Highly accelerated time-of-flight magnetic resonance angiography using spiral imaging improves conspicuity of intracranial arterial branches while reducing scan time.

OBJECTIVE: To systematically compare time-of-flight magnetic resonance angiography (TOF-MRA) acquired with Compressed SENSE (TOF-CS) to spiral imaging (TOF-Spiral) for imaging of brain-feeding arteries. METHODS: Seventy-one patients (60.2 ± 19.5 years, 43.7% females, 28.2% with pathology) who underwent TOF-MRA after implementation of a new scanner software program enabling spiral imaging were analyzed retrospectively. TOF-CS (standard sequence; duration ~ 4 min) and the new TOF-Spiral (duration ~ 3 min) were acquired. Image evaluation (vessel image quality and detectability, diagnostic confidence (1 (diagnosis very uncertain) to 5 (diagnosis very certain)), quantitative measurement of aneurysm diameter or degree of stenosis according to North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria) was performed by two readers. Quantitative assessments of pathology were compared to computed tomography angiography (CTA) or digital subtraction angiography (DSA). RESULTS: TOF-CS showed higher image quality for intraosseous and intradural segments of the internal carotid artery while TOF-Spiral better depicted small intracranial vessels like the anterior choroidal artery. All vessel pathologies were correctly identified by both readers for TOF-CS and TOF-Spiral with high confidence (TOF-CS (4.4 ± 0.6 and 4.3 ± 0.8), TOF-Spiral (4.3 ± 0.7 and 4.3 ± 0.8)) and good inter-reader agreement (Cohen's kappa > 0.8). Quantitative assessments of aneurysm size or stenosis did not significantly differ between TOF-CS or TOF-Spiral and CTA or DSA (p > 0.05). CONCLUSIONS: TOF-Spiral for imaging of brain-feeding arteries enables reductions in scan time without drawbacks in diagnostic confidence. A combination of spiral imaging and CS may help to overcome shortcomings of both sequences alone and could further reduce acquisition times in the future. KEY POINTS: • TOF-MRA with Compressed SENSE is superior in depicting arteries at the skull base while spiral TOF-MRA is able to better depict small intracranial vessels. • Both TOF-MRA with Compressed SENSE and TOF-MRA with spiral imaging provide high diagnostic confidence for detection of pathologies of brain-feeding arteries. • Spiral TOF-MRA is faster (by 25% for the sequence used in this study) than TOF-MRA with Compressed SENSE, thus enabling clear reductions in scan time for the clinical setting.

Extended Kalman filtering for continuous volumetric MR-temperature imaging.

Real time magnetic resonance (MR) thermometry has evolved into the method of choice for the guidance of high-intensity focused ultrasound (HIFU) interventions. For this role, MR-thermometry should preferably have a high temporal and spatial resolution and allow observing the temperature over the entire targeted area and its vicinity with a high accuracy. In addition, the precision of real time MR-thermometry for therapy guidance is generally limited by the available signal-to-noise ratio (SNR) and the influence of physiological noise. MR-guided HIFU would benefit of the large coverage volumetric temperature maps, including characterization of volumetric heating trajectories as well as near- and far-field heating. In this paper, continuous volumetric MR-temperature monitoring was obtained as follows. The targeted area was continuously scanned during the heating process by a multi-slice sequence. Measured data and a priori knowledge of 3-D data derived from a forecast based on a physical model were combined using an extended Kalman filter (EKF). The proposed reconstruction improved the temperature measurement resolution and precision while maintaining guaranteed output accuracy. The method was evaluated experimentally ex vivo on a phantom, and in vivo on a porcine kidney, using HIFU heating. On the in vivo experiment, it allowed the reconstruction from a spatio-temporally under-sampled data set (with an update rate for each voxel of 1.143 s) to a 3-D dataset covering a field of view of 142.5×285×54 mm(3) with a voxel size of 3×3×6 mm(3) and a temporal resolution of 0.127 s. The method also provided noise reduction, while having a minimal impact on accuracy and latency.

Robust adaptive extended Kalman filtering for real time MR-thermometry guided HIFU interventions.

Real time magnetic resonance (MR) thermometry is gaining clinical importance for monitoring and guiding high intensity focused ultrasound (HIFU) ablations of tumorous tissue. The temperature information can be employed to adjust the position and the power of the HIFU system in real time and to determine the therapy endpoint. The requirement to resolve both physiological motion of mobile organs and the rapid temperature variations induced by state-of-the-art high-power HIFU systems require fast MRI-acquisition schemes, which are generally hampered by low signal-to-noise ratios (SNRs). This directly limits the precision of real time MR-thermometry and thus in many cases the feasibility of sophisticated control algorithms. To overcome these limitations, temporal filtering of the temperature has been suggested in the past, which has generally an adverse impact on the accuracy and latency of the filtered data. Here, we propose a novel filter that aims to improve the precision of MR-thermometry while monitoring and adapting its impact on the accuracy. For this, an adaptive extended Kalman filter using a model describing the heat transfer for acoustic heating in biological tissues was employed together with an additional outlier rejection to address the problem of sparse artifacted temperature points. The filter was compared to an efficient matched FIR filter and outperformed the latter in all tested cases. The filter was first evaluated on simulated data and provided in the worst case (with an approximate configuration of the model) a substantial improvement of the accuracy by a factor 3 and 15 during heat up and cool down periods, respectively. The robustness of the filter was then evaluated during HIFU experiments on a phantom and in vivo in porcine kidney. The presence of strong temperature artifacts did not affect the thermal dose measurement using our filter whereas a high measurement variation of 70% was observed with the FIR filter.