Automatic selection of mask and arterial phase images for temporally resolved MR digital subtraction angiography

Abstract

For time-resolved background-subtracted contrast-enhanced magnetic resonance angiography, the bright and sparse arterial signal allows unique identification of contrast bolus arrival in the arteries. This article presents an automatic filtering algorithm using such arterial characterization for selecting arterial phase images and mask images to generate an optimal summary arteriogram. A paired double-blinded comparison demonstrated that this automatic algorithm is as effective as the manual process.Magn Contrast-enhanced magnetic resonance angiography (CEMRA) has become a routine clinical tool for pretreatment mapping of vasculature (1). Among data acquisition techniques for CEMRA, the time-resolved strategy offers a very useful option for many situations because of the elimination of the cumbersome timing of imaging to the contrast bolus arrival (2-13). The time-resolved CEMRA generates time series images in a manner similar to fluoroscopic X-ray angiography, where image postprocessing has been used frequently to improve vasculature display (14,15). For example, from the time-series images all mask images and arterial phase images can be summoned into one image of greater vascular detail with high signal-tonoise ratio (SNR) that is particularly useful for presentation in a surgical operation room where video display may not be available. Linear filtering techniques such as the matched filters can be used to produce a summation image (15,16) and have been attempted in time-resolved or dynamic CEMRA to generate a summary arteriogram (17-21). In practice, the major challenge for summarizing time series images is to identify the contrast bolus arrival and to avoid motion-corrupted mask images and arterial phase images that propagate severe motion artifacts into the final summation image (21). So far, this avoidance of motioncorrupted images and selection of optimal arterial phase and mask images for summation have been performed through a tedious manual procedure (21).Here we report an algorithm that can fully automate the linear filtering process, i.e., to select the set of arterial phase images and the set of mask images such that the subtracted image from the former to the latter is of the best quality. We describe in detail 1) how to quantify the notion of "image quality," and 2) how to effectively select the mask and arterial phase images based on the quantified measure of quality. POSTPROCESSING TECHNIQUES Problem StatementOur approach is basically an automated version of linear filtering that selects a mask image set and an arterial phase image set to generate a linear filtered image. For a given time series of images (I n , n ϭ 1ϳ n max , n max ϭ 35 -40, the total number of images in the time series), indexed by x and y, we want to select the mask image set (M) and the arterial phase image set (A), such that the subtracted or filtered image (S h ) is of "best quality," where S h is given by:where ͉M͉ and ͉A͉ are the number of mask images and arterial phase images, respectively. Note that all the arithmetic ...

Beyond Natural Numbers: Negative Number Representation in Parietal Cortex

“…The field of view was 20 cm, and the matrix size was 64 × 64, providing an in-plane spatial resolution of 3.125 mm. To reduce blurring and signal loss arising from field inhomogeneities, an automated high-order shimming method based on spiral acquisitions was used before acquiring functional MRI scans (Kim et al, 2002). …”

Adaptive subtraction as an aid in MR‐guided placement of catheters and guidewires

“…The proposed technique—adaptive subtraction—automatically selects the most suitable reference image from a moving subset of hitherto acquired dynamic images. This approach shows some similarity with the mask subtraction techniques that have recently been developed for MR angiography (2, 3) and MR temperature mapping (4). By way of in vitro and in vivo experiments it will be demonstrated that adaptive subtraction images, other than conventional subtraction images, are relatively insensitive to slow periodic motion, e.g., respiratory motion, and are only transiently degraded by gross subject motion and alterations of the scan parameters.…”

Automatic algorithm for correcting motion artifacts in time‐resolved two‐dimensional magnetic resoance angiography using convex projections

“…The fully automatic motion filtering presented here used input of contrast arrival frame number to classify precontrast frames for reference. The automated estimate of contrast arrival may be unreliable in the presence of motion (27). The POCS algorithm can be applied in nonautomatic mode, where an operator selects arterial, mask, and motion‐corrupted frames.…”

“…Next, for automatic motion correction, mask and arterial phases were identified using Ref. (27), a new, automatic MRDSA method. We incorporated the proposed POCS algorithm within this automated MRDSA algorithm.…”