiographic and diffusion-weighted magnetic resonance images were obtained from a patient with an incidental, 8-mm, anterior communicating artery aneurysm. T1-weighted images showed brain anatomic features with visible shifts induced by surgery.

3.4.5 3D-Navigator Echo MRA

A three-dimensional navigator echo (NE) sequence on an MR scanner with a high performance gradient system was used to evaluate MR coronary angiogsraphy [5]. For imaging, a navigated gradient-echo (GE) pulse sequence with an in-plane resolution between 0.63 × 0.63 and 0.78 × 0.78 mm2 with two overlapping slabs was acquired. The number of visualized coronary artery segments was estimated. In addition, signal-to-noise measurements were performed in the ascending aorta at the level of the proximal right and left coronary arteries.

176	Rakesh Sharma and Avdhesh Sharma

This technique visualized the left main, the right coronary artery up to segment 3, the LAD up to segment 8, and the RCX with segments 11 and 13. The average signal-to-noise value at the level of the right coronary artery was 11.4 ± 5.0, at the level of the left coronary artery 12.3 ± 4.5 with an in-place resolution of 0.63 × 0.63 mm2. This resulted in a too low signal-to-noise ratio so that an adequate assessment of coronary arteries was no longer possible. 3D-MR coronary angiography using the navigator technique is limited by the low signal-to-noise ratio.

3.4.6 Navigator Echo and Cine Gradient-Echo Technique

This technique was used to evaluate coronary artery stents with MR [6]. For both sequences the low-signal artifact was used to localize the stent, whereas the flow-related high signal before and distal to the stent was considered as a potency sign. All the stents were recognized as signal void with GE, and all but one with NE. Positive EET, with a stent on the left anterior descending coronary artery, presented low signal distal to the stent at both MR sequences. These suggested the dysfunction stenosis at conventional coronary angiography (CCA). Two sequential stents on the right coronary artery presented lack of signal distal to the stents at both MR sequences. It suggested occlusion (97% stenosis at CCA). However, negative EET, MR high signal before and distal to the stent suggested patency at both sequences. MR seems to be a safe and promising technique for noninvasive evaluation of coronary stents.

3.4.7 MR Phase-Contrast Doppler Flow Quantification

Determination of blood flow volume is useful in assessing ischemic cerebrovascular disease. Blood flow volume measurement was evaluated by three noninvasive imaging techniques, namely color velocity imaging quantification, spectral Doppler imaging quantification, and MR phase-contrast flow quantification, to see how well the flow values determined by each technique agreed with one another [7]. Flow volume quantification was tested experimentally using a flow simulator and three techniques. These techniques evaluated the vertebral and internal carotid arteries of patients with history of cerebral ischemia. In the flow simulation study, the flow values in each technique were compared with the phantom flow using Wilcoxon’s signed rank test. Flow volumes were measured by color velocity imaging quantification. MR phase-contrast flow quantification

Advances in Magnetic Resonance Angiography

177

agreed with the phantom flow simulation within the tested range, and spectral Doppler imaging quantification values were significantly overestimated. In patients, a large variation in the blood flow volume was obtained between each technique. Blood flow volume measurements determined by the three noninvasive imaging techniques on the same vessel can differ widely, and spectral Doppler imaging quantification consistently overestimated the flow volume. Color velocity imaging quantification or MR phase-contrast flow quantification can be used for clinical follow-up investigations in the future.

3.4.8 Breath-Hold Contrast Enhanced MRA

The purpose of this technique was to evaluate the effect of breathing on image quality of the aortic arch and carotid vessels during contrast-enhanced MR angiography. It showed that high-resolution breath-hold contrast-enhanced MR angiography combined with a timing-bolus technique can produce high-quality images of the entire carotid circulation [8]. High-resolution contrast-enhanced MR angiography on a 1.5-T Siemens imager was used for coronal three-dimensional gradient-echo sequence (TR/TE, 4.36/1.64; flip angle, 25◦) with asymmetric k- space acquisition. The 136 × 512 matrix yielded voxel sizes of 1.33 × 0.64 × 1.0 mm3. A timing-bolus acquisition, orientated in the coronal plane to include the aortic arch, was obtained initially during free-breathing. Twenty milliliters of gadopenetate dimeglumine was injected at 2 mL/sec. Unenhanced and enhanced 3D volumes were recorded. A subtracted 3D set was calculated and subjected to a maximum-intensity-projection algorithm. Half of the patients held their breath during angiography and the other half did not. Aortic arch motion was measured on the timing-bolus acquisition as the distance moved by a single pixel in both the x and y directions. Two observers assessed MIP MR images independently and vessel sharpness was scored on a scale of 1–5. Sharpness was also assessed quantitatively by generating a signal intensity profile across the aortic arch vessel wall and calculating the average of the upward slope and downward slope at full-width half maximum. Visualization of carotid branch vessels was scored on a scale of 0–5, and venous contamination was scored on a scale of 0–3. Average in-plane aortic arch movement was 10.3 mm in the x direction and 8.7 mm in the y direction. Quantitative and qualitative sharpness of the aortic arch and great vessel origins was better ( p < 0.05) during breath-holding than during non- breath-holding. No difference in the sharpness of the carotid vessels was noted

178	Rakesh Sharma and Avdhesh Sharma

between the two groups. Carotid branch vessels were well visualized from the aortic arch to the intracerebral circulation. The average venous contamination score was 0.56. Breath-holding greatly improves the sharpness of the aortic arch and great vessel origins but has no effect on visualization of the carotid vessels. High-resolution breath-hold contrast-enhanced MRA can produce high-quality, artifact-free images of the entire carotid circulation from the aortic arch to the intracerebral circulation.

3.4.9 K-space 3D Navigator-Gated MRA

To acquire the center of k-space while extending three-dimensional freebreathing navigator-gated coronary MRA by an initial single breath-hold [9], this approach was successfully applied. Resulting images were compared with conventionally acquired free-breathing navigator-gated MR angiograms. The acquisition of k-space center during the single breath-hold resulted in an appreciable increase in the signal-to-noise ratio. Visible length of the right coronary artery, as well as contrast-to-noise ratio between blood and the myocardial muscle were identical. The breath-hold extension was shown to be a valuable technique that may be combined with first-pass contrast-enhanced MR imaging. The other advantage was the creation of photographic freeze scans of coronary small arteries and heart within each heartbeat of approximately 80 msec intervals.

3.4.10MR Arterial Spin Tagging Projection Coronary MRA Technique

Conventional coronary MRA techniques display the coronary blood-pool along with the surrounding structures, including myocardium, the ventricular and atrial blood-pool, and the great vessels [10]. This representation of the coronary lumen is not directly analogous to the information provided by X-ray coronary angiography, in which the coronary lumen displayed by iodinated contrast agent is seen. Analogous “luminographic” data may be obtained using MR arterial spin tagging (projection coronary MRA) techniques. Such an approach was implemented using a 2D selective “pencil” excitation for aortic spin tagging in concert with a 3D interleaved segmented spiral imaging sequence with free-breathing and real-time navigator technology. This technique allows for selective 3D

Advances in Magnetic Resonance Angiography

179

visualization of the coronary lumen blood-pool, while signal from the surrounding structures is suppressed. However, there are technical hurdles for visualizing the coronal arteries because of small size and near-constant mobility, which is related to cardiac cycle and normal breathing.

3.4.11Collateral Flow Assessment with Contrast Enhanced MR Velocity Mapping

To correlate quantification of collateral flow in aortic coarctation with the morphological visualization of the collateral vessels and to compare different approaches to the measurement of collateral flow [11], coarctation was examined with T1-weighted spin-echo (T1-W SE) imaging and 3D contrast-enhanced MRA. MR velocity mapping was performed at four levels in the descending aorta. The flow immediately above and below the coarctation did not differ significantly. Measuring within the coarctation resulted in flow overestimation. The increase in flow from proximal to distal aorta was 12 ± 21% in patients with no or uncertain collaterals and 69 ± 55% in patients with pronounced collaterals. Spin-echo images and MRA were comparable in visualizing collateral vessels. The visual estimation of collaterals was correlated reasonably well with flow quantification MR velocity mapping. Collateral flow assessment with MR velocity mapping was an accurate technique for evaluating the hemodynamic importance of a coarctation and was recommended if abundant collaterals are not visualized with spin echo or MRA.

3.4.12(Gd)-Enhanced Three-Dimensional Magnetic Resonance Angiography

The goal of this technique was to evaluate the diagnostic value of gadolinium (Gd)-enhanced three-dimensional MRA in patients with congenital and acquired anomalies of the pulmonary and systemic veins [12]. Gadolinium-enhanced 3D MRA is a fast magnetic resonance imaging technique that has shown great promise in the evaluation of large and medium-sized arteries. However, its application to venous anomalies has not been established. Gd-enhanced 3D MRA examination was used for diagnosis with anomalies of the pulmonary or

180	Rakesh Sharma and Avdhesh Sharma

systemic veins and had additional diagnostic data available for comparison with the MRA findings. The technique did not detect anomalies of the pulmonary veins that were subsequently diagnosed by MRA. Intervention catheterization procedures and operations followed the 3D MRA diagnoses. 3D MRA either diagnosed previously unsuspected venous anomalies or added new clinically important information. The mechanism of pulmonary vein compression in eight patients was determined by MRA but not by other imaging modalities. Gadolinium-enhanced 3D MRA is rapid and accurate. 3D MRA has been shown to be the premier noninvasive technique for imaging large blood vessels in the body.

3.4.133D Time-Resolved Imaging of Contrast Kinetics Technique

This technique based on contrast-enhanced MR angiography was used by direct comparison with the fluoroscopic triggered 3D-elliptical centric view ordering (3D-ELLIP) technique [13]. 3D-TRICKS and 3D-ELLIP were directly compared on a 1.5-T MR unit using the same spatial resolution and matrix. In 3D-TRICKS, the central part of the k-space is updated more frequently than the peripheral part of the k-space, which is divided in the slice-encoding direction. The carotid arteries were imaged using 3D-TRICKS and 3D-ELLIP sequentially in 14 patients. Temporal resolution was 12 sec for 3D-ELLIP and 6 sec for 3D-TRICKS. The signal-to-noise ratio (S/N) of the common carotid artery was measured and the quality of MIP images was then scored in terms of venous overlap and blurring of vessel contours. No significant difference in mean S/N was seen between the two methods. Significant venous overlap was not seen in any of the patients examined. Moderate blurring of vessel contours was noted on 3D-TRICKS in five patients and on 3D-ELLIP in four patients. Blurring in the slice-encoding direction was slightly more pronounced in 3D-TRICKS. However, qualitative analysis scores showed no significant differences. When the spatial resolution of the two methods was identical, the performance of 3DTRICKS was found to be comparable in static visualization of the carotid arteries with 3D-ELLIP, although blurring in the slice-encoding direction was slightly more pronounced in 3D-TRICKS. 3D-TRICKS is a more robust technique than 3D-ELLIP, because 3D-ELLIP requires operator-dependent fluoroscopic triggering. Furthermore, 3D-TRICKS can achieve higher temporal resolution. For the

Advances in Magnetic Resonance Angiography

181

spatial resolution employed in this study, 3D-TRICKS may be the method of choice.

3.4.14 Autocorrected MRA for Motion-Induced Artifacts

This technique was used to investigate the efficacy of a retrospective adaptive motion correction technique known as autocorrection for reducing motion-induced artifacts in high-resolution three-dimensional time-of-flight MR angiography of the circle of Willis [14]. Gradient-recalled echo three-dimensional time-of-flight MR angiography sequence was used for MRA of the circle of Willis. Each volunteer was asked to rotate approximately 2◦ after completion of one third and one half of the acquisition in the axial, sagittal, and oblique planes (45◦ to the axial and sagittal planes). A single static data set was also acquired for each volunteer. Unprocessed and autocorrected maximum-intensity-projection images were reviewed as blinded image pairs by six radiologists and were compared on a five-point image quality scale. Mean improvement in image quality after autocorrection was 1.4 ( p < 0.0001), 1.1 ( p < 0.0001), and 0.2 ( p = 0.003) observer points (maximum value 2.0), respectively, for examinations corrupted by motion in the axial, oblique, and sagittal planes. All three axes had statistically significant improvement in image quality compared with the uncorrected images. Autocorrection can reduce artifacts in motion-corrupted MR angiography of the circle of Willis without distorting motion-free examinations.

3.4.15Multiphase Contrast-Enhanced Magnetic Resonance Angiography

A fast pulse sequence with spiral in-plane readout and conventional 3D partition encoding was reported for multiphase contrast-enhanced magnetic resonance angiography (CE-MRA) of the renal vasculature and compared to a standard multiphase 3D CE-MRA with FLASH readout [15]. An isotropic in-plane spatial resolution of 1.4 × 1.4 mm2 over 2.0 × 1.4 mm2 could be achieved with a high temporal resolution. The theoretical gain of spatial resolution by using the spiral pulse sequence and the performance in the presence of turbulent flow was evaluated in phantom measurements. A deblurring technique corrected the spiral raw data. Thereby, the off-resonance frequencies were determined by

182	Rakesh Sharma and Avdhesh Sharma

Figure 3.30: FFEP MRA.

minimizing the imaginary part of the data in image space. The correction algorithm reduced image blurring substantially in all MRA phases (see Fig. 3.30). The image quality of the spiral CE-MRA pulse sequence was comparable to that of the FLASH CE-MRA with increased spatial resolution and a reduced contrast-to-noise ratio. Additionally, artifacts specific to spiral MRI could be observed that had no impact on the assessment of the renal arteries.

3.4.16High-Resolution MRA with Phase/Frequency Flow Compensation

A newly developed pulse sequence 3D TOF-VTE was tested on clinical MRI systems, by performing scans of the cervical carotid artery and intracranial carotid artery at the carotid siphon. It required very long echo delay times (TE). Variable TE (VTE) was implemented into flow-compensated 3D TOF to minimize the effective TE and reduce the flow-related signal void. The k-space of the 3D TOF was divided into segment groups ranging from two to 32 segments with different TE. The TE were minimized and the flow-compensation gradient lobes were calculated to null the total first moment at the peak of the echo for each segment [16]. Possible artifacts and off-resonance effects were evaluated, with respect to the number of TE segments, using the point spread function

Advances in Magnetic Resonance Angiography

183

(PSF) and corresponding experiments. The optimal number of TE segments for the least artifact was determined to be one-half of the number of slices. Two types of artifacts caused by VTE were predicted and subsequently observed. The signal distribution near the bifurcation and the siphon was much more uniform with VTE, and the flow-related signal loss was greatly reduced (see Fig. 3.30). The resultant MR angiograms provided improved vessel detail. The results show that VTE improved the quality of flow-compensated 3D TOF MRA.

3.4.17Cardiac-Triggered Free-Breathing 3D Balanced Fast Field-Echo Projection MRA

A two-dimensional pencil-beam aortic labeling pulse was developed for the renal arteries [17]. For data acquisition during free breathing in eight healthy adults and seven consecutive patients with renal artery disease, realtime navigator technology was implemented. This technique allows high spatial resolution and high contrast renal MR angiography and visualization of renal artery stenosis without exogenous contrast agent or breath hold (see Fig. 3.31). Initial promising results warrant larger clinical studies.

3.4.18 Cervical MRA

Initial experience with intracranial and cervical MRA at 3.0 T was reported. Phantom measurement s (corrected for relaxation effects) show S/N (3.0T) = 2.14 + / − 0.08 × S/N (1.5 T) in identical–geometry head coils [18]. A 3.0 T TOF intracranial imaging protocol with higher-order autoshimming was developed and compared to 1.5 T 3D TOF in 12 patients with aneurysms. A comparison by two radiologists showed the 3.0 T to be significantly better (P < 0.001) for visualization of the aneurysms (see Fig. 3.29). The feasibility of cervical and intracranial contrast enhanced MR angiography (CEMRA) at 3.0 T was also examined. The relaxivity of the gadolinium contrast agent decreased by only about 4–7% when the field strength was increased from 1.5 T to 3.0 T. Cervical 3.0 T CEMRA was obtained in eight patients available for direct comparison. Image comparison suggested 3.0 T to be favorable field strength for cervical CEMRA. Voxel volumes of 0.62–0.73 mm3 were readily achieved at 3.0 T with the use of single-channel transmit-receive head or cervical coil, a 25 mL